HEALTH TECHNOLOGY ASSESSMENT

Transcription

1 HEALTH TECHNOLOGY ASSESSMENT VOLUME 18 ISSUE 29 MAY 2014 ISSN Rehabiitation aimed at improving outdoor mobiity for peope after stroke: a muticentre randomised controed study (the Getting out of the House Study) Phiippa A Logan, Sarah Armstrong, Tony J Avery, David Barer, Garry R Barton, Janet Darby, John RF Gadman, Jane Horne, Simon Leach, Nadina B Lincon, Samir Mehta, Ossie Newe, Katheen O Nei, Tracey H Sach, Marion F Waker, Hywe C Wiiams, Lisa J Woodhouse and Mat P Leighton DOI /hta18290

2

3 Rehabiitation aimed at improving outdoor mobiity for peope after stroke: a muticentre randomised controed study (the Getting out of the House Study) Phiippa A Logan, 1 * Sarah Armstrong, 2 Tony J Avery, 3 David Barer, 4 Garry R Barton, 5 Janet Darby, 1 John RF Gadman, 1 Jane Horne, 1 Simon Leach, 6 Nadina B Lincon, 7 Samir Mehta, 8 Ossie Newe, 9 Katheen O Nei, 4 Tracey H Sach, 5 Marion F Waker, 1 Hywe C Wiiams, 8 Lisa J Woodhouse 1 and Mat P Leighton 8 1 Division of Rehabiitation and Ageing, University of Nottingham, Nottingham, UK 2 East Midands Research Design Service, University of Nottingham, Nottingham, UK 3 Division of Primary Care, University of Nottingham, Nottingham, UK 4 Gateshead PCT, Gateshead Heath NHS Foundation Trust, Gateshead, UK 5 Facuty of Medicine and Heath Sciences, University of East Angia, Norwich, UK 6 United Linconshire Hospitas NHS Trust, Lincon, UK 7 Institute of Work, Heath and Organisations, University of Nottingham, Nottingham, UK 8 Nottingham Cinica Trias Unit, University of Nottingham, Nottingham, UK 9 Service user iving in Nottingham, *Corresponding author Decared competing interests of authors: Professor Marion Waker is a member of the Goba Stroke Community Advisory Pane (GSCAP), funded by Aergan. Hywe Wiiams is a member of the NIHR Journas Library Board. Pubished May 2014 DOI: /hta18290

4

5 This report shoud be referenced as foows: Logan PA, Armstrong S, Avery TJ, Barer D, Barton GR, Darby J, et a. Rehabiitation aimed at improving outdoor mobiity for peope after stroke: a muticentre randomised controed study (the Getting out of the House Study). Heath Techno Assess 2014;18(29). Heath Technoogy Assessment is indexed and abstracted in Index Medicus/MEDLINE, Excerpta Medica/EMBASE, Science Citation Index Expanded (SciSearch ) and Current Contents / Cinica Medicine.

6

7 Heath Technoogy Assessment HTA/HTA TAR ISSN (Print) ISSN (Onine) Five-year impact factor: Heath Technoogy Assessment is indexed in MEDLINE, CINAHL, EMBASE, The Cochrane Library and the ISI Science Citation Index and is assessed for incusion in the Database of Abstracts of Reviews of Effects. This journa is a member of and subscribes to the principes of the Committee on Pubication Ethics (COPE) ( Editoria contact: nihredit@southampton.ac.uk The fu HTA archive is freey avaiabe to view onine at Print-on-demand copies can be purchased from the report pages of the NIHR Journas Library website: Criteria for incusion in the Heath Technoogy Assessment journa Reports are pubished in Heath Technoogy Assessment (HTA) if (1) they have resuted from work for the HTA programme, and (2) they are of a sufficienty high scientific quaity as assessed by the reviewers and editors. Reviews in Heath Technoogy Assessment are termed systematic when the account of the search appraisa and synthesis methods (to minimise biases and random errors) woud, in theory, permit the repication of the review by others. HTA programme The HTA programme, part of the Nationa Institute for Heath Research (NIHR), was set up in It produces high-quaity research information on the effectiveness, costs and broader impact of heath technoogies for those who use, manage and provide care in the NHS. Heath technoogies are broady defined as a interventions used to promote heath, prevent and treat disease, and improve rehabiitation and ong-term care. The journa is indexed in NHS Evidence via its abstracts incuded in MEDLINE and its Technoogy Assessment Reports inform Nationa Institute for Heath and Care Exceence (NICE) guidance. HTA research is aso an important source of evidence for Nationa Screening Committee (NSC) poicy decisions. For more information about the HTA programme pease visit the website: This report The research reported in this issue of the journa was funded by the HTA programme as project number 08/14/51. The contractua start date was in August The draft report began editoria review in December 2012 and was accepted for pubication in June The authors have been whoy responsibe for a data coection, anaysis and interpretation, and for writing up their work. The HTA editors and pubisher have tried to ensure the accuracy of the authors report and woud ike to thank the reviewers for their constructive comments on the draft document. However, they do not accept iabiity for damages or osses arising from materia pubished in this report. This report presents independent research funded by the Nationa Institute for Heath Research (NIHR). The views and opinions expressed by authors in this pubication are those of the authors and do not necessariy refect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Heath. If there are verbatim quotations incuded in this pubication the views and opinions expressed by the interviewees are those of the interviewees and do not necessariy refect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Heath. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. Pubished by the NIHR Journas Library ( produced by Prepress Projects Ltd, Perth, Scotand (

8 Editor-in-Chief of Heath Technoogy Assessment and NIHR Journas Library Professor Tom Waey Director, NIHR Evauation, Trias and Studies and Director of the HTA Programme, UK NIHR Journas Library Editors Professor Ken Stein Chair of HTA Editoria Board and Professor of Pubic Heath, University of Exeter Medica Schoo, UK Professor Andree Le May Chair of NIHR Journas Library Editoria Group (EME, HS&DR, PGfAR, PHR journas) Dr Martin Ashton-Key Consutant in Pubic Heath Medicine/Consutant Advisor, NETSCC, UK Professor Matthias Beck Chair in Pubic Sector Management and Subject Leader (Management Group), Queen s University Management Schoo, Queen s University Befast, UK Professor Aieen Carke Professor of Pubic Heath and Heath Services Research, Warwick Medica Schoo, University of Warwick, UK Dr Tessa Criy Director, Crysta Bue Consuting Ltd, UK Dr Peter Davidson Director of NETSCC, HTA, UK Ms Tara Lamont Scientific Advisor, NETSCC, UK Professor Eaine McCo Director, Newcaste Cinica Trias Unit, Institute of Heath and Society, Newcaste University, UK Professor Wiiam McGuire Professor of Chid Heath, Hu York Medica Schoo, University of York, UK Professor Geoffrey Meads Professor of Heath Sciences Research, Facuty of Education, University of Winchester, UK Professor Jane Norman Professor of Materna and Feta Heath, University of Edinburgh, UK Professor John Powe Consutant Cinica Adviser, Nationa Institute for Heath and Care Exceence (NICE), UK Professor James Raftery Professor of Heath Technoogy Assessment, Wessex Institute, Facuty of Medicine, University of Southampton, UK Dr Rob Riemsma Reviews Manager, Keijnen Systematic Reviews Ltd, UK Professor Heen Roberts Professoria Research Associate, University Coege London, UK Professor Heen Snooks Professor of Heath Services Research, Institute of Life Science, Coege of Medicine, Swansea University, UK Pease visit the website for a ist of members of the NIHR Journas Library Board: Editoria contact: nihredit@southampton.ac.uk NIHR Journas Library

9 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Abstract Rehabiitation aimed at improving outdoor mobiity for peope after stroke: a muticentre randomised controed study (the Getting out of the House Study) Phiippa A Logan, 1 * Sarah Armstrong, 2 Tony J Avery, 3 David Barer, 4 Garry R Barton, 5 Janet Darby, 1 John RF Gadman, 1 Jane Horne, 1 Simon Leach, 6 Nadina B Lincon, 7 Samir Mehta, 8 Ossie Newe, 9 Katheen O Nei, 4 Tracey H Sach, 5 Marion F Waker, 1 Hywe C Wiiams, 8 Lisa J Woodhouse 1 and Mat P Leighton 8 1 Division of Rehabiitation and Ageing, University of Nottingham, Nottingham, UK 2 East Midands Research Design Service, University of Nottingham, Nottingham, UK 3 Division of Primary Care, University of Nottingham, Nottingham, UK 4 Gateshead PCT, Gateshead Heath NHS Foundation Trust, Gateshead, UK 5 Facuty of Medicine and Heath Sciences, University of East Angia, Norwich, UK 6 United Linconshire Hospitas NHS Trust, Lincon, UK 7 Institute of Work, Heath and Organisations, University of Nottingham, Nottingham, UK 8 Nottingham Cinica Trias Unit, University of Nottingham, Nottingham, UK 9 Service user iving in Nottingham *Corresponding author Pip.ogan@nottingham.ac.uk Background: One-third of stroke patients are dependent on others to get outside their homes. This can cause peope to become housebound, eading to increased immobiity, poor heath, isoation and misery. There is some evidence that outdoor mobiity rehabiitation can reduce these imitations. Objective: To test the cinica effectiveness and cost-effectiveness of an outdoor mobiity rehabiitation intervention for stroke patients. Design: Muticentre, parae-group randomised controed tria, with two groups aocated at a 1 : 1 ratio pus quaitative participant interviews. Setting: Fifteen UK NHS stroke services throughout Engand, Scotand and Waes. Participants: A tota of 568 stroke patients who wished to get out of the house more often, mean age of 71 years: 508 reached the 6-month foow-up and 10 were interviewed. Intervention: Contro was deivered prior to randomisation to a participants, and consisted of verba advice and transport and outdoor mobiity eafets. Intervention was a targeted outdoor mobiity rehabiitation programme deivered by 29 NHS therapists to 287 randomy chosen participants for up to 12 sessions over 4 months. Main outcome measures: Primary outcome was participant heath-reated quaity of ife, measured by the Short Form questionnaire-36 items, version 2 (Socia Function domain), 6 months after baseine. Secondary outcomes were functiona abiity, mobiity, number of journeys (from monthy trave diaries), satisfaction with outdoor mobiity (SWOM), psychoogica we-being and resource use [heath care and Persona Socia Services (PSS)] 6 months after baseine. Carer we-being was recorded. A outcome Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. vii

10 ABSTRACT measures were coected by post and repeated 12 months after baseine. Outcomes for the groups were compared using statistica significance testing and adjusted for mutipe membership to account for the effect of mutipe therapists at different sites. Interviews were anaysed using interpretive phenomenoogy to expore confidence. Resuts: A median of seven intervention sessions [interquartie range (IQR) 3 7 sessions], median duration of 369 minutes (IQR minutes) per participant was deivered. There was no significant difference between the groups on heath-reated quaity of ife (socia function). There were no significant differences between groups in functiona abiity, psychoogica we-being or SWOM at 6- or 12-month foow-ups. There was a significant difference observed for trave journeys with the intervention group being 42% more ikey to make a journey compared with the contro group [rate ratio 1.42, 95% confidence interva (95% CI) 1.14 to 1.67] at 6 months and 76% more ikey (rate ratio 1.76, 95% CI 1.36 to 1.95) at 12 months. The number of journeys was affected by the therapist effect. The mean incrementa cost (tota NHS and PSS cost) of the intervention was (95% CI to ), with an incrementa quaity-adjusted ife-year gain of (95% CI to 0.007) according to the European Quaity of Life-5 Dimensions and (95% CI to 0.006) according to the Short Form questionnaire-6 Dimensions. At baseine, 259 out of 281 (92.2%) participants in the contro group were dissatisfied with outdoor mobiity but at the 6-month assessment this had reduced to 77.7% (181/233), a 15% reduction. The corresponding reduction in the intervention group was sighty greater (21%) than 268 out of 287 (93.4%) participants dissatisfied with outdoor mobiity at baseine to 189 out of 261 (72.4%) at 6 months. Participants described osing confidence after stroke as being detrimenta to outdoor mobiity. Recruitment and retention rates were high. The intervention was deiverabe by the NHS but had a neutra effect in a areas apart from potentiay increasing the number of journeys. This was dependent on the therapist effect, meaning that some therapists were more successfu than others. The contro appeared to affect change. Concusions: The outdoor mobiity intervention provided in this study to these stroke patients was not cinicay effective or cost-effective. However, the provision of personaised information and monthy diaries shoud be considered for a peope who wish to get out more. Tria registration: Current Controed Trias ISRCTN Funding: This project was funded by the NIHR Heath Technoogy Assessment programme and wi be pubished in fu in Heath Technoogy Assessment; Vo. 18, No. 29. See the NIHR Journas Library website for further project information. viii NIHR Journas Library

11 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Contents List of tabes...xiii List of figures...xv List of boxes.... xvii List of abbreviations....xix Scientific summary...xxi Chapter 1 Introduction 1 Introduction 1 The effect of stroke on functioning, and rehabiitation 1 Outdoor mobiity after a stroke, and rehabiitation 1 Justification for the current study 2 Structure of the Heath Technoogy Assessment monograph 3 Chapter 2 Randomised controed study methods 5 Summary of study design 5 Primary objective 5 Secondary objectives 5 Primary outcome measure 6 Secondary outcome measures 6 Recruitment process 7 Sampe size 7 Revised sampe size 8 Setting and ocations 8 Primary care community stroke rehabiitation service 8 Secondary care community stroke rehabiitation service 9 Secondary care stroke service 9 Eigibiity criteria 9 Screening and baseine assessment 9 Consent 9 Contro information 10 Randomisation 10 Intervention 11 Training therapists to provide the intervention 11 Data coection 11 Eectronic case report form 12 Questionnaire bookets 12 Trave diary 12 Intervention records 12 Treatment fideity 13 Adverse events/safety evauation 13 Conceament of aocation during outcome assessments 13 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. ix

12 CONTENTS Participant retention and withdrawas 14 Statistica methods 14 Comparison between treatment arms 14 Descriptive anayses 15 Anaysing primary and secondary outcomes 15 Sensitivity anaysis adjusting for missing data 15 Mutipe membership random effect 16 Presentation of primary and secondary resuts 16 Exporatory/other anayses 16 Compiance in reference to CONSORT 17 Study management and oversight 17 Tria Management Group 17 Tria Steering and Data Committee 17 Amendments to the study 17 Chapter 3 Resuts: randomised controed study 19 Study recruitment and foow-up 19 Foow-up (CONSORT) 19 Carer questionnaire foow-up 22 Trave diary foow-up 22 Baseine characteristics 22 Deivery of the intervention 24 Number of therapists by site 24 Number and duration of intervention sessions 24 Description of set goas 25 Description of the key eements of the intervention 25 Treatment fideity 25 Competion of intervention 27 Contamination 27 Numbers anaysed 27 Primary outcome 28 Sensitivity anayses of the primary outcome 28 Secondary outcomes at 6 months 28 Outcome measures at 12 months 29 Robustness of resuts 30 Sensitivity anayses 30 Exporatory/other anayses 30 Nottingham Extended Activities of Daiy Living by category 30 Number of sessions by site 30 Number of trave journeys by intervention session 30 Fas by age category and aocation 31 Comparing changes in satisfaction with outdoor mobiity over time 32 Six-month foow-up by foow-up approach 32 Adverse events 33 Protoco deviations 33 Conceament of aocation 34 Summary 34 x NIHR Journas Library

13 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 4 Economic evauation: methods and resuts 35 Heath economic evauation: methods used 35 Methods 35 Measuring costs 35 Intervention 35 Other NHS and Persona Socia Services costs 36 Base-case anaysis 39 Resuts 41 Costs 41 Overa and incrementa costs 45 Outcomes 45 Base-case anaysis 45 Chapter 5 Quaitative study: methods and resuts 49 Introduction 49 Aims 49 Samping strategy and recruitment 49 Participant sampe characteristics 49 Data coection: interviewing approach 49 Data anaysis 50 Presentation of the findings 50 Robbed of ife 51 Fear of having another stroke 52 Fear of going out/socia confidence 52 Team confidence/coective efficacy 53 Roe confidence 54 It s not I can t, it s I can : ski mastery 54 Inner strength and confidence 54 Summary 55 Chapter 6 Discussion 57 Key findings 57 Intervention deivery 57 Effect of the intervention on heath-reated quaity of ife 58 Effect of the intervention on outdoor mobiity participation 59 Effect of the contro on outdoor mobiity 60 Effect of the outcome on functiona abiity and psychoogica distress 60 Was the intervention cost-effective? 60 Effect of the intervention on fas 60 What did we earn about confidence after stroke? 61 Adverse events 61 Methodoogica issues 61 Comparison with other studies 61 Generaisabiity of findings 61 Strengths and imitations 62 Economic evauation 62 Quaitative study 62 Choice of outcome measures 62 Participant identification and randomisation 62 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xi

14 CONTENTS Coection of foow-up data 63 Issues of unbinding 63 Differentia foow-up 63 Attention contro 64 Therapist and site effect 64 Chapter 7 Concusions 65 What the study found 65 Impications for cinica practice 65 Recommendations for research 65 Acknowedgements 67 References 69 Appendix 1 Statistica anaysis pan 75 Appendix 2 Summary of intervention training manua 109 xii NIHR Journas Library

15 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 List of tabes TABLE 1 Breakdown of participant identification, screening and recruitment numbers from two aternative contact approaches 20 TABLE 2 Proportion of who competed baseine and foow-up questionnaires 22 TABLE 3 Trave diary foow-up rates 22 TABLE 4 Participant baseine characteristics 23 TABLE 5 Summary of type of goas set with participants in the intervention group 26 TABLE 6 Summary of type of intervention and the number of sessions the intervention type was deivered 26 TABLE 7 Numbers anaysed for each outcome measure 27 TABLE 8 Primary outcome anaysis at 6 months 28 TABLE 9 Secondary outcomes at 6 months spit by adjusted measure of outcome 28 TABLE 10 Secondary outcomes at 12 months spit by adjusted measure of outcome 29 TABLE 11 Percentage intervention competed to satisfaction by site and average number of session deivered 31 TABLE 12 Number of journeys made within the intervention group by number of intervention sessions 31 TABLE 13 Fas data summary by age 32 TABLE 14 Fas data summary by treatment aocation 32 TABLE 15 Proportion of questionnaire received according to approach used, at 6 months 33 TABLE 16 Awareness of treatment aocation of RAs prior to, or, during RA-assisted participant visits of 6 months primary outcome coection 34 TABLE 17 Description of the costs associated with the intervention and contro 38 TABLE 18 Estimated eves of resource use and associated cost (mean per participant over the 12-month period) 42 TABLE 19 Unit costs attached to different items of resource use, with associated source 43 TABLE 20 Estimates of the mean cost ( ) and QALYs associated with each intervention over the 12-month study period 44 TABLE 21 Base-case and sensitivity anayses 47 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xiii

16

17 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 List of figures FIGURE 1 Study design fow chart. GP, genera practice 6 FIGURE 2 Recruitment rate 19 FIGURE 3 Fina CONSORT diagram. a, Participant remained in the tria but no data were coected at that time point 21 FIGURE 4 Range of intervention visits 25 FIGURE 5 Cost-effectiveness acceptabiity curve for the intervention (green ine) and contro group (back ine) (base case for EQ-5D data) 46 FIGURE 6 Cost-effectiveness acceptabiity curve for the intervention (green ine) and contro group (back ine) (base case for SF-6D data) 46 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xv

18

19 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 List of boxes BOX 1 Exampes of what confidence means 51 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xvii

20

21 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 List of abbreviations 95% CI 95% confidence interva CEAC CI CONSORT DMC ecrf EQ-5D GHQ-12 GP HTA ICC ICER INB IPA IQR MCID MI mph NCTU NEADL cost-effectiveness acceptabiity curve chief investigator Consoidated Standards of Reporting Trias Data Monitoring Committee eectronic case report form European Quaity of Life-5 Dimensions Genera Heath Questionnaire-12 items genera practice Heath Technoogy Assessment intracass correation coefficient incrementa cost-effectiveness ratio incrementa net benefit interpretive phenomenoogy anaysis interquartie range minimay cinicay important difference mutipe imputation mies per hour Nottingham Cinica Trias Unit Nottingham Extended Activities of Daiy Living NHS-IC NICE NIHR PCT PI PIS PROM PSS QALY RA RCT RMI SD SF-36v2 SF-6D SWOM TMG TSC TSDC NHS Information Centre Nationa Institute for Heath and Care Exceence Nationa Institute of Heath Research primary care trust principa investigator patient information sheet patient-reported outcome measure Persona Socia Services quaity-adjusted ife-year research assistant randomised controed tria Rivermead Mobiity Index standard deviation Short Form questionnaire-36 items, version 2 Short Form questionnaire-6 Dimensions satisfaction with outdoor mobiity Tria Management Group Tria Steering Committee Tria Steering and Data Committee Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xix

22

23 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Scientific summary Background Impaired mobiity affects haf of a stroke patients, with one-third sti being dependent on others to get outside 6 months after stroke. As a resut, stroke patients become housebound, eading to increased immobiity, poor heath, isoation and misery. This diminution of quaity of ife provides the justification for an intervention aimed at enhancing outdoor mobiity for those with mobiity restrictions. Stroke guideines do not contain evidence-based recommendations to treat patients who have outdoor mobiity imitations. In the UK, the routine care for outdoor mobiity imitations is provision of eafets and verba information. A Cochrane review concuded that the passive provision of eafets is not effective. An outdoor mobiity intervention for peope with stroke was deveoped and evauated in a singe-centre piot randomised controed study. This piot study found cear benefits in peope who received the intervention, with 65% being abe to get out of the house as much as they wished compared with 35% in the contro group. In addition, the participants who received the intervention took significanty more journeys. Aim Athough the intervention and its components in reation to the care pathways and socia and heath contexts were we deveoped, it was not cear whether: 1. the benefits coud be repicated by other therapists across the heath system 2. the intervention coud improve heath-reated quaity of ife 3. such an approach was cost-effective. Design A muticentre, mutitherapist, mutisetting, parae-group randomised controed study [randomised controed tria (RCT)] with economic evauation and nested quaitative study. Randomisation was provided by the Nottingham Cinica Trias Unit (NCTU). The groups were compared at 6 and 12 months, accounting for baseine differences and adjusting for the mutipe membership random effects caused by having numerous therapists deivering the intervention in severa sites. The incrementa cost-effectiveness of the intervention compared with contro, was anaysed using a UK NHS and Persona Socia Services (PSS) perspective. Quaitative interviews were competed with a subset of intervention participants to expore feeings of confidence. Setting The study was conducted within the stroke pathway in 15 NHS stroke services throughout Engand, Scotand and Waes. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xxi

24 SCIENTIFIC SUMMARY Participants Peope who had experienced a stroke were identified between November 2009 and August 2011 through genera practices, primary care therapy teams, community stroke teams or outpatient cinics. Overa, 11,126 patient invitations were sent, with 1448 (13%) interested peope repying. Research assistants contacted respondents to arrange 852 (8%) baseine visits. Peope were eigibe if they provided written informed consent, were > 18 years od, had a stroke at east 6 weeks previousy and wished to get out of the house more often. Peope were not eigibe if they were not abe to compy with the protoco and therapy programme or if they were in active rehabiitation. In tota, 568 invited peope (5.1%) were eigibe to take part and were randomy aocated to the rehabiitation intervention group (287) or contro group (281). Participants ranged in age from 32 to 96 years [mean of 71.6 years, standard deviation (SD) 12.1 years] and the time since stroke was from 2 to 479 months (mean of 40 months, SD 52.7 months). Overa, 253 (44%) were men and 315 women (56%). Ten intervention participants took part in the quaitative study. Intervention and contro A participants received the contro, which was verba advice and eafets given at the baseine assessment visit. Research assistants at each site provided a personaised pack of oca trave information containing, for exampe, eafets and bus timetabes. The intervention group received additiona face-to-face rehabiitation from NHS therapists, up to 11 times over 4 months. This was a mixture of exercise and practica activities to increase outdoor mobiity; psychoogica interventions to improve confidence and targeted information with therapist training; and a treatment manua. Treatment fideity was assessed by a research assistant in 10% of intervention sessions, who compared treatment records and treatment sessions with a checkist. Main outcome measures Outcomes were coected 6 and 12 months after recruitment and by monthy trave diaries. The primary outcome was heath-reated quaity of ife, as measured by Socia Function domain score from the Short Form questionnaire-36 items, version 2 (SF-36v2) at 6 months foow-up. These two questions ask participants to rate their socia activity away from home. The secondary outcome measures at 6 and 12 months were (1) functiona activity using the Nottingham Extended Activities of Daiy Living Scae; (2) mobiity using the Rivermead Mobiity Index; (3) the number of journeys (trave journeys) made outside the house, using participant-competed trave diaries; (4) satisfaction with outdoor mobiity using a yes/no question: Do you get out of the house as much as you woud ike? ; (5) psychoogica we-being using the Genera Heath Questionnaire-12 items (GHQ-12); (6) carer psychoogica we-being using the GHQ-12; (7) quaity-adjusted ife-years (QALYs) using the European Quaity of Life-5 Dimensions (EQ-5D) and Short Form questionnaire-6 Dimensions (SF-6D); (8) resource use (NHS, PSS, carer input and some patient-borne costs); and (9) participant mortaity coected from NHS Information Centre (NHS-IC)/NHS Centra Register. The statistica methods of anaysis were stated prior to the start of the study. A main outcomes were presented descriptivey by group and then the resuts of the mutivariabe anaysis adjusting for age, baseine outcome vaue, therapy effect and site effect were presented as differences in means for continuous data, odds ratios for binary data or rate ratios for count data (adjusted resuts), with 95% credibe intervas as a measure of significance. The anaysis was adjusted for therapist effect xxii NIHR Journas Library

25 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 and site effect, as recommended for trias of compex interventions in which one participant might receive the intervention from a number of therapists. Site and therapy effect sizes were presented for each of the outcomes, apart from trave journeys (anaysed as rate ratios) for which these were not cacuabe. No serious adverse events were recorded. Study-specific adverse events were coected from the intervention group: any fa that resuted in assistance of a heath-care professiona. In addition, a participants had the opportunity to record fas on the trave diary. Interview study Ten intervention participants were interviewed. A semistructured interview was used to guide a digitay recorded interview in the participants own homes. The interviews were transcribed and anaysed using interpretive phenomenoogy by two researchers. Resuts Foow-up rates In tota, 264 out of 287 (92%) intervention participants and 239 out of 281 (85%) contro participants competed the 6-month foow-up, and 232 out of 287 (81%) intervention participants and 211 out of 281 (75%) contro participants competed the 12-month foow-up. The differences in foow-up rates between the two groups did not affect the power of the study. A tota of 192 carers competed the baseine assessment: 148 out of 192 (77%) competed the 6-month foow-up and 127 out of 192 (66%) competed the 12-month foow-up. Foow-up was competed in August Characteristics of the participants The two groups were we matched in age, ethnicity, residence, functiona abiity, functiona abiity and psychoogica we-being. Time from stroke to recruitment was 6 months ess in the contro group (mean 37 months vs. 43 months) than the intervention group and there were more men in the contro group (47%) than intervention group (42.2%). Adjustment for gender had no effect on the primary outcome. Characteristics of the intervention In tota, 29 therapists deivered the intervention a median of seven times [interquartie range (IQR) 3 11 times] per participant, with a median duration of minutes (IQR minutes). The intervention was competed satisfactoriy 67.3% of the time and deivered 100% of the time according to the protoco. Primary outcome measure The variabiity of the socia function score was simiar in the two groups, athough the mean score was sighty higher in the intervention group (47.0) at 6 months compared with the contro group (43.9). The adjusted difference in means between groups was 4.630, with a 95% credibe interva of to Secondary outcome measures No significant difference was observed for the secondary outcome measures at 6 or 12 months for psychoogica distress, activity, mobiity or satisfaction with outdoor mobiity (52 of the contro group said yes, 72 of the intervention group said yes ). Adjusting for therapist and site effect did not affect these resuts. Participants in the intervention group took more journeys than the contro group when the resuts were adjusted for therapist and site effect. Those in the intervention group were 42% more ikey to make a journey than those in the contro group at 6 months [rate ratio 1.42, 95% confidence interva (95% CI) 1.14 to 1.67] and 76% were more ikey to make a journey at 12 months (rate ratio 1.76, 95% CI 1.36 to 1.95). Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xxiii

26 SCIENTIFIC SUMMARY Economic evauation In the base-case anaysis, the mean incrementa cost (tota NHS and PSS cost) of the intervention was (95% CI to ) with an incrementa QALY gain of (95% CI to 0.007) according to the EQ-5D and (95% CI to 0.006) according to the SF-6D. Thus, the intervention was not estimated to be cost-effective compared with the contro. The probabiity that the intervention was cost-effective, compared with the contro, was 5.2% at a threshod of 20,000 per QALY (based on the EQ-5D). The sensitivity anayses support these concusions, as the 95% CI surrounding the incrementa net benefit was never whoy positive at the 20,000-per-QALY threshod. Exporatory anaysis Effect of the contro There was very strong evidence that the contro group improved markedy. At baseine 259 out of 281 (92.2%) participants were dissatisfied with outdoor mobiity but at the 6-month assessment this had reduced to 78% (160/205), a 15% reduction. The corresponding reduction in the intervention group was sighty greater (18%), with 268 out of 287 (93.4%) expressing dissatisfaction with outdoor mobiity at baseine and 17 out of 227 (75.5%) expressing this at 6-month assessment. Fas The tota group had a median of three (IQR 1 6.5) fas per year. Fa rates between the contro and intervention groups were the same. Quaitative interview study A participants said they understood the term confidence but found it difficut to describe. However, they were abe to describe how oss of confidence had robbed them of identity, made them fearfu, made them reiant on others, and that they had ost their roe, ost skis and fet ow sef-worth. Interviews identified that fear of faing and fear of another stroke was a huge barrier to participating in outdoor mobiity. Avoidance behaviours further imited competence and confidence in these activities. Confidence appeared to have a tempora component and increased confidence in one domain impacted on other areas. Meaningfu roes, such as, gardener, cook or driver, or engagement in repacement roes such as vounteer or card maker, were associated with a positive increased confidence. Socia confidence, fear of socia interactions and stigma were identified as eading to a poor psychosocia outcome. Discussion Main findings An intervention provided by NHS therapists to improve outdoor mobiity neither improved heath-reated quaity of ife (socia function) or any other heath outcomes measured nor did it prove more cost-effective than the contro intervention. There was strong evidence that some participants were abe to make significanty more journeys and that personaised outdoor mobiity information and sef-competed daiy trave diaries coud improve satisfaction with outdoor mobiity (SWOM). Stroke patients wish to improve their confidence. xxiv NIHR Journas Library

27 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 We concude that at present the intervention evauated in this study does not improve outdoor mobiity for a stroke patients. However, the provision of personaised information and monthy diaries shoud be considered for a peope who wish to get out more and some patients coud increase their outdoor mobiity through face-to-face intervention sessions from therapists with certain skis. Research in context to other studies There is ony one study to which we can compare the resuts. This was the singe-centre study on which this muticentre study was based. The intervention provided in the muticentre study was simiar in number and duration to that in the singe-centre study [median of six sessions (IQR 4 6 sessions), duration 240 minutes (IQR minutes)] but was deivered by 29 therapists over 15 sites instead of one therapist in one site. The participants were simiar in age. The two studies differed in two areas. Contro-group participants in the present study were provided with personaised outdoor mobiity information and monthy trave diaries. This was an augmented version of that provided in the singe-centre study and competey different to routine care ate after stroke, which woud be no intervention. The second difference was participants in the singe-site study were 1 year after stroke, whereas participants in the present study were 3.5 years after stroke. Strengths and imitations Strengths Recruitment rates were consistent over the duration of the study, the target sampe size was reached, participant retention rates were high, eigibiity criteria was incusive and the intervention was deivered in a pragmatic manner by NHS therapists in numerous ocations, making the resuts generaisabe across the UK. A strength was that the sampe size and between-group comparisons were anaysed using a therapist and site adjustment. This type of anaysis is not yet routiney used but is recommended for this type of tria. We accept that the ony one secondary outcome became significanty in favour of the intervention when the adjustment was appied, so the resuts need to be read with caution. Limitations A imitation of this study was that the contro participants were inadvertenty provided with an intervention that may have affected a change in the SWOM scores. Another imitation was that no process evauation was competed to et us understand how some therapists were abe to increase journeys. Cinica impications Stroke patients are not getting out of the house as much as they woud ike. An outdoor mobiity intervention can be deivered by NHS occupationa therapists and physiotherapists at a range of geographica ocations. There was no observed improvement in heath-reated quaity of ife (socia function), psychoogica distress and functiona activity in the intervention group over the contro group. The intervention was not cost-effective compared with contro. Some therapist/participant combinations are more successfu than others. The provision of information and the daiy competion of trave diaries improved participant SWOM. Outdoor mobiity does not cause fas. Loss of confidence can affect outdoor mobiity. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. xxv

28 SCIENTIFIC SUMMARY Research impications Recruitment, retention and foow-up rates provide evidence that high-quaity stroke rehabiitation research can be competed. Muticentre studies of rehabiitation interventions are possibe. Stroke patients are abe to successfuy compete outcome assessments and monthy trave diaries. Using data adjustments for therapist and site effect needs further exporation. The contro intervention needs evauation. The reationship between number of journeys and quaity of ife needs examining. Tria registration This study is registered as ISRCTN Funding Funding for this study was provided by the Heath Technoogy Assessment programme of the Nationa Institute for Heath Research. xxvi NIHR Journas Library

29 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 1 Introduction Introduction In 2009 the Nationa Institute for Heath Research (NIHR) Heath Technoogy Assessment (HTA) programme funded a proposa to investigate the cinica effectiveness and cost-effectiveness of deivering a targeted outdoor mobiity therapy provided within stroke services aimed at heping peope get out of the house more often foowing a stroke. This report detais the resuts of the Getting out of the House Study. The effect of stroke on functioning, and rehabiitation Stroke can have a devastating effect on peope s ives, with haf of survivors being dependent on others 6monthsater, 1 one-third feeing sociay isoated, one-quarter having abnorma moods, and haf not getting out of their houses as much as they woud ike. 2 These are wordwide heath-care issues but, with 130,000 peope a year having a stroke in Engand and Waes, 3 and stroke increasing with age, the number of peope needing UK heath and socia care services is set to rise. The cost of stroke to the UK NHS is estimated to be over 2.5B per year 3 in addition to the cost to socia care services and to the stroke survivors. In the UK most peope who have a stroke are admitted to hospita for acute care but are discharged on average 19.5 days ater. 4 Treatment is then transferred to community-based teams and socia care services. Stroke unit care has been found to be the best way to organise inpatient care for stroke patients, 5 aowing them to be transferred to evidenced-based Eary Supported Stroke Discharge services 6 when they go home. Once home, stroke rehabiitation interventions provided by occupationa therapists have been shown to increase activity 7,8 by heping peope to regain independence in persona and domestic activities of daiy iving. Most stroke rehabiitation research has been conducted on peope in the first year after their stroke. Athough many consequences of stroke persist beyond a year there is itte evidence for the benefit of interventions deivered after 1 year. 9 The eary treatment and rehabiitation of stroke necessariy focuses on reducing neuroogica impairments but as time goes by after a stroke the activity and participation become more important. Aso, the first goa of patients in hospita tends to be to get home, which requires a focus on sef-care activities, but, once home, they set goas in terms of being abe to optimise participation, and these tend to require an abiity to get out of the house or to engage in socia activities. However, these rehabiitation areas are rarey targeted. Outdoor mobiity after a stroke, and rehabiitation Stroke can eave peope with ong-term and persistent impairments, eading to activity imitations and restriction in participation. These restrictions occur in addition to other impairments resuting from comorbidities, such as osteoarthritis or dementia, making recovery and evauation of treatments difficut and compicated. In addition, stroke often eaves peope with perceptua and cognitive impairments that can reduce performance of everyday activities, such as shopping, hobbies, cooking or using the bus. Lowering of mood, feeings of becoming a different person and ack of confidence are aso common after stroke. 10 These hidden impairments often go untreated and, aongside the more obvious physica ones, contribute to activity imitations especiay when peope are discharged from heath services. There is evidence that stroke patients may deay getting back to a norma ife, even athough they may have made a good physica recovery. 11 Hestrom et a. 12 argued it is not unti peope go home after a stroke and attempt to try everyday activities that the rea impact is reaised. 12 In response to this, it is recommended that a stroke patients receive a 6-month review, with rehabiitation being avaiabe to enabe peope to maintain Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 1

30 INTRODUCTION participation, reduce ong-term misery and improve quaity of ife. These are priority areas identified in the Department of Heath of Engand and Waes Stroke Strategy. 13 Limitations in outdoor mobiity affect 42% of stroke patients, 2 with 33% remaining permanenty dependent on others ong term. 3 Stroke patients can become housebound, eading to poor menta heath, isoation and misery. 14 This can be devastating for patients and carers, and increase their reiance on home visits from medica staff, home care, socia services and home adaptations. Stroke tends to affect oder peope, median age 77 years [interquartie range (IQR) years], 15 and, as outdoor mobiity incuding waking outside decines with age, mobiity disorders may aso be due to impairments and socia factors associated with ageing and comorbidity. Peope aged 80 years of age make haf the number of outdoor journeys and trave ess than one-quarter of the distance of those aged years. 16 There is evidence that edery peope find it difficut to participate in transport services, owing to an inabiity to carry heavy oads and a fear of crime when outside at night. 14 Peope who are dependent on waking frames are, on average, getting out of the house ess than twice each week. 17 Outdoor mobiity is usuay achieved by a combination of waking, cars and pubic transport, with very few peope using speciaist transport such as the provision of a taxi service for peope with disabiities. 16,18 Specific outdoor mobiity rehabiitation interventions exist for peope with imited vision, earning disabiities and those who use wheechairs 19 but these are not routiney used for peope who have had a stroke. The Nationa Cinica Stroke Guideines do not contain any evidence-based recommendations to guide therapists in how to treat outdoor mobiity imitations. 20 However, the cinica guideines recommend that stroke patients are encouraged to take physica exercise, as there is an increasing body of evidence that exercise is of benefit to stroke survivors. 21 In the UK the routine treatment for outdoor mobiity probems is to provide information in a eafet and verbay, deivered by a heath-care professiona. A Cochrane review 22 concuded that the passive provision of eafets is not effective and eafets can be difficut to understand after a stroke. It is more ikey that a mutimoda rehabiitation intervention that incudes exercises to decrease impairments (these can be physica and menta), education to provide information and behaviour change regimes, use of adaptive equipment and support from other peope, both forma or informa sources, woud be more effective at improving outdoor mobiity. This study aimed to fi this evidence gap. Justification for the current study An outdoor mobiity intervention for peope with stroke was deveoped using quaitative interview findings, pubished iterature, correspondence with the Department of Transport and expert opinion. 23 This intervention was evauated in a singe-centre randomised controed study. 24 The study found cear benefits in outdoor mobiity for peope who received the intervention. The patients who received the intervention took twice as many journeys, many of these waking, as those in the contro group, 4 and 10 months after recruitment into the study, and were significanty more ikey to be satisfied with their outdoor mobiity activity. Athough the study was based on a rea cinica need, recruited to target, had exceent foow-up and was pubished in a peer-reviewed journa, it had some imitations. The intervention was deivered by a singe therapist, who was a stroke speciaist, with experience in the deivery of community-based treatment. It was deivered in one UK city andnoheatheconomicevauationwascompeted. Athough this study has aready infuenced practice, aone it is insufficient to inform evidence-based guideines. Thus, the next step was to conduct a muticentre randomised controed tria (RCT), using the essons earned from the singe-centre study. We competed a muticentre RCT to try and answer the foowing questions. 1. Whether: i. the resuts of the piot study were generaisabe to other therapists and other sites. ii. the intervention coud be impemented across a heath system. 2 NIHR Journas Library

31 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 iii. the intervention improves overa heath and, if so, whether such an approach is conventionay cost-effective. Structure of the Heath Technoogy Assessment monograph The monograph has been separated into methods and resuts for the randomised controed study, incuding deivery of the intervention and cinica effectiveness, foowed by methods and resuts for both the economic evauation and the quaitative substudy. Finay, a resuts are discussed, foowed by overa concusions. Appendices incude the summary of the intervention manua and the statistica anaysis pan. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 3

32

33 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 2 Randomised controed study methods This chapter describes the methods for the deivery and assessment of the intervention and coection of data to assess cinica effectiveness. Summary of study design The Getting out of the House Study was a muticentre, parae-group RCT comparing cinica effectiveness and cost-effectiveness for two groups, an intervention group and a contro group. The methodoogy used was, as far as possibe, a repication of that used in the singe-centre study. 24 Figure 1 summarises the study design. Patients aged 18 years, who had experienced a stroke > 6 weeks previousy were considered initiay eigibe for the study. The majority of patients were contacted by a patient invitation etter sent from either their genera practitioner or from the oca community rehabiitation or oca hospita stroke service register. If interested, patients repied to their oca research team, who contacted them and screened them by teephone ca to further determine eigibiity. Fina confirmation of eigibiity and subsequent consenting was carried out by the research assistant (RA) at the patient s home. Foowing consent, a baseine assessment was competed and a participants received the contro. They were provided with trave diaries in the form of 12 months of trave caendar to record the number of journeys that they made and any fas that they had. A participants were then randomised by the oca therapist, using the web-based randomisation system that was controed by the Nottingham Cinica Trias Unit (NCTU), to either targeted outdoor mobiity therapy (the intervention) or no further therapy contact (contro). Those randomised to the intervention received up to 12 visits over a maximum of 4 months (from baseine). No further therapy contact was received after 4 months, aowing 2 months before the primary and secondary outcome data were coected. Outcomes were assessed at 6 and 12 months after recruitment by posta questionnaire and by monthy trave diaries. Loca RAs binded to treatment aocation were avaiabe to assist the participant with competion of questionnaires if needed. RAs recorded incidents where they were unbinded. A subset of participants at a singe site were asked to participate in a nested quaitative study investigating confidence issues surrounding stroke. Primary objective To test the cinica effectiveness and cost-effectiveness of treating peope who have had a stroke with an outdoor mobiity rehabiitation intervention in addition to routine care compared with routine care aone. Secondary objectives The secondary objectives were to measure whether the intervention was associated with: improved: mobiity in and outside the house patient we-being participation in everyday activities carer we-being heath-reated quaity of ife. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 5

34 RANDOMISED CONTROLLED STUDY METHODS Stroke registers GP invites Outpatient cinic visit Pre-consent screening (teephone) by RA Consent by RA Coect baseine data, deiver contro Randomisation by therapist via NCTU web-based system Intervention deivered within 4 months of baseine Contro 6-month foow-up with primary end point via: (a) Posta questionnaire direct to participant (b) Loca RA-assisted competion Subset approached for quaitative confidence interviews 12-month foow-up via: (a) Posta questionnaire direct to participant (b) Loca RA-assisted competion FIGURE 1 Study design fow chart. GP, genera practice. Primary outcome measure The primary outcome measure was the Socia Function domain score of the Short Form questionnaire-36 items, version 2 (SF-36v2) 25 measured at 6 months, as recorded in their questionnaire response. This is one of the eight domains of the SF-36v2 and was scored and transformed on to a scae of 0 (worst possibe heath state) to 100 (best possibe heath state). Secondary outcome measures The secondary outcome measures were: 1. functiona activity in participants measured by the Nottingham Extended Activities of Daiy Living (NEADL) scae 26,27 2. mobiity reported by participants in the competion of the Rivermead Mobiity Index (RMI) the number of journeys as recorded by participants in the trave caendars 4. satisfaction with outdoor mobiity (SWOM) as reported by participants in the competion of a yes/no question: Do you get out of the house as much as you woud ike? 5. psychoogica we-being as reported by participants in the competion of Genera Heath Questionnaire 12 (GHQ-12) NIHR Journas Library

35 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO carer psychoogica we-being as reported by participants in the competion of GHQ heath-reated quaity of ife as reported by participants in the competion of European Quaity of Life-5 Dimensions (EQ-5D) 30 and Short Form questionnaire-6 Dimensions (SF-6D) 25 (based on a subset of questions from the SF-36 v2 31 ) 8. the resource use of heath and socia care, carer input and provision of equipment as reported by participant in the competion of the service use questionnaire and data coected reating to intervention visits 9. participant mortaity (death data) as coected from NHS Information Centre (IC)/NHS Centra Register. Recruitment process There were three methods of patient identification: searching genera practice (GP) databases, searching stroke registers in secondary care hospitas, community hospitas or primary care community teams and approaching patients attending post-stroke outpatient cinics as they were discharged from rehabiitation. If the search team aso had access to oca stroke service rehabiitation records they cross referenced for any active stroke-reated rehabiitation that may be ongoing. A potentia participants received a version of the participant information sheet (PIS) prior to the baseine visit and had sufficient time to consider the information. In addition, if required, an aphasic PIS and a summary PIS were avaiabe for participants and their carers. Other recruitment strategies incuded pacing adverts within oca newspapers, oca trust pubications and reevant websites (e.g. the Stroke Association), estabishing a study website, various artices within oca trust pubications and oca press, visiting oca stroke groups and widespread distribution of the study poster to reevant areas (e.g. GPs, rehabiitation teams, stroke groups, etc.). Sampe size Our sampe size cacuations were based on the primary outcome measure, the Socia Function domain of the SF-36v2 at 6-month foow-up. Athough this was not used in the piot study, a recent study had suggested a minimay cinicay important difference (MCID) for the Socia Function domain was 12.5 points. 32 The estimate of the standard deviation (SD) was taken from Britte s study 33 and this cosey matched the standard deviation obtained by assuming that the distribution of the nine possibe integer-vaued scores foowed a roughy right-anged trianguar distribution (i.e. that was positivey skewed). This foowed the genera approach suggested by Deming 34 for schematic distributions. Assuming a power of 90% and a two-sided significance eve of 5%, we estimated that to detect a difference in mean SF-36 scores of 12.5 points, assuming a common SD of 28.2 points, 33 a sampe size of 135 patients per group was required. This cacuation assumed an attrition rate of 20% over a 6-month period. Custering by deivery of treatment was aowed for by generating simuated data to mimic the mutieve structure of the proposed study, with site-specific variation between therapists. The ikey existence of sharing of therapists by patients was addressed by a mutipe membership mode, and the variance of the treatment effect so found was compared with that of a regression mode ignoring custering (both modes treating centre as a fixed effect). The data sets and anayses were performed in Stata v10 (StataCorp LP, Coege Station, TX, USA). The variance infation observed was insensitive to variation in the simuated between-site variance and the average between-therapist variance, and a rounded vaue of 2.5 was seected as the mutipier for a sampe size based on a naive anaysis. After the aowance for custering by deivery of treatment, a sampe size of 338 participants per arm was required, with an overa sampe size of 676, aowing for 20% ost to foow-up. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 7

36 RANDOMISED CONTROLLED STUDY METHODS Revised sampe size Owing to an increase in the expected number of sites, the sampe size was revised. As more information had become avaiabe after the origina sampe size cacuation was performed we deveoped an aternative method of cacuating a sampe size that was easier to repicate. This anaytic approach was based on the same assumptions of the origina sampe size cacuation and assumed that within each site there woud potentiay be four therapists deivering the intervention in the intervention group and one therapist deivering the contro in both groups. In these cacuations we assumed a therapist effect intracass correation coefficient (ICC) of 0.02 (Professor Stephen Waters, University of Sheffied, 30 October 2008, persona communication) and a site effect ICC of 0.04 [obtained from Aberdeen University s database of ICCs ( For comparison, based on seven sites this method gave a sampe size of 746, aowing for 20% ost to foow-up. This method was adopted to provide a revised sampe size. As the number of sites had increase from the assumed 7 to 14 (at that time) we conducted further power cacuations assuming a ower sampe size and investigated the effect of varying the number of sites. This was based on detecting a difference in means of 12.5 (as in the origina sampe size cacuation) and assuming a tota sampe size of 440 with the number of sites ranging from 7 (as originay panned) to 14. These cacuations were based on the same assumptions made in the anaytic sampe size above (i.e. SD = 28.2 sites, two-sided significance eve of 5%, ICC of 0.02 for therapist effect, and ICC of 0.04 for centre effect). These indicated that a sampe size of 440 woud have a power of 86% (assuming no attrition) and a power of 82% (assuming ost to foow-up rate of 20%) to detect a difference in means of 12.5 if there were seven sites and if there were 12 or more sites the corresponding power woud be at east 90%. There were a tota of 15 sites in this study so this met the atter cacuations. Sampe size cacuations can be performed ony if a number of assumptions are made and athough our cacuations were based on the best estimates we coud obtain at the time that there was some uncertainty about the true vaues of the ICCs used to adjust for custering. Therefore, we decided to be cautious and aimed to recruit a sampe size of 506 participants ( %) to ensure that we had recruited enough participants. This was agreed with the Tria Steering Committee (TSC). Setting and ocations The study was coordinated by the NCTU and invoved 15 sites from across Engand, Scotand and Waes. Initiay seven sites were panned but during the course of recruitment an additiona eight were added. The sites were a mixture of settings, spit in broad terms as foows. Primary care community stroke rehabiitation service 1. NHS Nottingham City. 2. NHS Nottinghamshire County. 3. Gateshead Primary Care Trust (PCT). 4. NHS North Somerset. 5. Woverhampton PCT. 6. Kent Community Heath Trust. 7. Tower Hamets PCT. 8. Norfok Community Heath and Care NHS Trust/NHS Norfok. 9. NHS Bristo. 8 NIHR Journas Library

37 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Secondary care community stroke rehabiitation service 1. NHS Lanarkshire. 2. NHS Grampian. 3. Cardiff and Vae University Heath Board. 4. Cwm Taf Heath Board. Secondary care stroke service 1. United Linconshire Hospitas NHS Trust. 2. Southend Hospita NHS Foundation Trust. Eigibiity criteria Patients were eigibe for the study if they provided written informed consent and if they: i. were aged 18 years ii. had experienced a stroke at east 6 weeks previousy, and iii. wished to get out of the house more often. Patients were not eigibe for the study if they were: i. not abe to compy with the requirements of the protoco and therapy programme in the opinion of the assessor ii. sti in post-stroke intermediate care or active rehabiitation, or iii. previousy enroed in this study. Patients who had suffered a transient ischaemic attack were not incuded. Screening and baseine assessment The majority of preconsent screening was undertaken over the teephone by the RA in response to a patient repy to a patient invitation, and invoved a genera discussion of the study background and participation requirements. In addition, key eigibiity questions were asked in reation to wishing to get out of the house more often and estabishing potentia outdoor mobiity goas reated to the intervention manua. Foowing this process, if the potentia participant remained eigibe then a baseine visit woud be arranged. The baseine visit was conducted in the potentia participant s home. The visit consisted of four eements: first, confirming eigibiity against incusion and excusion criteria; second, consenting (see Consent, beow); third, baseine data coection; and, fourth, deivery of the contro (see Contro, beow). Ony imited numbers of data were coected, incuding basic demographics, time since stroke and contact detais. Consent Participant information sheets were avaiabe in severa different formats, incuding fu ength, abbreviated eafet stye, summary and aphasic. In practice, a combination of a four of these were often used, as appropriate, for each individua participant. Ony peope who had the menta capacity to consent to Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 9

38 RANDOMISED CONTROLLED STUDY METHODS participation, defined against the criteria in the Engish Menta Capacity Act, 35 were eigibe to take part. If there were communication probems then the participant s carer or reative was invoved in the process. In addition, owing to potentia physica imitations with competion of the consent forms, it was acceptabe for the participant to mark the consent form, accompanied by a witness signature to compete the consent process. If the physica chaenges were more severe then a witness (usuay a member of the participant s famiy, carer or neighbour) coud compete the consent form on the participant s behaf. During the baseine visit, carer information sheets and questionnaires were provided to a carers. They were aowed to compete them immediatey or ater, and the process of competing and returning the questionnaire was considered to be impied consent. Contro information A participants in the study received the contro information during the baseine visit by the RA. The contro information was provision of verba and written oca mobiity and transport information. This information was site specific. Each site coated a pack of oca information and eafets as guided by the research team. They were asked to incude information about bus times, oca community transport, taxi services, wheechair services, disabed persons car badges, wheechair borrowing schemes and mobiity equipment. These individuaised packs were given to the participant after their content had been discussed. Randomisation Foowing consent and competion of the baseine assessments and deivery of the contro information, the participant s detais were passed to the oca therapist. The therapist accessed the remote, secure web-based eectronic case report form (ecrf) and randomisation system deveoped and maintained by the NCTU and entered basic demographic detais. Once these detais had been entered irrevocaby into the system, the group to which the participant was randomy aocated was provided to the therapist. Randomisation was created using Stata/SE version 9 statistica software with a 1 : 1 aocation to intervention group or contro. Participants and therapists were aware of the intervention aocation, whereas the RA responsibe for assisting with coection of the outcome measures was kept binded to the aocation throughout the study. Randomisation was based on a computer-generated pseudorandom code using random permuted baanced bocks of randomy varying size, set up and maintained by the NCTU in accordance with its standard operating procedure and hed on a secure server. The randomisation was stratified by age (< 65 years and 65 years) and site. Sixty-five years of age was chosen as a cut-off point because Mobiity Aowance a monetary aowance was no onger avaiabe for peope aged > 65 years and over. Access to the sequence was confined to the NCTU Informartion Technoogy (IT) Manager. The sequence of treatment aocations was conceaed unti a participants had been randomy assigned to a treatment group and recruitment, data coection and a other study-reated assessments were compete. If the participant was randomised to the intervention group then the therapist contacted the participant to arrange an initia goa-setting visit. If the participant was randomised to the contro group the therapist woud either, based on site staff preference, not contact the participant or contact the participant to expain their aocation. It was not recorded which approach was used for which participant. 10 NIHR Journas Library

39 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Intervention The deveopment of the intervention foowed the guideines. 36 Quaitative interviews 23 were used to identify barriers and need theories. A group of senior and speciaist therapists met to discuss the interviews and deveoped an intervention based on the resuts, their cinica skis and training. The intervention training manua that was used to train therapists in each site was deveoped using a systematic approach. Initiay, a iterature review was performed using the search strategy for the foowing terms rehabiitation, stroke, occupationa therapy and outdoor mobiity, with the resuts reviewed by a pane of expert stroke cinicians from which a draft training manua was deveoped. This was then reviewed by a focus group of therapists (n = 15) who had either deivered the intervention in the piot study or were cinica experts in stroke. The intervention training manua used for this study contained: standardised assessments case vignettes of treatment pans goa panning and activity anaysis guideines protoco for a first outing waking, using the bus and using a taxi checkist of benefits and barriers to going outside motivationa and confidence-buiding strategies exampes of skis needed to catch a bus or train exampes of skis needed to be abe to operate an outdoor mobiity scooter. The aim of the intervention was to increase outdoor mobiity participation by aeviating physica difficuties, deveoping skis to maximise the individua s potentia and overcoming psychoogica barriers. The main component of the intervention was for participants to repeatedy practise outdoor mobiity. This incuded buses, taxis, waking, vountary drivers and mobiity scooters unti they fet confident to go aone or with a companion. The number of intervention sessions depended entirey on the participant. If they fet they did not require any further intervention, for whatever reason, then the intervention stopped. If they fet they required additiona intervention for whatever reason, they coud continue the intervention up to a maximum of 12 visits. We recorded the duration of each intervention visit, incuding the time taken to deiver the intervention, and trave to and from the participant s home. A description of the intervention has been pubished 37 and a training manua deveoped. A more detaied summary of the training manua can be found in Appendix 2. Training therapists to provide the intervention Therapists in each site were trained by Pip Logan [chief investigator (CI) and soe therapist in the piot study] to deiver the intervention using the intervention training manua and a standard presentation over the course of 2 hours. Foowing this face-to-face training, the majority of therapists attended an update session each year. On an ongoing basis they were abe to e-mai other therapists on the study to discuss treatments and were abe to contact the study CI with any queries reating to the intervention. Each therapist was registered on the ecrf database and when the record of a particuar intervention visit was added to the ecrf then a therapist, or therapists, had to be assigned to that particuar visit, therefore providing an accurate record of the deivery of the intervention, which coud be correated against the training records. Data coection There were four methods of data coection: ecrf, questionnaire bookets, trave diaries and intervention records. The NCTU performed quaity contro checks on a aspects of data coected and entered onto the Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 11

40 RANDOMISED CONTROLLED STUDY METHODS ecrf to ensure accuracy and competeness. Specificay, a 100% check of primary end points and a 10% check of a other data were competed. Eectronic case report form Ony the therapist [and principa investigator (PI)] at each site had access to the ecrf. The baseine demographic detais, incuding the NHS number, were recorded and suppied to the NHS IC/NHS Centra Register to aow mortaity checks prior to foow-up. In addition, data for duration of visit, the occurrence of any adverse events and end of intervention detais was aso coected and directy entered on the ecrf, often transcribed from paper records (either study worksheets or the intervention records). Questionnaire bookets Baseine participant questionnaires were competed by participants, with assistance from the RA if needed, during the baseine screening assessment. Aternativey, the RA eft the questionnaire for the participant to compete ater and post direct to the NCTU. We asked participants to record at baseine and both the 6- and 12-month assessments, whether they competed the questionnaire themseves or required assistance. This was to indicate who physicay competed the questionnaire and it was assumed that the participant answered the questionnaires regardess of who fied in the form. If the RA assisted with competion of the baseine questionnaire, they informed the NCTU that the participant woud aso require assistance at 6 and 12 months. The RA sent the baseine questionnaire bookets to the NCTU for entry onto the database. Baseine carer questionnaires were either competed during the baseine screening assessment, if the carer was avaiabe, or eft with the participant for competion ater. Six- and 12-month participant and carer questionnaire data were coected either by post or with assistance of a bind-to-aocation RA. Foowing baseine assessment, each participant was assigned to the posta approach uness otherwise indicated by the RA. The NCTU tracked a questionnaires with the option to send a reminder etter or to teephone to assess the situation. If difficuties were encountered at any time during the study then participants were switched from the posta approach to the RA approach. These methods were found to be successfu in the singe-centre piot study, with a 90% posta return rate of active participants (i.e. those who had not withdrawn, died or been ost to foow-up) with the further 10% coected by an independent assessor, the equivaent of the RA in this study. Trave diary Trave diaries were competed by a participants on a daiy basis. This was in the form of a caendar with the participants entering the number of journeys made for each day. In addition, in order for the study team to assess safety, participants indicated whether they had experienced a fa on each day. The participant was provided with a 12- or 13-month caendar, depending on the date of the baseine visit. Training in the competion and return of the trave diaries was provided by the RA at baseine. A prepaid enveope was sent to participants at the end of each month for them to return the diaries. No reminder etters were sent. Data from trave diaries were entered on to the database on a month-by-month basis but ony if a journey was recorded (i.e. 1 journey for any given day). Months that contained ony 0 were aso recorded. If the month was bank then no data were entered for that particuar month. Intervention records As we as the intervention duration being recorded in the ecrf, therapists aso competed a paper intervention record form detaiing goas set, visit-by-visit cinica detais pus a breakdown of therapy activities. These were coected at the end of the study and coated into a separate database. Therapists were trained to use the form as part of the intervention training. The type and duration of intervention within any given visit were recorded under the foowing headings: goa setting, mobiity, confidence, adaptive equipment, information, other rehabiitation and referra on to other agencies. For each 12 NIHR Journas Library

41 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 intervention visit the therapist recorded how many minutes they spent on each activity, they isted the agreed outdoor mobiity goas and kept running records of each treatment session as they woud in routine cinica care. The intervention records were kept in the NHS cinica setting unti the end of the study when they were copied and coated by the centra coordinating centre. Treatment fideity To assess the fideity of the intervention therapy provided, 20% of cinica records were monitored (10% of a participants). A RA who was independent of the site visited the site and checked the treatments being deivered against a predefined checkist by accompanying the therapist on a number of treatment visits and by monitoring the intervention records. The ast question on the checkist (a yes/no question as to whether the assessor beieved that treatment provided met the required standard) was used to indicate whether the intervention had been deivered as specified and, therefore, impied fideity of treatment. Adverse events/safety evauation The main risk identified from this study was the potentia to increase the chance of a participant having a fa, as a resut of increased mobiity. For adverse event purposes we adopted the foowing definition: A recordabe adverse event is any fa that occurs during or foowing the first outdoor mobiity intervention and for the remaining period of intervention in the study that specificay requires the assistance of a heath-care professiona. If these criteria were met then an adverse event form was competed. We adopted the foowing approach to capture this information the therapist woud ask, during an intervention visit, the foowing questions: Have you had any fas since my ast visit? If yes, How many? How many required heath-care professiona hep? Therefore, adverse events were ony recorded from the intervention group, as the contro group did not receive any visits post baseine. No other adverse events were coected. The occurrence of a serious adverse event as a resut of participation within this study was not expected and no serious adverse event data were coected. In order to capture information from both groups for comparison and a further measure of safety, we adopted a second approach. Participants indicated a fa on the trave diary by marking or circing the preprinted f, denoting that 1 fa had occurred for that participant on that day. These woud not be cassed or recorded as adverse events. Aso, a comparison was made of any untoward changes in the we-being score of a participants by comparing the baseine GHQ-12 score with the 6- and 12-month score. In the event of a pregnancy occurring in a study participant then they were aowed to continue in the study regardess of which arm they had been randomised to. Conceament of aocation during outcome assessments Participants and therapists were aware of treatment aocation. However, the treatment aocation was conceaed from the RA so that they were abe to assist the participant with competion of the 6- and 12-month questionnaires, if necessary. In order to measure conceament of aocation we asked the RAs to compete a binding assessment for each post-randomisation visit, indicating whether they were unbinded prior to or during the visit. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 13

42 RANDOMISED CONTROLLED STUDY METHODS Participant retention and withdrawas To try and reduce the bias in participant retention caused by peope wishing to withdraw before the primary end point because they were not aocated to the intervention group, we aowed participants to stop competing the trave diaries and asked them whether they woud be happy to remain in the study and compete the 6-month questionnaire. No specific withdrawa criteria were defined for this study. If a participant eft the study prematurey (i.e. prior to competion of the protoco), the primary reason for discontinuation was determined and recorded if at a possibe. Withdrawn participants were not repaced. Participants were made aware (by the information sheet and consent form) that shoud they withdraw, the data coected prior to the withdrawa date woud not be erased and woud sti be used in the fina anaysis. For this study, and in reference to the CONSORT (Consoidated Standards of Reporting Trias) diagram (see Figure 3), the foowing definitions were appied in reation to the participants and the data: Withdrawn consent The participant formay withdrew from the study and no further data were coected from that point. This was measured in reation to competion of 6- or 12-month foow-up and not the time of withdrawa, for exampe, if a participant withdrew after 7 months and did not compete their 6-month foow-up then they woud be recorded as withdrawn prior to 6-month foow-up. Lost to foow-up The study coordinating centre and/or oca site staff were unabe to contact the participant. This was recorded in reation to competion of 6- or 12-month foow-up as above. No foow-up data The study coordinating centre and/or oca site staff were in contact with the participant but foow-up data were not received; however, the participant was not formay withdrawn from the study. These participants at 6-month foow-up were cassed as no foow-up data on the CONSORT diagram and continued to participate in the study. These participants at 12 months foow-up were aso cassed as no foow-up data on the CONSORT diagram. Statistica methods Fu descriptions of the statistica methods of anaysis are detaied in the statistica anaysis pan (see Appendix 1). Intervention and goas descriptive anaysis A description of the components of the intervention was undertaken with the type of intervention therapy received by participants (e.g. mobiity, goa setting, confidence, etc.) summarised in terms of number and percentage of participants who received the different types of therapy sessions and the mean, SD, median, IQR, and the minimum and maximum of the number of sessions for each type of therapy session. The goas set by each participant during the intervention sessions were recorded verbatim and this information was then sorted into categories of the goa types. These data were then anaysed through cacuating the number and percentage of participants who set each type of goa. Comparison between treatment arms A anayses were undertaken on an intention-to-treat basis, in that participants were anaysed according to the group to which they were randomised, regardess of whether they received the intervention or not. Anayses were conducted on avaiabe case data. A sensitivity anaysis, using mutipe imputation (MI) to repace missing vaues, was performed for a outcome measures except trave journeys (see Sensitivity anaysis adjusting for missing data section beow for further information). 14 NIHR Journas Library

43 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Basic data exporations using summary and graphica statistics (i.e. box pots, etc.) were conducted to check for outiers. If, after additiona investigation, there was no evidence that these were errors then the anayses were repeated after Winsorising the data to assess the robustness of the resuts. Winsorising invoves repacing extreme vaues (outiers) with a specified percentie of the data. 38 In this case, we repaced data vaues beow the 5th percentie with the 5th percentie vaue and to repace data vaues above the 95th percentie with the 95th percentie vaue. Descriptive anayses Continuous data that were approximatey normay distributed were summarised in terms of the mean, SD and number of observations. Skewed data were presented in terms of the median, ower and upper quarties, minimum, maximum, and number of observations. Categorica data were summarised in terms of frequency counts and percentages. Anaysing primary and secondary outcomes The mean socia function score of the SF-36v2 was compared between treatment groups using a mutipe membership form of mixed-effects mutipe regression anaysis, adjusting for site (as a random effect), age and baseine socia function score as covariates, and therapist as a mutipe membership random effect. A three-eve hierarchica regression mode was used with site at eve 3, mutipe membership effect of therapists at eve 2, and participants at eve 1. Regression coefficients and 95% credibe intervas were presented. The robustness of these findings was assessed by repeating the anaysis and incuding baseine variabes that are ikey to be associated with the outcome variabe (namey gender and residentia status) as covariates in the mode. Satisfaction with outdoor mobiity was compared between treatment groups using a mutipe membership form of a mixed-effects ogistic regression mode adjusting for site (as a random effect), baseine answer to the question and age as covariates, and therapist as a mutipe membership random effect. Odds ratios and 95% credibe intervas were presented. Trave journeys were compared between treatment groups using a mutipe membership form of a mixed-effects Poisson regression mode, adjusting for site (as a random effect) and age as covariates, and therapist as a mutipe membership random effect. Rate ratios and 95% credibe intervas were presented. Trave journeys were presented as number of journeys per day. The anaysis assumed that for a particuar month for which journeys had been recorded then a other days were imputed with 0. A months that were returned bank were not incuded in the anaysis. When no trave diary was returned then that month was not incuded in the anaysis. A remaining secondary outcome variabes were anaysed using the same methods as for the primary outcome measure. We acknowedge the potentia for type 1 errors associated with significance testing for mutipe end points and, therefore, we consider the anayses of the secondary outcome measures to be party exporatory in nature, and party confirmatory of the findings for the primary outcome measure. Sensitivity anaysis adjusting for missing data We assessed the effect of missing data on the overa concusions of the study using MI. We used the Stata ice command to generate 10 imputed data sets, and estimated the intervention effect in each imputed data set. We then used Rubin s Rues to combine these estimates into a pooed estimate of the intervention effect. 39 This process was conducted for a outcomes (except the trave journeys) for 6- and 12-month data. We considered the use of this process for the trave journeys to be inappropriate because, in addition to imputing the tota number of journeys made, we woud aso have had to impute the number of days on which journeys were taken. Furthermore, there was no prior information avaiabe at baseine on the number of journeys made or the numbers of days on which the participants had made journeys that coud be used to predict the data for 6 and 12 months. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 15

44 RANDOMISED CONTROLLED STUDY METHODS Mutipe membership random effect In many studies, outcome measures are infuenced by an intervention, such as a drug or a singe person deivering an intervention, and, in these studies, assessing the effects of the drug or person on the outcome is cear. In our study, participants in the contro group were treated by ony one therapist but in the intervention group participants may have been treated by mutipe therapists and some therapists may have deivered more therapy sessions to a specific individua than others. This means that the outcome obtained from a participant receiving treatment from different therapists may have been infuenced by more than one person and, as some therapists may have deivered more therapy sessions to a participant than others (and the therapists may differ in their effectiveness), it is important to take account of this in the anaysis. We used mutipe membership random-effects modes to dea with this situation and adjusted for the effect of the therapist by pacing weights on the therapists deivering the intervention. We paced the greatest weight on the therapist who deivered the most therapy sessions for a particuar participant. If two therapists deivered the same number of therapy sessions to a participant, equa weight was appied to each therapist, whereas if one therapist deivered one-third of the therapy sessions and another therapist deivered two-thirds of the therapy sessions (i.e. twice as many), the therapist who provided the most sessions was given the greatest weight (in this case, twice the weight of the other therapist). For participants treated by ony one therapist, a weight was paced on that therapist, as ony that therapist infuenced the outcome. The weight appied to each therapist treating an individua therefore refected the potentia amount of infuence that they may have had on the outcome with ow weights indicating ess infuence on the outcome than higher weights. Presentation of primary and secondary resuts A main 6- and 12-month primary and secondary outcome data were presented descriptivey by treatment group using means and SDs for continuous data or frequency counts and percentages for categorica data (unadjusted resuts), and then the resuts of the mutivariabe anaysis adjusting for therapy effect and site effect were presented as differences in means for continuous data, odds ratios or rate ratios for binary and count data, respectivey (adjusted resuts), with 95% credibe intervas as a measure of significance. In addition, the ICCs indicating the size of the therapy effect and site effect were presented, apart from trave journey, for which these effects were not cacuabe. The interpretation of the resuts of the mutivariabe anaysis of the 6- and 12-month primary and secondary outcome data are as foows: 1. Difference in means Where a score of 0 indicates no difference and hence if the 95% credibe interva crossed the vaue 0 then the resut (difference in means between the two treatment groups) was not statisticay significant. This approach was used for the primary outcome, NEADL, RMI and GHQ-12 outcome measures. 2. Odds ratio Where a score of 1 indicates no difference and hence if the 95% credibe interva crossed the vaue 1 then the resut (the odds of getting out of the house as often as they woud ike in the intervention group reative to the contro group) was not statisticay significant. This approach was used ony for the outcome SWOM. 3. Rate ratio Where a score of 1 indicates no difference and hence if the 95% credibe interva crossed the vaue 1 then the resut was not considered to be statisticay significant. This parameter was used for trave journeys. Exporatory/other anayses The foowing anayses were prespecified in the anaysis pan: 1. NEADL by category Descriptive summaries (by treatment group) for the NEADL scae categories Mobiity, Kitchen, Domestic and Leisure. 2. Number of sessions by site Descriptive anaysis of number of intervention visits by site compared with perceived successfu competion of the intervention. 3. Number of trave journeys by intervention session Descriptive anaysis comparing the number of trave journeys within the intervention group by participants who had received fewer than six sessions or 16 NIHR Journas Library

45 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 six or more sessions. The piot study found that a median of six sessions were effective, hence the pre-existing threshod of fewer than six or six or more sessions was used in this anaysis. 4. Fas by age category and treatment aocation The tota number (%) of fa-days was described by treatment arm and age category (< 60 years of age and 60 years). 5. Satisfaction with outdoor mobiity changes over time To expore the effect of the contro information a descriptive comparison in the change of the SWOM scores from baseine to 6 months for each group. 6. Six-month foow-up by approach A descriptive comparison of coection of 6-month foow-up data from posta or RA approach. Compiance in reference to CONSORT Compiance in this study was effectivey a measure of competeness of foow-up data. Baseine questionnaires were required to be competed before randomisation so that baseine CONSORT-presented numbers randomised. The CONSORT diagram presented compiance for this study as the number of returned questionnaires for the 6- and 12-month foow-ups, regardess of competeness of the questionnaire bookets. For individua anayses of outcome measures, as the anaysis reies on a comparison between foow-up data and the baseine data, the n presented in the resuts section was a measure of the number of participants who competed the outcome measure at both baseine and foow-up. Data from participants who answered some, but not a, of the questions within a questionnaire based outcome measure were sti incuded in the anaysis; however, if the whoe of the questionnaire was missing then their data were not used. Study management and oversight A number of committees were assembed to ensure the management and conduct of the study, and to uphod the safety and we-being of participants. Tria Management Group The Tria Management Group (TMG) oversaw the operationa aspects of the study, which incuded the processes and procedures used and the day-to-day activities invoved in study conduct. Tria Steering and Data Committee The Tria Steering and Data Committee (TSDC) had the overa responsibiity for ensuring a scientificay sound study design, a we-executed study, and accurate reporting of the study resuts. The data monitoring part of this committee evauated the outcome and safety data in the context of the overa study and the currenty existing information about the study. They considered the appropriate time frame for reviewing the data during the course of the study. The TSDC had access to grouped study data. However, as part of the process of cacuating a revised sampe size, an ad hoc independent Data Monitoring Committee (DMC) was convened, which reviewed the suitabiity and egitimacy of the proposed revision. Amendments to the study Eigibiity To simpify the screening process we removed Abe to compy with the requirements of the protoco as an incusion criterion, and Significant cognitive impairment which wi impede abiity to compete the assessments and Diagnosis ikey to interfere with rehabiitation or outcome assessments e.g. termina iness as excusion criteria. In their pace we added a singe excusion Not abe to compy with the requirements of the protoco and therapy programme, in the opinion of the assessor/gp/investigator. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 17

46 RANDOMISED CONTROLLED STUDY METHODS These amendments were made prior to the start of recruitment. We amended At east 6 weeks but no onger than 5 years since stroke to At east 6 weeks since stroke as an incusion, as there was no vaid reason for the 5-year imit and participant invitations asked participants whether the mobiity issues were as a resut of their stroke, so time frame was not reevant. Hence it was more incusive to potentia participants. This amendment was made after 4 months of recruitment. Sampe size Foowing consutation with the funders and the TSC, it was agreed that we woud cacuate a revised sampe size, based on an increase in the number of sites. This revision is described in the Revised sampe size section earier in the chapter. Independent statisticians from outside the study checked the vaidity of a cacuations and assumptions. These amendments were made after 16 months of recruitment. Summary of changes to recruitment materia After initiay using a standard PIS, in fu and summary form, we introduced a suppementary PIS eafet, in addition to rewording the PIS, consent form and patient invites. These were amended in order to make the aims of the study more transparent and patient friendy. Another significant change was the introduction of an aphasia-specific PIS consisting of images and key words. Though initiay designed for aphasic patients it was used extensivey, in conjunction with more standard formats, for other stroke patients. These amendments were made after 4 months of recruitment. Quaitative study A study investigating confidence after stroke was incorporated into the existing study protoco with a view to interviewing existing study participants about issues reating to confidence and how it affected their recovery from stroke. This part of the study was ony in Nottingham for 10 participants. The quaitative study is described in more detai in Chapter 5. This amendment was made after 11 months of recruitment. Recruitment timeines Initia recruitment timeines were 12 months; however, fina recruitment timeines were extended to 22 months. This was managed through a combination of efficiencies within the overa project timeines and an extension to the overa project from 36 months to 40 months. 18 NIHR Journas Library

47 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 3 Resuts: randomised controed study This chapter reports recruitment and retention of participants, and coection and anaysis of study data specificay reating to the deivery and fideity of the intervention, primary outcome measured at 6 months, and secondary outcomes measured at 6 and 12 months. Study recruitment and foow-up The study recruited 568 out of 506 (112.3%) of the required sampe size from November 2009 to August Participants were recruited from 15 sites across Engand, Scotand and Waes, with an average recruitment of 38 participants (range from 5 to 99) per site, with nine sites recruiting > 30 participants. The participant invovement in the tria ended in August 2012, foowing competion of recruitment and foow-up. Figure 2 shows that, after an initia ag phase, steady recruitment was achieved with no noticeabe seasona variation. Tabe 1 provides a breakdown of screening figures coected from a sites, in terms of number of invitations sent, number of yes repies, number of home screening visits, number randomised and the proportion of recruits per 100 invites. The different approaches were used to varying extents: GP-based approaches being responsibe for recruiting 158 out of 568 (27.8%) of randomised participants at a rate of 2.5 per 100 invites, and stroke register-based approaches recruiting 410 out of 568 (72.2%) of randomised participants at a rate of 8.5 per 100 invites. The stroke register data incudes participants recruited at outpatient cinics. Foow-up (CONSORT) Figure 3 shows the CONSORT diagram for the study. The identification of participants was from 11,126 invites, either by etters of invitations or face-to-face approach. The exact breakdowns were not recorded; however, the majority were by etter of invitation. We had yes repies from 1448 out of 11,126 (13.0%) approaches to participate in the study with 9678 out of 11,126 (87%) either not responding or providing negative repies. We did not record the number of no repies. After initia pre-consent screening, usuay by teephone, a further 596 out of 11,126 (5.4%) were considered ineigibe. The main reason for ineigibiity Number of participants FIGURE 2 Recruitment rate October 2009 December 2009 February 2010 Apri 2010 June 2010 August 2010 October 2010 December 2010 Time period February 2011 Apri 2011 June 2011 August 2011 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 19

48 RESULTS: RANDOMISED CONTROLLED STUDY TABLE 1 Breakdown of participant identification, screening and recruitment numbers from two aternative contact approaches Approach Invites Yes repies (% of invites) Screening visits (% of yes repies) Randomised (% of yes repies) Recruits/ 100 contacts GP database (10.7) 261 (38.4) 158 (23.3) 2.5 Stroke register (16.0) 591 (76.9) 410 (53.3) 8.5 Overa 11, (13.0) 852 (58.8) 568 (39.2) 5.1 was that the potentia participants fet they got out of the house as much as they wished aready. Therefore, 852 out of 11,126 (7.7%) had a further screening visit in their own home from the RA. Overa, 568 out of 852 (66.7%) were found to be eigibe to take part. Of the 11,126 etters sent, 568 peope were incuded (5.1%). Of the 284 out of 11,126 (2.6%) who did not enter the study, ony 1 out of 11,126 (0.01%) expicity refused to participate (owing to their desire to be in the intervention group ony). The baseine questionnaire was competed by 567 out of 568 (99.8%) participants, with the one outstanding being eft with the participant and never being received by the study coordinating centre. Foowing randomisation, 287 out of 568 (50.5%) participants were aocated to the intervention group and 281 out of 568 (49.5%) to the contro group. In the contro group, 1 out of 281 (0.4%) participants received two intervention visits in error. In the intervention group, 281 out of 287 (97.9%) of participants received at east one intervention visit. There was no evidence of crossover between the two groups. Six-month foow-up was measured by the receipt of the 6-month questionnaire booket by the research team. For the intervention group, 264 out of 287 (92.0%) reached 6-month foow-up. Of the 23 out of 287 (8.0%) who did not reach 6-month foow-up, 3 out of 287 (1.0%) were ost to foow-up, 14 out of 287 (4.9%) withdrew consent and 4 out of 287 (1.4%) died. A further 2 out of 287 (0.7%) did not compete the foow-up but remained in the study, with 266 participants remaining in the intervention group after 6-month foow-up. For the contro group, 239 out of 281 (85.1%) reached 6-month foow-up. Of the 42 out of 281 (14.9%) who did not reach 6-month foow-up, 8 out of 281 (2.8%) were ost to foow-up, 29 out of 281 (10.3%) withdrew consent and 5 out of 281 (1.8%) died. Therefore, there were 239 participants remaining in the contro group after 6-month foow-up. Competion of 6- and 12-month foow-ups was defined as return of the fuy/partiay competed questionnaire booket to the NCTU. It does not indicate the eve of competeness. Six- and 12-month resuts for individua measures presented ater in the chapter indicate the number of participants from which we have anaysabe data. There was a differentia foow-up rate between the two groups for the 6-month foow-up, with 92.0% coected for the intervention group and 85.1% coected for the contro group. However, these were both ess than the predefined 20% attrition rates. Successfu data coection at 12-month foow-up was measured by receipt of the 12-month questionnaire booket. For the intervention group, 232 out of 287 (80.8%) reached 12-month foow-up. Of the 34 out of 287 ( = 34; 11.8%) who did not reach 12-month foow-up, 3 out of 287 (1.0%) were ost to foow-up, 17 out of 287 (5.9%) withdrew consent and 8 out of 287 (2.8%) died. A further 6 out of 287 (2.1%) did not compete the foow-up but remained in the study. For the contro group 211 out of 281 (75.1%) reached 12-month foow-up. Of the 28 out of 281 ( = 28; 10.0%) who did not reach 12-month foow-up, 4 out of 281 (1.4%) were ost to foow-up, 15 out of 281 (5.3%) withdrew consent and 7 out of 281 (2.5%) died. A further 2 out of 281 (0.7%) did not compete the 12-month foow-up but remained in the study. 20 NIHR Journas Library

49 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Patient invites sent (n = 11126) Repied yes to invite (n = 1448) No repy or repied no (n = 9678) Did not meet incusion criteria (n = 596) Had a screening visit (n = 852) Did not meet incusion criteria (n = 283) Decined to participate (n = 1) Number randomised (n = 568) Aocated to intervention group (n = 287) Received intervention (n = 281) Did not receive intervention (n = 6) Aocated to contro group (n = 281) Received contro (n = 281) Receive intervention (n = 6) No foow-up data a (n = 2) Lost to foow-up (n = 3) Withdrew (n = 14) Deceased (n = 4) Lost to foow-up (n = 8) Withdrew (n = 29) Deceased (n = 5) 6-month foow-up (n = 264) 6-month foow-up (n = 239) No foow-up data a (n = 6) Lost to foow-up (n = 3) Withdrawn (n = 17) Deceased (n = 8) No foow-up data a (n = 2) Lost to foow-up (n = 4) Withdrawn (n = 15) Deceased (n = 7) 12-month foow-up (n = 232) Reached 12 months (n = 211) FIGURE 3 Fina CONSORT diagram. a, Participant remained in the tria but no data were coected at that time point. As with the 6-month foow-up, there were differentia foow-up rates between the two groups for the 12-month foow-up. A tota of 232 out of 287 (80.8%) of questionnaires were coected for the intervention group and 211 out of 281 (75.1%) coected for the contro group. Athough the contro group foow-up rate was within the defined 20% attrition rate, the overa number of participants with 12-month foow-up data (n = 443) was within the threshod to ensure that the study had adequate power. Tabe 2 summarises the resuts of who competed the questionnaire and indicates a consistent proportion of participants competed the questionnaires themseves, whereas at 6 and 12 months there were consistent competion rates for both carer and other (16% and approximatey 37%, respectivey). The other competions at baseine were due to the presence of the RA. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 21

50 RESULTS: RANDOMISED CONTROLLED STUDY TABLE 2 Proportion of who competed baseine and foow-up questionnaires Who competed the questionnaire Questionnaire You (%) Carer (%) Other (%) Baseine (n = 567 a ) 234 (41.3) 32 (5.6) 301 (53.1) 6-month foow-up (n = 503) 230 (45.7) 81 (16.1) 192 (38.2) 12-month foow-up (n = 443) 208 (47.0) 74 (16.7) 161 (36.4) a No baseine questionnaire data were returned for one participant. Carer questionnaire foow-up We received 192 carer questionnaires at baseine, with 100 out of 192 (52.1%) and 92 out of 192 (47.9%) from carers of participants in the intervention and contro groups, respectivey. We received 148 carer questionnaires at 6-month foow-up, with 84 out of 100 (84.8%) and 64 out of 92 (69.6%) from carers of participants in the intervention and contro groups, respectivey. We received 127 carer questionnaires at 12-month foow-up, with 71 out of 100 (71.0%) and 56 out of 92 (60.9%) from carers of participants in the intervention group and contro, respectivey. Trave diary foow-up Owing to the voume of trave diaries and the chaenges of tracking their status, a decision was taken not to send reminders if the trave diary was not received. Overa, 70.6% of a expected trave diary months were received and assigned, regardess of whether they contained data, to 508 out of 568 (89.4%) of participants (Tabe 3); 73.6% were received for the intervention group and 67.5% from the contro group. Overa, 55.1% of participants returned diaries for the fu 12 months. Baseine characteristics Tabe 4 presents a summary of key baseine characteristics of each group. Baseine characteristics of the sampe showed the majority of participants to be women, iving with others, with a mean age of 71 years (SD 12.1 years). At baseine there was a sight gender imbaance, with a higher proportion of men in the contro group than the intervention group (47% vs. 42.2%). Residentia status was we baanced between groups with ony a sighty higher proportion iving aone in the contro group (35.2% vs. 33.4%). The average time since stroke was 37 months (SD 43.8 months) for the contro group and 43 months (SD 60.1 months) for the intervention group. Baseine scores for mobiity, activity measures, GHQ-12 and SWOM were simiar in the two groups. However, for the primary outcome measure of quaity of ife (socia functioning score), the contro group had a higher mean (50.1 vs. 45.9) and median (50 vs. 37.5) than the intervention group. Satisfaction with outdoor mobiity scores indicated that overa 93.3% of participants were not currenty satisfied with their eve of outdoor mobiity. The remaining participants who indicated SWOM (n = 38) were spread across both groups, in a sites except one, and had no other unusua characteristics. TABLE 3 Trave diary foow-up rates Overa (expected n = 7363) Contro group (expected n = 3644) Intervention group (expected n = 3719) Trave diaries n % n % n % Tota received NIHR Journas Library

51 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 TABLE 4 Participant baseine characteristics Aocation Variabe Parameter Contro Intervention Tota n (%) 281 (49.5) 287 (50.5) Age at incusion (years) Mean (SD) 71.5 (12.1) 71.7 (12.1) Median (IQR) 75 (63 81) 73 (64 81) Min. to max Sex Women n (%) 149 (53) 166 (57.8) Men 132 (47) 121 (42.2) Ethnicity White n (%) 263 (93.6) 270 (94.1) Back: Caribbean 7 (2.5) 4 (1.4) Back: African 2 (0.7) 1 (0.3) Back: other 1 (0.4) 3 (1) Pakistani 0 (0) 0 (0) Indian 5 (1.8) 4 (1.4) Bangadeshi 2 (.7) 2 (0.7) Chinese 0 (0) 0 (0) Mixed 0 (0) 1 (0.3) Not given 1 (0.4) 1 (0.3) Other 0 (0) 1 (0.3) Residentia status Lives aone n (%) 99 (35.2) 96 (33.4) Lives with others 170 (60.5) 179 (62.4) Living in care home 12 (4.3) 12 (4.2) Time since stroke (months) Mean (SD) 37 (43.8) 43.2 (60.1) Median (IQR) 21.3 ( ) 24.5 ( ) Min. to max Socia functioning score Mean (SD) 50.1 (30.7) 45.9 (30.3) Median (IQR) 50 (25 75) 37.5 ( ) Min. to max NEADL score Mean (SD) 10.1 (5.7) 8.8 (5.2) Median (IQR) 10 (5 14) 9 (4 13) Min. to max RMI score Mean (SD) 8.9 (4.1) 8.1 (3.9) Median (IQR) 10 (6 12) 8 (5 11) Min. to max continued Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 23

52 RESULTS: RANDOMISED CONTROLLED STUDY TABLE 4 Participant baseine characteristics (continued) Aocation Variabe Parameter Contro Intervention SWOM Yes n (%) 20 (7.1) 18 (6.3) No 259 (92.2) 268 (93.4) Genera Heath Questionnaire (participant) Mean (SD) 15.1 (6.8) 14.9 (6.5) Median (IQR) 13 (10 19) 14 (10 19) Min. to max Max., maximum; min., minimum. It is important to carify the reationship between the SWOM (Do you get out of the house as much as you woud ike?) question as part of the baseine questionnaire and the eigibiity criteria of Wishing to get out of the house more often. On 38 out of 568 (6.7%) occasions, 20 out of 281 (7.1%) in the contro group and 18 out of 287 (6.3%) in the intervention group, the participant indicated that he or she met the eigibiity criteria during the consent process but answered Yes to the SWOM question during competion of the baseine questionnaire. On further investigation it appeared that certain participants were viewing these questions within a different context. The eigibiity question was reated to aspirationa view of getting out of the house (i.e. Wishing to get out of the house more often ), whereas the SWOM question was reated to day-by-day coping views of getting out of the house. Hence these participants were not cassed as ineigibe and were incuded in the anayses. Deivery of the intervention Number of therapists by site There were 29 therapists taking part in this study, who deivered at east one treatment session. There was no restriction on how many therapists deivered the intervention at each site. This was mainy determined by how the service was structured, avaiabiity of staff to perform the research activity, and deivery of research and treatment costs to the reevant department. The therapists ranged from junior to senior (Agenda for Change bands ). Three were physiotherapists, 17 were occupationa therapists and nine were assistant practitioners. The number of therapists per site ranged from 1 to 4, with a median of two per site. A sma proportion of visits were attended by two therapists (65/1939; 3.4%). These were mainy as a resut of training junior staff and training as part of handover to repacement staff. Generay, ony one therapist deivered the intervention whie the second therapist observed as part of training. Number and duration of intervention sessions The intervention group received a median of seven intervention sessions (IQR 3 11 sessions), mean 6.80 sessions (SD 4.01 sessions). Figure 4 iustrates the distribution of intervention visits received by participants, ranging from zero intervention sessions (6/287, 2.1%) to the maximum 12 (61/287, 21.3%); 138 out of 287 (48.1%) received ess than the median of seven visits, with 149 out of 287 (51.9%) receiving 7 visits. These were cacuated from ecrf intervention visit data. Of the 287 intervention participants, paper intervention records competed by the therapists were returned for 269 out of 287 (93.7%) participants, 264 out of 287 (92.0%) of whom had received at east one treatment session. The median duration in tota, in minutes, of intervention provided for these 264 participants was minutes (IQR minutes), mean minutes (SD minutes). These were cacuated from intervention records data. 24 NIHR Journas Library

53 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO Number of participants Number of interventions FIGURE 4 Range of intervention visits. Description of set goas Tabe 5 summarises the types of goas set for each participant. Of the 287 participants in the intervention group, information on goas set was recorded for 243. Participants were abe to set more than one goa during the process. However, we have no direct measure, reating to individua goas set, to indicate the proportion of goas that were achieved. Instead, we have an overa measure of whether the intervention was deivered to the satisfaction of the therapist, detaied ater in Tabe 11. The most common goa was a ong wak of > 100 m, set by 55.1% of participants, whereas increasing confidence was set for 36.2%. There was a wide range of goas set and the vast majority were considered appropriate for an intervention aimed at getting participants out of the house and were aso considered attainabe. Description of the key eements of the intervention The average duration of an intervention visit, incuding trave to and from the participant s home, was 96.6 minutes with a range of minutes. A tota of 1856 out of 1939 (95.7%) intervention visits were competed within the protoco guideines of 4 months post baseine visit. The other 4.3% were outside the 4-month protoco guidance, athough not a were recorded as protoco deviations by the therapists. Tabe 6 summarises the proportion and type of intervention deivered to participants. Goa setting was deivered for 243 (92.1%) participants, with a median of two sessions provided (IQR 1 4 sessions). Mobiity training was deivered the most often for a median of 5.5 visits for 222 participants (84.1%), and had the ongest duration, with a median of minutes (IQR minutes) (data not reported in the tabe). Confidence buiding was used with 202 (76.5%) participants, with a median of four sessions provided (IQR 2 8 sessions). The east used treatment method was adaptive equipment training, with 63 (23.9%) participants receiving a median of one session (IQR 1 2 sessions). A treatment techniques isted on the record form were used at east once by each site. Treatment fideity Of the 15 sites deivering the intervention, 14 were assessed for fideity of treatment. Treatment fideity forms were competed for 59 out of 287 (20.6%) intervention participants. Fideity of treatment assessments were not performed equay across sites, ranging from the intervention sessions for 10 out of 59 (17.0%) participants assessed at one site to 1 out of 59 (1.7%) participant at each of two sites, and no participants at one site. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 25

54 RESULTS: RANDOMISED CONTROLLED STUDY TABLE 5 Summary of type of goas set with participants in the intervention group Goa type n participants Percentage of 243 participants Long wak > 100 m Increase confidence Short wak < 100 m Others a Stamina Access oca shop Bus Attending socia cubs/socia activity Training outdoor mobiity equipment Mobiity scooter Increase independence Access town centre Powered wheechair Crossing roads Taxi Car transfers Increase journeys Driving Dia-a-Ride Wak inside Day centre Shopmobiity Car a Unabe to categorise. TABLE 6 Summary of type of intervention and the number of sessions the intervention type was deivered No. of sessions Intervention type Participants, n (%) a Mean (SD) Median (IQR) Min. to max. Goa setting 243 (92.1) 3.32 (3.13) 2 (1 4) 1, 12 Mobiity 222 (84.1) 6.06 (3.89) 5.5 (2 10) 1, 12 Information 205 (77.7) 3.89 (3.28) 3 (1 10) 1, 12 Confidence 202 (76.5) 5.13 (3.53) 4 (2 8) 1, 12 Other rehabiitation 139 (52.7) 3.55 (2.91) 2 (1 5) 1, 11 Referra 104 (39.4) 2.04 (1.32) 2 (1 3) 1, 7 Adaptive equipment 63 (23.9) 1.92 (1.24) 1 (1 2) 1, 6 Max., maximum; min., minimum. a Percentage out of 264 participants from which we have data. 26 NIHR Journas Library

55 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Initiay, the fideity of treatment assessment was a combination of checking treatment records against a predefined checkist and aso accompanying the treating therapists on visits. However, owing to the possibiity that accompanying the therapists on visits might infuence how a therapist undertook these sessions, it was decided that ony reviewing the participants notes woud be competed for the remainder. Of the 59 fideity of treatment assessments, 12 (20.3%) were competed using the visit and the treatment notes and 47 out of 59 (79.7%) were competed by just assessing the treatment notes. The fina question asked the assessor to make a judgement as to whether the intervention had met the standard based on the checkist. They indicated that in 100% of the cases they beieved it had. Competion of intervention At the end of the participant s intervention period the therapist indicated in 193 out of 287 (67.3%) of participants that they, the therapists, fet that the participant had competed the intervention to the therapist s satisfaction (i.e. a surrogate marker for achieving their set goas). Contamination There was one participant aocated to the contro group who received two intervention visits in error. This was recorded as a protoco deviation and is reported ater in the chapter. The nature of the study, with the intervention requiring a visit to the participant s home, as we as the fact that a participants were not currenty within the rehabiitation service, meant that participants in the contro group were unikey to have any contact with oca site therapists and hence contamination was unikey to occur. There is no evidence of any contamination within this study. Numbers anaysed Figure 2 iustrates the fow of participants and their data; however, it does not provide the detais on an outcome-by-outcome basis. A partiay competed outcome at the particuar foow-up was required for unadjusted measures. In order for any unadjusted outcome measure (apart from trave journeys) to be cacuated, the particuar measure must be at east partiay competed at both baseine and at the particuar foow-up. For trave journeys, the participant had to return at east one trave diary to be incuded in either unadjusted or adjusted anaysis. Tabe 7 detais the number of participants anaysed for both adjusted and unadjusted outcome measure anaysis. TABLE 7 Numbers anaysed for each outcome measure Numbers anaysed 6 months 12 months Outcome Unadjusted Adjusted Unadjusted Adjusted SF-36 (SF domain) NEADL RMI GHQ GHQ-12 (carer) SWOM Trave journeys Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 27

56 RESULTS: RANDOMISED CONTROLLED STUDY Primary outcome The primary outcome was the Socia Function domain within the SF-36v2 reating to an improvement in participant s quaity of ife. The predefined MCID was Tabe 8 summarises the primary outcome anaysis. The variabiity of the socia function score was simiar in the two groups athough the mean score was sighty higher in the intervention group (47.0) at 6 months compared with the contro group (43.9). The adjusted difference in means between groups was with a 95% credibe interva of to This suggests a sighty higher socia function score in the intervention group but this was not significant and it is ess than the pre-defined MCID. The ICCs for the therapist (ICC = ) and site effects (ICC = ) are sma, and ess than we anticipated in our sampe size cacuations. Sensitivity anayses of the primary outcome Sensitivity anaysis was conducted, as per the statistica anaysis pan, to assess the effect of outiers and missing data. The resuts of the sensitivity anayses were consistent with those from the main anaysis and did not resut in any difference in the concusions drawn. Secondary outcomes at 6 months Tabe 9 summarises the secondary outcome anaysis at 6 months. The unadjusted mean scores in both groups were simiar at 6 months for a secondary outcome measures. In the adjusted anayses there was a TABLE 8 Primary outcome anaysis at 6 months Unadjusted Adjusted ICC Outcome Contro, mean (SD) Intervention, mean (SD) Difference in means 95% credibe interva Therapist Site Socia Function domain 43.9 (29.8) (n = 239) 47.0 (30.5) (n = 261) to TABLE 9 Secondary outcomes at 6 months spit by adjusted measure of outcome Unadjusted Adjusted ICC Outcome Contro mean (SD) Intervention mean (SD) Difference in means 95% credibe interva Therapist Site NEADL 9.9 (5.7), (n = 233) RMI 8.4 (4.2), (n = 236) GHQ-12 (participant) 15.4 (6.6), (n = 234) GHQ-12 (carer) 14.4 (6.4), (n = 64) SWOM Yes = 52, no = 181, (n = 233) Trave journeys/day 1.1 (1.2), (n = 241) 9.0 (5.5), (n = 261) 7.8 (4.2), (n = 263) 14.2 (6.0), (n = 261) 13.6 (5.9), (n = 84) Yes = 72, no = 189, (n = 261) 1.0 (1.0), (n = 263) to to to to to to NIHR Journas Library

57 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 statisticay significant difference observed between groups for trave journeys, with the intervention group being 42% more ikey to make a journey than the contro group [rate ratio 1.42, 95% confidence interva (95% CI) 1.14 to 1.67]. There were no other outcome differences between groups. There was evidence of a therapist effect on the trave journeys taken but not for any of the other outcome measures. ICCs for therapist and centre effects were sma, with ICCs for therapist effect being no greater than and ICCs for site effect being no greater than Therapy effect and site effect coud not be cacuated for trave journeys. The other domains from the SF-36v2 were aso anaysed at 6 months and showed no significant difference between groups. Outcome measures at 12 months Tabe 10 summaries the secondary outcomes anaysis at 12 months. The variabiity of the socia function score in both groups at 12 months was simiar to that observed at 6 months but, unike the 6-month findings, the unadjusted mean socia function score at 12 months was sighty ess in the intervention group (45.5) compared with the contro group (48.1) (see Tabe 10). The adjusted difference in means ( 1.24) indicated that the intervention group scores were ower than the contro, athough the difference was not significant and no therapy or site effect was observed for this outcome measure. The unadjusted mean scores in both groups were simiar for a outcome measures. In the adjusted anaysis there was a significant difference for trave journeys in the intervention group; rate ratio 1.76 journeys (95% CI 1.36 to 1.95 journeys). NEADL, RMI, SWOM and GHQ-12 (participant and carer) did not show any observabe differences between groups. The ICCs for therapist effect were arger than those cacuated for a secondary outcome measures at 6 months, with the exception of GHQ-12 (carer). The ICCs for site TABLE 10 Secondary outcomes at 12 months spit by adjusted measure of outcome Unadjusted Adjusted ICC Outcome Contro, mean (SD) Intervention, Mean (SD) Difference in means 95% credibe interva Therapist Site Socia Function domain 48.1 (28.7), (n = 207) NEADL 9.9 (5.9), (n = 210) RMI 8.4 (4.2), (n = 210) GHQ-12 (participant) 14.3 (6.6), (n = 208) 45.5 (31.6), (n = 221) 9.0 (5.8), (n = 228) 7.6 (4.4), (n = 232) 15.2 (6.6), (n = 228) to to to to GHQ-12 (carer) 13.7 (5.9), (n = 56) Outcome SWOM Yes = 45, no = 160, (n = 205) Outcome 13.7 (5.3), (n = 71) to Unadjusted Adjusted ICC Contro Intervention Odds ratio 95% credibe interva Therapist Site Yes = 56, no = 171, (n = 227) to Unadjusted Adjusted ICC Contro, mean (SD) Intervention, mean (SD) Rate ratio 95% credibe interva Therapist Site Trave journeys/day 1.1 (1.3) (n = 241) 1.0 (1.1) (n = 263) to 1.95 Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 29

58 RESULTS: RANDOMISED CONTROLLED STUDY effect, however, were smaer at 12 months than at 6 months, apart from NEADL and participant-reported GHQ-12. There was a significant therapist effect for trave journeys at 12 months; however, we are unabe to estimate the size of the effect because this is not cacuabe for this type of data (see Tabe 10). Robustness of resuts We assessed the robustness of the findings by repeating a anayses and adjusting for the baseine variabes gender and residentia status in addition to the other covariates aready adjusted for in the modes. The resuts of the anayses were consistent with those obtained from our origina anaysis. We were unabe to conduct this anaysis for the GHQ-12 carer outcome because the modes woud not converge. Sensitivity anayses The ony potentia outiers we found were for the GHQ-12 outcome measures for the participant and carers. After Winsorising these potentia outiers we re-anaysed the data but the findings were consistent with the origina anaysis. We aso conducted a sensitivity anaysis using MI to repace missing data but this did not change the findings of the study. Exporatory/other anayses Nottingham Extended Activities of Daiy Living by category No significant differences between groups were found for any of the four categories (mobiity, kitchen, domestic and eisure) at 6 or 12 months. Number of sessions by site Tabe 11 detais a site-by-site breakdown of the number of sessions deivered by therapists and the corresponding percentage of participants who competed the intervention to satisfaction. The median number of visits for this study was seven so this is a suitabe point of comparison within the context of this study. Of the seven sites with a median of < 7 visits, ony 3 out of 7 sessions (42.9%) had > 50% of participants who competed the intervention to satisfaction. Of the eight sites whose median was 7 visits, 8 out of 8 sessions (100.0%) had > 50% of participants who competed the intervention to satisfaction. The piot study, however, showed that the intervention was effective with a median of six visits, 24 so this is a suitabe predefined point of comparison in terms of potentia cinica effect. Of the five sites that had a median of < 6 visits, ony 2 out of 5 sessions (40.0%) had > 50% of participants who competed the intervention to satisfaction. Of the 10 sites that had a median of 6 visits, 9 out of 10 sessions (90.0%) had > 50% of participants who competed the intervention to satisfaction. The resuts are exporatory but are suggestive that the more intervention visits a participant had the more ikey they were to compete the intervention to satisfaction. Number of trave journeys by intervention session A further exporatory comparison of trave journeys within the intervention group, based on the number of sessions they received, was carried out (Tabe 12). From the piot study there was evidence that receiving six intervention visits woud increase the number of journeys made, therefore it is considered to be a cinicay important pre-existing threshod. The difference between trave journeys made by participants in the intervention group was summarised by < 6 and 6 or more intervention visits at both 6 and 12 months. These data show that participants who had six or more intervention visits were more ikey to have a higher number of journeys at both 6 months (167 vs. 148) and 12 months (316 vs. 291). The resuts are exporatory but suggest that the more intervention visits a participant has the more ikey they are to take more journeys beyond the competion of the intervention. 30 NIHR Journas Library

59 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 TABLE 11 Percentage intervention competed to satisfaction by site and average number of session deivered Site Percentage who competed intervention to satisfaction No. of sessions Mean (SD) Median (IQR) Min. to max. A (3.79) 2 (1 6) 1, 12 B (4.16) 4 (2 9) 1, 12 C 40 5 (4.18) 4 (1 10) 1, 11 D (2.67) 5 (2 5) 2, 10 E (3.79) 5 (4 11) 4, 11 F (3.84) 6 (4 12) 2, 12 G (3.01) 6.5 (4.5 8) 1, 11 H (3.14) 7 (4 9) 1, 12 I (3.89) 7 (5 11) 1,12 J (3.47) 8 (4 10) 1, 12 K (3.24) 8 (5 11) 2, 12 L (3.32) 9.5 (6 12) 2, 12 M (3.63) 10 (5.5 12) 1, 12 N (3.55) 10 (5 12) 3, 12 O (3.37) 12 (11 12) 2, 12 Max., maximum; min., minimum. TABLE 12 Number of journeys made within the intervention group by number of intervention sessions No. of intervention sessions Foow-up Outcome measure <6(n = 109) 6(n = 156) 6 months No. of journeys 16,141 25,979 Mean (SD) 148 (176) 167 (169) Median (IQR) 92 (36 199) 111 (61 247) Min. to max months No. of journeys 31,679 49,214 Mean (SD) 291 (385) 316 (321) Median (IQR) 160 (54 381) 208 (94 464) Min. to max Max., maximum; min., minimum. Fas by age category and aocation A data described in this section were coected from trave diaries. Tabe 13 presents fas data described by age, either < 60 years of age or 60 years. At randomisation, 104 participants (18.3%) were < 60 years of age, whereas 464 (81.7%) were 60 years. A tota of 52 out of 104 (50%) of the participants aged < 60 years and 216 out of 464 (46.6%) of the participants aged Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 31

60 RESULTS: RANDOMISED CONTROLLED STUDY TABLE 13 Fas data summary by age Fa-days < 60 year 60 years Tota Participants, n (%) 104 (18.3) 464 (81.7) 568 (100) Participants with recorded fas, n (%) 52 (50) 216 (46.6) 268 (47.2) Tota of fa days, n (%) 704 (36.1) 1246 (63.9) 1950 (100) Mean (SD) 13.5 (22.2) 5.8 (10.2) 7.3 (13.7) Median (IQR) 3 (1 13) 3 (1 5.5) 3 (1 6.5) Min. to max. 1, 89 1, 79 1, 89 Max., maximum; min., minimum. 60 years had fas. In tota, there were 1950 fa-days: 704 (36.1%) in the < 60 years age group and 1246 (63.9%) in the 60 years group. The median number of fa-days overa was 3 per year (IQR fa-days). There appeared to be no difference between the median fa days of the two age groups. Tabe 14 presents fas data described by treatment aocation. The proportion of participants who had a fa was simiar in each group: 133 out of 281 (47.4%) in the contro group and 135 out of 287 (47%) in the intervention group. Comparing changes in satisfaction with outdoor mobiity over time There was very strong evidence that the contro group improved markedy. At baseine, 259 out of 281 (92.2%) participants were dissatisfied with outdoor mobiity, but at the 6-month assessment this had reduced to 78% (160/205), a 15% reduction. The corresponding reduction in the intervention group was sighty greater (18%), with 268 out of 287 (93.4%) expressing dissatisfaction with outdoor mobiity at baseine and 171 out of 227 (75.5%) expressing this at 6-month assessment. This suggests that the contro (consisting of the baseine visits and competion of the trave diary) may have affected a change. Six-month foow-up by foow-up approach Tabe 15 detais the questionnaire data coection from the different methods of approach. For 6-month questionnaires we received the questionnaire bookets for 503 out of 568 (88.6%) participants. The overa average difference in days from actua competion to expected due date was +4.7 days (range 33 to +133 days). Of those received, 280 out of 503 (55.7%) were via posta approach, 185 out of 503 (36.8%) were via the RA approach, with the remaining 38 out of 503 (7.6%) via the RA approach after switching from posta approach. RA assistance did not necessariy mean that the RA competed the TABLE 14 Fas data summary by treatment aocation Aocation Fa-days Contro Intervention Tota Participants, n (%) 281 (49.5) 287 (50.5) 568 (100) Participants with recorded fas, n (%) 133 (47.3) 135 (47.0) 268 (47.2) Tota of fa days, n (%) 934 (47.90) 1016 (52.10) 1950 (100) Mean (SD) 7.0 (13.0) 7.5 (14.5) 7.3 (13.7) Median (IQR) 2 (1 7) 3 (1 6) 3 (1 6.5) Min. to max. 1, 80 1, 89 1, 89 Max., maximum; min., minimum. 32 NIHR Journas Library

61 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 TABLE 15 Proportion of questionnaire received according to approach used, at 6 months Six months Tota received Percentage of tota Tota posta approach Posta to RA a Tota RA approach Tota a Posta to RA indicates that posta approach was used initiay before switching to RA approach. questionnaire, just that assistance was provided; however, in the majority of cases the RA woud ask the questions and compete on the participant s behaf. Adverse events Adverse events records (i.e. a fa that required the assistance of a heath-care professiona) were coected from ony participants in the intervention group over the course of deivering the intervention visits. There were 20 adverse events from 17 participants, with the majority of fas occurring at home. Two were recorded as ongoing, with the average mean duration of event being 5.6 days and median duration of 1 day (range 1 44). Overa, there was no effect on the deivery of the intervention and ony one ed to permanent discontinuation of the intervention. Overa, 24 participants died, 12 in each group, indicating that the intervention group did not have an increased risk of death compared with what woud be expected for this patient group. As part of the safety anaysis there was aso no significant difference in GHQ-12 score observed between the groups. Protoco deviations There were 68 protoco deviations recorded, with 37 out of 68 (54.4%) for visits not performed within the 4-month window (the majority of these were as a resut of adverse weather conditions or temporary participant or therapist unavaiabiity). However, from ecrf intervention data there were a tota of 83 visits deivered outside the 4-month window, so there was an issue recording these as protoco deviations. Twenty-five out of the 68 participants (36.8%) recorded other intervention therapy, which refers the intervention being deivered via non-protoco-defined methods (in a of these cases the intervention was deivered either by teephone or etter). One out of 68 participants (1.5%) in the contro group received the intervention (two visits); 1 out of 68 participants (1.5%) was in active rehabiitation at point of recruitment; and 1 out of 68 participants (1.5%) was an eigibiity deviation, as the participant was unwe and not fit to receive treatment, whereas 1 out of 68 participants (1.5%) had a consent issue that was ater resoved. Finay, 2 out of 68 participants (2.9%) had an intervention deay owing to unforeseen circumstances. We considered these a to be minor deviations and that they were unikey to have any effect on the outcome measure. As a resut, no action was taken and a of these participants remained in the study and in the fina anaysis. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 33

62 RESULTS: RANDOMISED CONTROLLED STUDY Conceament of aocation There was a tota of 223 RA-assisted participant visits at 6 months (see Tabe 15), for which the participant-reported primary outcome measure was coected. We received binding assessments for 171 out of 223 (76.7%) [intervention group 88 out of 171 (51.5%); contro group 83 out of 171 (48.5%)] RA-assisted participant visits at 6 months (Tabe 16). The RA was unbinded for 87 out of 171 (50.9%) of those visits, with 13 out of 171 (7.6%) unbinded prior to the visit and 74 out of 171 (43.3%) unbinded during the visit. The RAs were unbinded more frequenty for participants in the intervention group (48/87, 55.2%) than for those in the contro group (39/87, 44.8%), athough this is not indicative of any significant difference in unbinding rates between the two groups. Summary The study reached its recruitment targets and achieved strong retention rates at 6 months. Differentia foow-up occurred but not to an extent as to affect the power of the study. The attrition rates were greater within the first 6 months of foow-up. The quaity of ife (socia function) measure (primary outcome) showed no significant difference between groups at 6 months or at 12 months. Six- and 12-month measures for functiona mobiity, SWOM, and participant and carer we-being showed no significant differences between groups. The 6- and 12-month measure of trave journeys showed a significant difference in favour of the intervention group when the therapist and site effect was taken into consideration. There was no evidence of a therapy or site effect in any of the outcome measures at either 6 or 12 months, apart from trave journeys at 6 and 12 months. The intervention was deivered a median of seven times across the study as a whoe, with 67.3% competed to the satisfaction of the participant; however, there was considerabe variation amongst the sites for both these outcomes. Exporatory data suggest that the more intervention visits a participant receives, the more ikey they are to take more journeys beyond the competion of the intervention. Exporatory data suggest that the more intervention visits a participant had, the more ikey they were to compete the intervention to satisfaction. The contro group appears to have benefitted from incusion in the study by becoming more satisfied with their outdoor mobiity over time. TABLE 16 Awareness of treatment aocation of RAs prior to, or during, RA-assisted participant visits of 6 months primary outcome coection Six months Prior During Tota n visits 171 n unbinded % n visits NIHR Journas Library

63 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 4 Economic evauation: methods and resuts Heath economic evauation: methods used This component of the study aimed to extend the evidence base by estimating for the first time the incrementa cost-effectiveness of an outdoor mobiity rehabiitation intervention, compared with contro, from a heath and persona socia services perspective. In addition, carer time and certain patient-borne costs were aso estimated. It was decided that medication costs woud not be monitored, as it was considered that the intervention woud not infuence these. A within-tria anaysis was conducted over a 12-month period, where an avaiabe case anayses approach 41 was adopted. Thus, for each variabe, we anaysed a avaiabe data, which meant a different sampe n was used for different variabes. For exampe, the EQ-5D and SF-6D may have different response rates. Additionay, in the base-case anaysis, it shoud be noted that when cacuating eves of cost-effectiveness [see Cost-effectiveness, beow, for a definition of incrementa cost-effectiveness ratio (ICER)] we incuded ony those participants who had both compete cost and effect data. As such, the number of participants for whom these data are avaiabe may we differ from the n for which separate cost and effect data are avaiabe. Methods Measuring costs Overview For each participant the tota NHS and Persona Socia Services (PSS) costs were estimated by summation of the intervention cost and other NHS and PSS costs. Carer input (coated in terms of ost productivity) and certain patient-borne costs were aso estimated. Costs were estimated in UK stering ( ) at financia year eves. Intervention Training Therapists were provided with training in order to deiver the intervention. Three types of training were provided: (1) group training in Nottingham; (2) individua training to a particuar site therapist; and (3) within-site cascade training. The time inputs by a members of staff were estimated for each of these training methods, incuding preparation and trave time. The unit cost of NHS community therapy time, as estimated by Curtis et a. 42 was assumed to appy to a time inputs, where this was adjusted to be grade 7 for the individua and group trainer (compared with grade 5 for the cascade trainer and for those receiving the training and subsequenty deivering the intervention). Trave distances were aso estimated, and assigned a trave cost of 54p per mie. 42 Tota training costs (the sum of staff and trave costs) were apportioned across a participants aocated to the intervention arm of the study. Therapy contacts Participants in both arms received a baseine visit; this was assumed to ast 1 hour, incuding preparation, trave time and the writing up of notes, as we as the patient contact. Generay, this was provided by one therapist, athough two therapists may undertake this (or subsequent visits) if there were issues reating to handover/safety, thus those who provided the intervention were asked to note which therapist(s) attended each visit (if this was not recorded it was assumed that one therapist provided the intervention). The associated trave costs for the baseine, and a subsequent therapy visits, were estimated in the foowing Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 35

64 ECONOMIC EVALUATION: METHODS AND RESULTS manner. After discussion with those who deivered the service, it was assumed that a trave time of 12 minutes woud appy to each visit (this is in ine with a previous assumption in reation to GP home visits 42 ). Assuming an average speed of 29.9 mies per hour (mph) (the average free-fow vehice speed in a 30-mph imit, 43 this woud equate to a tota mieage of 5.98 mies per visit. Appying a trave cost of 54p per mie, 42 this woud equate to a trave cost of 3.23 per visit. Those in the intervention arm received extra intervention visits, where the time associated with these visits (incuding trave to and from the participant s home) was prospectivey recorded by the therapist providing the intervention. It was additionay assumed that, for each visit, there woud be 5 minutes preparation time and a further 10 minutes to write up associated patient records. Again for these visits the associated trave time was assumed to be 12 minutes, with a trave cost of 3.23 per visit. Thirty-minute supervisor meetings were aso assumed to occur (one per site per month) for the duration of the study, where these were assumed to be 1 : 1 (therapy grade 7 and grade 5). Tota supervisor meeting costs were equay apportioned across a visits. No other costs were incuded, as these were considered negigibe (e.g. occasiona bus trips with the participant, as part of the intervention). Other NHS and Persona Socia Services costs Leves of resource use The UK Nationa Institute for Heath and Care Exceence (NICE) recommends that costs can be cacuated from the perspective of the NHS and PSS. 44 Accordingy, at both 6 and 12 months post randomisation, participants were asked to compete a resource-use questionnaire and return it by post. In this they were asked the number of times they had received different NHS and PSS services, any other care and certain patient costs that had been incurred. Specific questions with regard to heath-care professiona visits (and where they took pace), hospita attendances and admissions, residentia/nursing home admissions, home hep (from community care assistant/someone who ives with them/other friends or famiy), Meas on Whees and equipment purchased (to hep with a heath probem) were incuded. Assumptions made in order to assign costs to items of resource use Previousy estimated unit costs 42,45 were assigned to eves of resource use, where the foowing assumptions were made. Within Curtis et a., 42 the ength of time and/or cost associated with patient contacts (incuding home visit) is not stated for many heath professionas, athough the cost per hour of empoyment is generay avaiabe. Costs in terms of per hour of empoyment, per practice visit and per home visit are avaiabe, however, for the GP. Thus, the ratio of 1 hour of empoyment compared with (1) a practice visit and (2) a home visit can be cacuated. This ratio was appied to the costs per hour of empoyment of other staff to estimate associated practice/hospita visit and home visit costs. Where the cost per hour of empoyment, for a particuar heath professiona, was not reported within Curtis et a. 42 the average cost across the foowing heath professionas was used: genera practitioner, practice nurse, district nurse, dietitian, physiotherapist, occupationa therapist, socia worker, speech and anguage therapist, for the respective type of visit (GP, home or hospita). If the pace of a heath professiona visit was not reported then it was assumed that the patient traveed to the heath professiona (patient trave costs were not estimated/incuded). With regard to hospita admissions (in the past 6 months), if the ength of a hospita admission was not reported the mean ength of stay (per admission) for other respondents (who reported both the number of admissions and the accompanying ength of stay) was appied to each admission that was reported but had no accompanying ength of stay. We aso asked about the number of times a person had been admitted to both a residentia home and a nursing home (in the past 6 months). We did not request that participants report the ength of any associated stays in such care home we thereby made the assumption that each time a person reported they were admitted to a residentia/nursing home they had stayed there for three of the preceding 6 months (the average ength of stay in a care home has been estimated to be 801 days 46 and we 36 NIHR Journas Library

65 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 assumed that participants were on average admitted haf-way through the 6-month period). If more than one admission was reported, it was assumed the participant had been in the home for the whoe 6-month period in question. Participants were asked whether they had attended a day-care centre in the past 6 months, and if they had, how many times per week they attended. The number of reported day centre attendances per week was assumed to appy to a weeks in the 6-month period in question. Participants were asked to report the number of times (in the past week) they had home hep or a visit from a community care assistant, and how ong that person stayed. We requested that participants report the average time per visit. If the average time per visit was not reported, the average time per visit for other respondents (who reported both the number of times and the accompanying average time per visit) was appied to each visit that was reported but had no accompanying time per visit. Additionay, some of the responses were higher than what we considered to be a possibe visit ength. For exampe, one participant reported 28 visits and a visit ength of 960 minutes (16 hours). This equates to more hours than there are in a week. We thereby assumed that this participant, and a others for whom the product of the number of visits and the associated time was greater than the number of hours in a week (a tota of four participants at 6 months and five participants at 12 months), had misinterpreted the question and reported the tota ength of contact in the week, rather than the average per visit. Thus for these participants the reported visit ength was assumed to be the tota for the whoe week. Again, it was assumed that the number of visits/hours reported for the week in question appied to a weeks in the past 6 months. In ine with the therapist intervention, the associated trave time for each home hep visit was assumed to be 12 minutes, with a trave cost of 3.23 per visit. Participants were additionay asked how many times they had received hep from someone they ived with, in the past week and the average associated ength of time. They were aso asked to report the eve of such hep from peope they do not ive with. In both these questions, we requested that peope report the average ength of time (per visit) for the hep they received. If the average time per visit was not reported, the average time per visit for other respondents (who reported both the number of times and the accompanying average time per visit) was appied to each visit that was reported but had no accompanying time per visit. Aso, for some participants, the product of the number of visits and the associated time was greater than the number of hours in a week (with regard to someone they ived with this occurred for a tota of two participants at 6 months and two participants at 12 months, and for peope they do not ive with this occurred for one participant at 6 months). We thereby assumed that these participants had misinterpreted the question and reported the tota ength of contact in the week, rather than the average per visit. Thus for these participants the reported visit ength was assumed to be the tota for the whoe week. Again it was aso assumed that the number of visits/hours reported for the week in question was equivaent to the average per week across a weeks in the past 6 months. Within both these questions, participants were asked whether the person who provided the hep had had to take time off work to provide this hep. In order to provide an estimate of ost productivity, a cost was ony appied to the tota number of hours they were estimated to have received (both for peope they ive with, and do not ive with) if they reported that the person that provided the care had to take time of work to provide such care. In this case, the average houry earnings 47 was appied to these times, consistent with the human capita approach. 48 However, as the costing of such care is sometimes considered controversia, 49 these costs are aso reported separatey to the aforementioned NHS and PSS costs. Participants were asked to report if they had Meas on Whees and, if so, the number that they had received in the past week, and, if they paid for them, the amount they cost. Meas on Whees that were not reported to be paid for by participants were assigned a previousy estimated unit cost 50 and cassified as a PSS cost. The costs for those who reported that they paid for them themseves were cassified as patient-borne costs. Again, it was assumed that the number of meas reported for the week in question was equivaent to the average per week across a weeks in the past 6 months. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 37

66 ECONOMIC EVALUATION: METHODS AND RESULTS Participants were aso asked to report any equipment they had bought, or been given, to hep with a heath probem and, if so, to state the equipment, who paid for it (the participant/socia services) and the cost of the item. Items that were not reported to be paid for by socia services were cassified as a patientborne cost. When a cost was not stated, where possibe, previousy estimated unit costs were assigned to items (where these were taken from, e.g. Curtis et a. 42 ). When a unit cost for an item coud not be identified, or the type of item was not reported, either the cost of what was considered to be a simiar item or the average cost of a items for which a unit cost was identified was assigned to the item in question. This question did not specify the time frame over which it was interested in equipment purchases; consequenty, even athough a other questions specified the previous 6 months, it is possibe that some of the reported items may have been purchased before the participant joined the study. The potentia impact of this was reduced, however, as the equivaent annua cost of equipment purchases was cacuated, 48 for which the discount rate was assumed to be 3.5% per annum and the ifespan of the equipment was assumed to be 7 years. Categorisation of costs The above enabed a cost to be assigned to each of the resource-use questions. These were then categorised, as foows, where a costs reate to the 12-month post-intervention period: the costs associated with heath professiona or home hep visits; visits to accident and emergency, wak-in centres, outpatients, day centres; and admissions to hospita, residentia homes and nursing homes. Meas on Whees and equipment, for which the participant did not pay, were summed to estimate other NHS and PSS costs. Tabe 17 provides detais of the resources monitored within each question. These were, TABLE 17 Description of the costs associated with the intervention and contro Resource item Leve of resource use Associated unit cost ( ) Associated tota cost ( ) Per-participant cost ( ) Provision of training hours of (centra) trainer time Associated trave costs (800 mies) 14 hours of cascade training per hour a per mie a per hour a Receipt of training 62 hours per hour a Associated trave costs (900 mies) 0.54 per mie a Overa training cost Therapy contacts Therapist time per hour a 125, (mean = 6.76 visits) b,c Supervision meetings Associated trave costs (12 mies per visit) 1 per month (over 18 months) in each of the 15 sites 0.54 per mie a per meeting 10, Overa intervention cost 146, Contro visit Incrementa cost of the intervention 1 visit, 1 hour in duration c a Based on Curtis et a. 42 b Average contact time equas minutes with 12 minutes trave. c Costs for more than one therapist have been incuded, if appicabe. 38 NIHR Journas Library

67 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 in turn, added to the intervention costs to provide an estimate of tota NHS and PSS costs (base-case anaysis). The costs associated with hep from peope they ive or do not ive with were added together to provide an estimate of ost productivity. Simiary, the costs associated with Meas on Whees and equipment, for which the participant did pay, provided an estimate of the costs borne by the patient. Finay, overa costs were estimated by the summation of intervention costs, other NHS and PSS costs, ost productivity and patient-borne costs. Overa and incrementa costs For each of the aforementioned cost categories, the mean incrementa cost of the intervention (over the 12-month foow-up period) was cacuated by subtracting the mean cost for the contro group from the mean cost for intervention group. Measuring outcomes To estimate the impact on heath-reated quaity of ife, participants were asked to compete the EQ-5D 51 at baseine, 6 and 12 months post randomisation. The EQ-5D has five questions, through which the respondent is asked to report the eve of probems they have (no probems, some/moderate probems, and severe/extreme probems) with regard to mobiity, sef-care, usua activities, pain and anxiety/depression. 30 The three-eve version of the EQ-5D (EQ-5D-3L) was used. Responses to these five dimensions are converted into one of 243 different EQ-5D heath-state descriptions, which range between no probems on a five dimensions (11111) and severe/extreme probems on a five dimensions (33333). A utiity score (a scae where death = 0 and fu heath = 1) was assigned to each of these 243 heath states using the York A1 tariff 52 (associated EQ-5D scores range between and 1.00). Competion of the EQ-5D enabed a cost utiity anaysis to be undertaken, in which the benefits of different heath-care treatments can be compared on a common utiity scae. 53 The area-under-the-curve method 53 was used to estimate the mean quaity-adjusted ife-year (QALY) gain/oss over the 12-month tria period for both groups. Within these QALY cacuations, those who died within the study period were assigned a utiity score of 0 upon death. In a simiar way, responses to 11 of the questions on the SF were used to estimate a score on the SF-6D. 25 The SF-6D is composed of six dimensions (physica functioning, roe imitations, socia functioning, pain, menta heath and vitaity), which have between four and six eves. A non-parametric mode (which uses Bayesian methods) 55 was used to estimate SF-6D heath-state utiity vaues for each of 18,000 potentia heath states (associated SF-6D scores range between and 1.00). QALY gains/osses were again cacuated, as for the EQ-5D, and those who died were, again, assigned a score of 0. As the EQ-5D and SF-6D utiity measures are based on both different heath-state descriptions and use different vauation techniques, they coud produce different utiity scores for the same group of patients. This study therefore sought to expore the impact the choice of utiity measure had on estimates of cost utiity, as further research has been argued to be necessary in this area. 56 It shoud be noted, however, that NICE currenty recommends that the EQ-5D be used within the reference case anaysis, 39 and thus this constituted our main base-case anaysis (see beow). Base-case anaysis Mutipe regression 57 was used to estimate the mean cost difference (incrementa cost) and mean QALY difference (incrementa effect) between the two treatment groups, where both the overa cost and the mean QALY gain/oss over the 12-month period were adjusted for baseine utiity, age, sex and residentia status (the ast three variabes were chosen as they were the ony participant variabes that were avaiabe for a participants who had compete cost and QALY data). The mean QALY difference was estimated for both the EQ-5D and SF-6D. In ine with the cinica anaysis, within the base-case anayses ony participants with compete cost and QALY data were incuded. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 39

68 ECONOMIC EVALUATION: METHODS AND RESULTS Cost-effectiveness After checking that dominance was not apparent (this woud occur if one intervention were ess costy and more effective than another), 53 the incrementa cost per QALY gain (ICER) associated with the intervention was cacuated (mean incrementa cost/mean incrementa QALY gain). In ine with NICE guidance 44 we compared the ICER to a cost-effectiveness threshod (λ) of 20,000 per QALY. Decision uncertainty The bootstrap technique 58 (with 5000 repications) was used to estimate the 95% CIs surrounding the incrementa cost and incrementa effect (where appropriate), the 95% CI was estimated using the percentie method. 59 As the ICER has the potentia to be misinterpreted, 60 we aso estimated the incrementa net benefit (INB) (and associated 95% CI) at a threshod of 20,000 per QALY. A negative INB woud indicate that the intervention was not cost-effective at this threshod. The bootstrap sampes were aso used to estimate the cost-effectiveness acceptabiity curve (CEAC) for each group, where the CEAC depicts the probabiity that an intervention is cost-effective at different eves of λ. 61 The probabiity of the intervention being cost-effective was specificay estimated at the (λ) of 20,000 per QALY. Sensitivity anaysis We assessed how robust concusions were to the foowing changes: 1. Received six or more intervention visits: i. On the assumption that those who received more visits might benefit more, here, ony intervention participants who were incuded in the base-case anaysis and had six or more visits were incuded in the anayses. The contro group was the same as for the base case. 2. MI: i. To impute missing data in this data set, we used regression methods to predict these vaues based on their reationship with other covariates (age, sex, residentia status, cost and utiity data). Imputation took pace in 10 cyces, the estimates from which were then pooed and cacuated using Rubin s Rues. A MI was performed for incompete cost and outcomes components at the patient eve using the mi impute mvn procedure in Stata Winsorising: i. As outined previousy, for the cinica data, we repaced data vaues beow the 5th percentie with the 5th percentie vaue and to data vaues above the 95th percentie with the 95th percentie vaue. This was appied to the cost and QALY data for those individuas incuded in the base-case anaysis. One reason for undertaking this anaysis was the wide variation in some of the resource-use data that was reported, e.g. home hep. This approach reduces the infuence of extreme vaues and may partiay test whether or not some of the assumptions in reation to these costs were correct. ii. Different cost perspective. Resuts were re-estimated from an overa cost perspective. 40 NIHR Journas Library

69 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Resuts Costs Training Training costs were as foows (see Tabe 17). Tota trainer time associated with group training (provided to eight sites at once) and individua training (to a further nine sites) was estimated to be hours (incuding trave and preparation). Tota trainer time for the cascade training (provided at seven sites) and time for receipt of a types of training (incuding trave) was estimated to be 14 and 62 hours, respectivey. Trainer (group and individua) unit costs were estimated to be 45, compared with 31 per hour for cascade trainers and trainees. The tota cost of a staff time associated with a aspects of training was thereby estimated to be , with the addition of 918 for trave costs to sites (eight peope attending the group training and nine site visits, with an average return trip of 100 mies), this gave a tota training cost of Apportioned across a 287 participants in the intervention arm, this equates to per participant. Therapy contacts A participants received a baseine visit. This was the ony contact for those in the contro arm and the associated mean tota cost (staff time and trave costs) was estimated to be per participant (see Tabe 17). (It shoud be noted that one participant in the contro received two intervention visits no associated intervention costs were assigned to this participant as this was provided in error.) The costs associated with therapist visits to those in the intervention arm are summarised in Tabe 17. On average, participants in the intervention arm received a further 6.76 visits (range 0 12). The associated time for these visits was recorded for a but 1 of the 1939 visits, for which the mean time was minutes (assuming a trave time of 12 minutes, this equates to an average contact time of minutes). This mean vaue was assumed to appy to the visit where the time was not recorded. Each supervisor meeting was estimated to cost 37.87, one per month per site were estimated to occur across the 18-month period for which the intervention was provided. This equates to a tota cost of 10, across a sites: 5.24 per visit undertaken. Visit, preparation and records write-up time were each costed at 31 per hour and, after adding the supervision cost ( 5.24 per visit) and trave cost ( 3.23 per visit), the mean cost of the intervention was estimated to be per participant, where this increased to after incuding the aforementioned training costs. The incrementa intervention cost, compared with the baseine visit provided to the contro arm (cost 34.78) was thereby estimated to be per participant. Other NHS and Persona Socia Services costs The 6-month resource-use questionnaire was returned by 259 out of 287 participants in the intervention arm and 235 out of 281 in the contro arm; the numbers at 12 months were 230 and 209, respectivey. However, not a returned questionnaires were fuy competed and we accordingy note the response rate to each of the individua questions in Tabe 18. (Note: A response was required at both 6 and 12 months in order for the 12-month cost to be estimated.) Mean eves of resource use, reating to each of the cost questions, are shown in Tabe 18, in which it can be seen that participants frequenty visit heath professionas, outpatients and receive home hep. The unit costs appied to the reported eves of resource use are summarised in Tabe 19. Subsequenty, costs were categorised into the foowing groups: other NHS and PSS costs, tota NHS and PSS costs, ost productivity, and overa costs (see Tabe 18 for detais of which questions contributed to each cost category). The mean cost (per participant) for the intervention and contro groups, for each of these cost categories, are presented in Tabe 20, in which it can be seen that, for each of these cost categories, the mean costs are estimated to be higher for the intervention group. It shoud be noted, however, that the number of participants for whom compete cost data are avaiabe fas when a broader perspective is taken. This can be expained argey by the fact that responses are required to a greater number of resource-use questions. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 41

70 ECONOMIC EVALUATION: METHODS AND RESULTS TABLE 18 Estimated eves of resource use and associated cost (mean per participant over the 12-month period) Respondents, n Leves of resource use Mean cost ( ) Item Intervention Contro Intervention Contro Intervention Contro No. of heath professiona visits a No. of hospita admissions a No. of A&E visits a No. of wak-in centre visits a No. of outpatient visits a No. of admissions to residentia homes a No. of admissions to nursing homes a No. of day centre visits a No. of home hep visits a Hours of hep from someone they ive with b Percentage who took time off work % 3.10% Hours of hep from someone they do not ive with b Percentage who took time off work % 3.10% Meas on Whees Cost when the participant did not pay a Cost when the participant paid c Equipment % 77.80% Cost when the participant did not pay a Cost when the participant paid c A&E, accident and emergency. a Reate to other NHS and PSS costs. b Reate to estimated carer input (ost productivity costs are appied ony when time off work was reported). c Participant-borne costs. 42 NIHR Journas Library

71 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 TABLE 19 Unit costs attached to different items of resource use, with associated source Item Estimated unit cost ( ) GP visit a Practice nurse visit a 9.90 District nurse visit a Dietitian visit a 9.30 Physiotherapist visit a 9.60 Occupationa therapist visit a 9.60 Socia worker visit a Speech and anguage therapist visit a 9.30 Cost per day in hospita (non-eective inpatient excess bed-day cost) A&E visit (non-admitted cost used) Wak-in centre visit (non-admitted cost used) Outpatient visit Admission to residentia home (weeky cost) Admission to nursing home (weeky cost) Day centre (per visit cost) Home hep b (cost per hour of face-to-face contact) Hep from someone they ive with (who took time off work) (cost per hour) Meas on Whees (where the participant did not pay) Equipment: c Grab rai Hoist Aarm system Stair ift Concrete ramp Wheechair A&E, accident and emergency. a Costs presented are for a visit to the professiona by the participant (houry empoyment costs were taken from Curtis: 42 see Methods for assumptions used to estimate unit costs). Estimated home visits costs, and the costs associated with visits to other heath professionas, are avaiabe from authors. b Assumed to cost equivaent to a home care worker. c Participants were asked to report the cost of equipment if they paid for it themseves. Exampes of unit costs appied to equipment reported to be provided by socia services are provided. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 43

72 ECONOMIC EVALUATION: METHODS AND RESULTS TABLE 20 Estimates of the mean cost ( ) and QALYs associated with each intervention over the 12-month study period Estimated variabe Intervention Contro Difference Intervention costs, (1) (n = 287), (95% CI to ) (n = 281), (95% CI to 34.23) (95% CI to ) Other NHS and PSS costs (2) 11, (n = 162), (95% CI to 55,886.08) (n = 148), (95% CI 133 to 44,954.34) (95% CI 35, to 48,398.27) Tota NHS and PSS costs (1 + 2) 12, (n = 162), (95% CI to 59,692.78) (n = 148), (95% CI to 44,988.56) (95% CI 38, to 51,687.98) Lost productivity (3) (n = 178), (95% CI 0.00 to ) (n = 165), (95% CI 0.00 to ) (95% CI to ) Participant costs (4) (n = 203), (95% CI 0.00 to ) (n = 189), (95% CI 0.00 to ) (95% CI to ) Overa cost 14, (n = 143), (95% CI to 70,841.91) (n = 135), (95% CI to 36,952.20) (95% CI 30, to 58,969.86) ( ) a EQ-5D score: Baseine (n = 281), (95% CI to 0.848) (n = 280), (95% CI to 0.850) 6 months (n = 255), (95% CI to 0.848) (n = 229), (95% CI to 0.883) 12 months (n = 223), (95% CI to 0.850) (n = 204), (95% CI to 1.000) Change (over 12-month period) (n = 218), (95% CI to 0.435) (n = 203), (95% CI to 0.651) (95% CI to 0.691) QALY b (n = 223), (95% CI to 0.832) (n = 207), (95% CI to 0.844) (95% CI to 0.630) SF-6D score: Baseine (n = 271), (95% CI to 0.665) (n = 267), (95% CI to 0.668) 6 months (n = 239), (95% CI to 0.651) (n = 226), (95% CI to 0.660) 12 months (n = 209), (95% CI to 0.649) (n = 198), (95% CI to 0.687) Change (over 12-month period) (n = 198), (95% CI to 0.161) (n = 190), (95% CI to 0.170) (95% CI to 0.255) QALY c (n = 191), (95% CI to 0.649) (n = 195), (95% CI to 0.647) (95% CI to 0.246) a The reported costs do not exacty sum to the overa cost as the n differs across cost categories; simiary the n differs across outcome measures and time points. b This incudes n = 11 intervention participants, and n = 11 contro participants, who died and were assigned a utiity score of 0. c This incudes n = 9 intervention participants, and n = 12 contro participants, who died and were assigned a utiity score of NIHR Journas Library

73 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Overa and incrementa costs From the perspective of the NHS and PSS, the mean cost (per participant) was estimated to be approximatey 2500 higher for the intervention group than the contro group, and mean overa costs were estimated to be approximatey 4500 higher. The confidence intervas in Tabe 20 do show, however, the arge variations in reation to these figures. It can aso be seen that the mean cost of the intervention is sma in reation to other NHS and PSS costs incurred by this popuation group. Outcomes The mean baseine 6- and 12-month EQ-5D scores for both groups are shown in Tabe 20. It can be seen that in the intervention arm, compared with baseine, the mean EQ-5D scores were ower by at 12 months. Conversey, the mean EQ-5D score for the contro group improved by over the same period. Based on those who had compete EQ-5D data at baseine, 6 and 12 months, the mean QALY gain was for the intervention group and for the contro group. The baseine score, however, was sighty higher for the contro arm (we adjust for this in subsequent anayses). With regard to the SF-6D, both groups had sighty ower mean scores at the 12-month foow-up point than at baseine (see Tabe 20). The mean QALY gains were aso simiar in both groups. Base-case anaysis For those who had both compete cost and QALY data (based on the EQ-5D), after adjusting for covariates, the mean incrementa cost (tota NHS and PSS cost) was estimated to be (95% CI to ), with an incrementa QALY gain of (95% CI to 0.007) (see Tabe 21 for detais of the numbers incuded in the anaysis). An ICER was not cacuated for this group, as the intervention was, on average, both more expensive and ess effective. With regard to the SF-6D, the incrementa cost was (95% CI to ), with an incrementa effect of (95% CI to 0.006). The intervention was thereby estimated to be dominated by the contro group. The associated CEACs are shown in Figures 5 and 6, for the EQ-5D and SF-6D, respectivey. The probabiity that the intervention was cost-effective was < 20% at a cost-effectiveness threshods. Sensitivity anaysis The resuts of each of the sensitivity anayses are presented in Tabe 21. Within a these anayses it can be seen that the 95% CI surrounding the INB is never whoy positive. Thus, in ine with the base-case anaysis, we are unabe to concude that the intervention is significanty (p < 0.05) more cost-effective at a λ of 20,000 per QALY. Indeed, the INB was more commony negative and there was no suggestion that the intervention was more cost-effective for those who received six or more intervention visits. One additiona point to note is that athough there is some consistency in these resuts (at a λ of 20,000 per QALY the 95% CI surrounding the INB is never whoy positive), there is some variation in the mean estimates. For exampe, in the base-case anaysis (based on avaiabe data), compared with contro, the intervention is estimated to be (on average) more costy and ess effective. Conversey, when we ook at the resuts based on MI, the intervention is estimated to be (on average) ess costy and more effective. As such, at a λ of 20,000 per QALY, there is wide variation in the probabiity of the intervention being cost-effective (see Tabe 21). Summary In the base (compete)-case anaysis, the mean incrementa cost of the intervention (tota NHS and PSS costs) was estimated to be (95% CI to ), with an incrementa QALY gain of (95% CI to 0.007), according to the EQ-5D. This suggests that the intervention was not cost-effective. The sensitivity anayses tended to support this concusion as, at a cost-effectiveness threshod of 20,000 per QALY, the CIs around the mean INB were never whoy positive. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 45

74 ECONOMIC EVALUATION: METHODS AND RESULTS Probabiity of being cost-effective Contro group Intervention group Cost-effectiveness threshod ( 000/QALY) FIGURE 5 Cost-effectiveness acceptabiity curve for the intervention (green ine) and contro group (back ine) (base case for EQ-5D data). Probabiity of being cost-effective Contro group Intervention group Cost-effectiveness threshod ( 000/QALY) FIGURE 6 Cost-effectiveness acceptabiity curve for the intervention (green ine) and contro group (back ine) (base case for SF-6D data). 46 NIHR Journas Library

75 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 TABLE 21 Base-case and sensitivity anayses Anaysis Outcome measure Incrementa cost ( ) QALY gain INB ( ) Base-case anaysis (tota NHS and PSS costs) EQ-5D (n = 151; 139) a (95% CI to ) a (95% CI to 0.013) (6.5%), a (95% CI to ) (95% CI to ) (95% CI to 0.007) (5.2%), (95% CI to 50.26) SF-6D (n = 137; 139) a (95% CI to ) a (95% CI to 0.005) (20.6%), a (95% CI to ) (95% CI to ) (95% CI to 0.006) (18.1%), (95% CI to ) Received six intervention visits (tota NHS and PSS costs) EQ-5D (n = 91; 139) (95% CI to ) (95% CI to 0.003) (4.8%), (95% CI to ) SF-6D (n = 85; 139) 1937 (95% CI to ) (95% CI to 0.004) (22.9%), (95% CI to ) MI EQ-5D (n = 287; 281) (95% CI to ) (95% CI to 0.001) (40.7%), (95% CI to ) SF-6D (n = 287; 281) (95% CI to ) (95% CI to 0.013) (61.6%), (95% CI to ) Winsorising EQ-5D (n = 151; 139) (95% CI 6,35.90 to ) (95% CI to 0.008) (1.3%), (95% CI to ) SF-6D (n = 137; 139) (95% CI to ) (95% CI to 0.005) (9.1%), (95% CI to ) Overa costs EQ-5D (n = 133; 129) (95% CI to ) (95% CI to 0.016) (2.8%), (95% CI to ) SF-6D (n = 122; 126) (95% CI to ) (95% CI to 0.009) (8.8%), (95% CI to ) INB, incrementa net benefit at the threshod of 20,000 per QALY. The probabiity of being cost-effective is aso estimated at this vaue. a Unadjusted costs. Adjusted resuts are presented aong with unadjusted a resuts for the base case; n refers to the number of participants who had compete data for both costs and outcomes. (This is dependent on the perspective taken and outcome measure used in the anaysis. Numbers for the intervention group are dispayed first.) Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 47

76

77 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 5 Quaitative study: methods and resuts Introduction This chapter presents the aims and methods of the quaitative study, which expores the meaning of confidence after stroke, as described by the intervention participants. The principes of interpretive phenomenoogy anaysis (IPA) were appied in both the coection and anaysis phases of the study, resuting in broadening our understanding of the meaning of confidence after having a stroke. Aims The aims of the quaitative study were to answer the foowing questions: 1. How does having a stroke affect sef-confidence? 2. How do stroke survivors describe their experiences of regaining confidence after stroke? 3. What do stroke survivors identify as barriers to regaining confidence? Samping strategy and recruitment A participants were seected from the getting out of the house tria as IPA recommends that participants shoud be experts in the phenomenon being studied. Lack of confidence was cited in the singe centre 24 as a reason for not getting out of the house, so it was fet that participants in the muticentre study were ikey to have a perspective of individua confidence. To achieve maximum diversity a purposefuy seected sampe, ranging from no symptoms at a after stroke to severe disabiity, were recruited. The Modified Rankin Scae 63,64 was used to identify this range, which was considered refective of the stroke popuation. Ten potentia participants were contacted by the quaitative researcher and invited to consent. A 10 participants agreed to be interviewed and a wiingness to te their stroke story was observed. Participant sampe characteristics The sampe was drawn from two tria sites for geographica convenience. Five men and five women were incuded in the study. One presented with no symptoms after stroke, three with sight disabiity, four with moderate disabiity, and two with moderatey severe disabiity. Five ived aone, three ived with their spouses, and the remaining two ived with their spouses and chidren. Nine participants were of white British origin and one was of back Caribbean origin. Data coection: interviewing approach Semistructured, in-depth interviews that sought the perceptions of stroke survivors were the favoured method of data coection. It is advocated that in-depth interviews are the best ways to attain rich, comprehensive, first-person experience. 65 Interviews were conducted in participants own homes, which aimed to provide a safe and famiiar context. A refective journa to capture feeings and thoughts throughout the interview process was maintained. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 49

78 QUALITATIVE STUDY: METHODS AND RESULTS An interview guide was deveoped for two piot interviews, and improved for next eight interviews. This proved a usefu too for steering the interview. However, the first interview question eicited the majority of the data. Te me about your stroke, what happened when you had your stroke? The natura responses and fow of the interview as a resut of this question, enabed participants to articuate meanings around osing confidence and regaining confidence, through their experiences, in their own words. Ten competed interviews were transcribed verbatim, using digita dictation transcription software. Transcripts refected words, aughter, significant pauses and siences. Data anaysis The framework for the data anaysis was seected on the basis of the methodoogy. A six-stage process foowing IPA principes was chosen in favour of using a computer software package, such as, NVivo. IPA is interested in understanding the content of the data, rather than measuring frequency of words, or imposing a more tapered approach to the data. Therefore, principes advocated by an IPA expert 65 were appied, and six-stage anaysis process foowed: 1. Reading and re-reading the data, making initia notes in margins. 2. Initia noting, taking one participant s data at time, appying no rues and, therefore, making this stage as exporatory as possibe. 3. Key data were then captured during the next anaytica stage and the deveopment of emergent themes progressed. 4. Stage four invoved ooking for patterns and connections across emergent themes and interpretation within these data was appied. 5. Stage five invoved a repetition of stages 1 4 across a participants. 6. The fina stage was a process that aimed to capture the essence of each individua story and aso seek out differences and simiarities within the data. This process was audited by an experienced researcher in order to verify that a stages had been foowed, adding to the trustworthiness of the data findings. Presentation of the findings The findings section foows with six key emergent themes that describe the phenomenon of confidence after a stroke. Athough each participant provided a unique account, surprisingy, the emergent themes identified shared many concepts, despite difference in stroke impairments and persona context. The themes represent the essence of confidence for the participants. As there is an argument for the definition of confidence being different dependent upon who is asked, 66 embedded in the interview data were a direct question about what confidence meant to participants in this study, conducted on the basis of providing some introductory data. Box 1 iustrates these findings: 50 NIHR Journas Library

79 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 BOX 1 Exampes of what confidence means What does confidence mean to you? Bob Going out and doing what you want to do Ryan 9.13 Confidence to me, in reation to this, is doing something you want to do, when you want to do it. Mick 26.5 It s some way of going back to norma ife. Being confident to eave the house, waking again, trying to tak erm... Ted 5.14 Phewww God [quiety] [ong pause] I m thinking. Difficut one and I can t reay answer the question. June Confidence [ong pause]... Can you say that again for me pease? Aison I don t know, I ve never got it back competey. Scared to go out because I thought I was going to make a foo of mysef... Leon 16.2 I ose a ot since my strokes. Thing is, confidence is something I beieve in, but you know, ike my waking, for instance, how can I expain it? I wak inside here because the door is there. I hod on to there, right? But if I get outside, there s nothing to hod on to... To three of the participants, confidence is about doing, choice and engaging in everyday activity. Two participants strugged with its definition in this direct context; however, were abe to define the impact of confidence esewhere in the data. A further two immediatey describe osing their confidence and to these participants, confidence has a negative connotation and reates to fear and safety. Further anaysis and interpretation is embedded within the themes that foow. Robbed of ife The notion that having a stroke questions who you are was articuated by many of the participants. Ski oss, decreased competency and ack of engagement in activities were described as contributing to a genera feeing of being a esser person, and uncertainty as to how competent one fees after having a stroke. June iustrates a perceived ink between activity and identity, associating what she does, to who she is: Can t wak far, can t pay badminton. I m just a totay different person. June 11.9 The data indicated that the participants sense of competence after stroke was chaenged. Not feeing abe to engage in previous, often famiiar, skis or roes, such as washing the pots or waking the dog [aughter] (Ryan 9.17) contributed to oss of confidence. However, embedded in the data, successfuy regaining skis contributed to increased competence and sef-beiefs, resuting in a regaining of confidence in their abiities. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 51

80 QUALITATIVE STUDY: METHODS AND RESULTS Fear of having another stroke Some eement of fear was identified by a participants; however, the impact of fear on osing confidence and aso regaining confidence varied between participants. The first fear that participants described was the fear of having another stroke:... Every time I had a headache, I fee ightheaded, my eg hurts, there is aways that question, erm, maybe [aughs], just kinda maybe. Mick you never know what is going to happen tomorrow.... I am going to have another stroke? So that was on my mind. June 17.6 Ted aert to any changes in my body at a, you know? Anything, because I think now, perhaps this is going to be the big one, you know. Barbara 19.4 Participants described iving with this fear and how it prevented them from participating in everyday activities. Mick experienced a second stroke, and articuates this period, as: cementing the fear in my brain Mick This second event had a huge impact on his recovery and confidence to eave the house, and resuted in a period of avoiding going out. For Mick, confidence has been about graduay overcoming this fear to enabe him to regain confidence, to enabe him to engage in what he chooses. He describes this process as gradua, and experiences good and bad days, suggesting confidence has a tempora component. Simiar meanings evoved from June, indicating that fear is a factor that underpins her participation in going out of her home environment. Ted s fear of having another stroke directy impacted on feeing anxious in crowds, and cites shopping centres as a pace that hed a particuar anxiety. Avoidance was his coping strategy. Fear of going out/socia confidence Fear of going out and being sociay active after a stroke was a major concern voiced throughout the interviews by a of the 10 participants. Some were abe to address their fears eary on in their recovery, whereas others found it more difficut and, years after recovery, have not fuy been abe to achieve this, resuting in an avoidance of going out and a genera sense of a diminished socia ifestye and socia confidence. I don t know whether it was that I was sef-conscious about the fact that I was strugging with my wheechair, or strugging to wak, peope ooking at me or not, but... Ryan Eventuay, as your confidence grows, your bubbe starts to get bigger, erm, the garden, then the street, then the shops, eventuay. Mick NIHR Journas Library

81 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Participants aso attributed positive encouragement from others as an enaber to going out and not avoiding activities that caused fear. Athough participants acknowedged huge fears about going out and socia interactions with others, couped with this fear was an immense desire to overcome these fears. For exampe, Ryan states I had coping strategies pinned a over the wa (15.17). It is cear from the descriptions evoving from participants that confidence improves over time. Nevertheess, for some this is a sow process that may never be fuy resoved. In summary, fear has a huge impact on daiy activities and ives after stroke, often inked with ow sef-efficacy and avoidance of engagement in tasks, roes and events. Team confidence/coective efficacy The infuence of significant others, as touched on within the previous theme, impact on eves of confidence. Athough, not surprisingy, participants described positive infuences as enabers to regaining confidence and conversey, negative infuences act as barriers, the negatives and positives are not aways overt in the context of participant s daiy ives and often require a period of refection. Infuence from a supportive friend, couped with sef-determination, one participant s confidence eves increased to a point that enabed her to pursue ambitions she woud not have considered prior to her stroke: He [Friend] said Let s go to Cyprus. I said I m not going to CYPRUS [increased tone: aughter] you see I ve never fown before, you see, so we few to Cyprus. [aughter] Freya 41.7 Conversey, another participant describes her famiy s hep as being restrictive: They don t et me, we, I can t go out of the back door without someone going with me. Heen Independenty mobie and competent, this participant describes feeings of inadequate opportunities, to fufi her potentia. A message underpinning success for another participant, fundamenta in gaining the confidence to mobiise outside after stroke, was a physiotherapy intervention:... the best thing that happened to me, and the most erm, usefu one to me, was being referred to a physiotherapist. Barbara 11.1 Initiay, unabe to mobiise to the kerb without fear of faing, and apprehensive about physiotherapy intervention, she described factors such as humour, empathy, positive reinforcement and encouragement as the components that ed to a positive treatment outcome. Keen to feedback her perceived success, once her treatment was competed, she describes teephoning the physiotherapist who deivered her treatment: And I said to him, I can wak into town now, I can wak as far as Marks and Spencer s, and he said I can t beieve it, that s wonderfu. But I can, and I can now get as far as the shopping centre. Barbara Increased confidence after stroke appears to have a component that is infuenced by the actions of others, in addition to the actions of sef. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 53

82 QUALITATIVE STUDY: METHODS AND RESULTS Roe confidence Participants perceived that oss of roes had an impact on decreased confidence. Loss of a driving roe was considered hugey important to two participants and both these participants reported being more confident and active once they were abe to resume their driving roe. A gardening roe, for another participant, despite a memory issue since stroke, sti enabes her to participate in an enjoyabe roe: Coudn t remember when they were supposed to fower, if they fowered and a those kind of things, but it didn t bother me. June Confidence in meaningfu roes appeared to make a difference to a participants. Increased confidence was evident when participants were motivated to engage in simiar roes that were important pre-stroke or when they had found repacement roes that generated simiar benefits. It s not I can t, it s I can : ski mastery Participants iustrated many exampes of how reearning skis and becoming successfu in mastering a new ski as being one of the biggest factors in regaining confidence. Some participants were sow to start to regain skis and gave exampes of how they avoided activities that evoked fear and uncertainy. Others described being successfu in a particuar task or ski, which enabed them to beieve they coud increase their range of activities. This correates cosey to the sef-efficacy theory. Participants who described themseves as confident pre-stroke tended to aso describe a higher eve of sef-efficacy. This process is best described by participants: Once I coud get up and take a step, I knew I coud do it, I know it seems daft. Ryan Graduay your confidence buids. The first time is aways kinda nervous [pause] we it is for me now anyway. Mick 31.4 Eventuay, started cooking for mysef, I thought, oh I can manage to do different things instead of using the microwave a the time. Freya 30.2 Once participants began to succeed and achieve by repetitive practice, their motivation and confidence eves improved enabing them to move on to other activities. Practising skis was described as increasing competence and confidence; nonetheess, when practice did not improve ski, participants described becoming frustrated and found that their confidence decreased. Inner strength and confidence The fina theme examines how the phenomenon of confidence is associated with other components of psychoogica distress. For three participants, episodes of depression were prevaent after stroke. Other participants identified periods of ow mood, and most had some experience of anxiety. Low confidence was cited as underpinning their daiy ives throughout these experiences. One participant stated being depressed makes you fee a bit useess (Heen 23.7) and another describes simiar sentiments: Some days I fee hepess, aone (Mick 20.10). Interestingy, a participant describing hersef as having very ow 54 NIHR Journas Library

83 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 sef-esteem and a history of depression prior to her stroke, perceives she has a raised sef-esteem and higher sef-efficacy beiefs as a resut of her stroke journey an inner strength amost that I didn t reaise I had (Freya 35.19). She iustrates this by teing us:... the main barrier is within yoursef, I think. The biggest one is within yoursef, you think I can t do that and you think about it and you think I try. But you don t try very hard because you think you can t do it, you see? So you have to try a bit harder and then you reaise you can do it, you know. So often the biggest barrier is within yoursef. Freya Athough focusing on doing and achieving, a process that increased sef-efficacy beiefs was experienced by some participants; others ceary described it being a strugge to achieve the same: I know in the back of my head there was something teing me you had to try and do these things, because, if you don t June, you just sit and vegetate, and ife is too important for that. June As soon as I see a crowd of peope I start to ose my baance. Go hod me someone hod me [panic in tone of voice]. Aison 15.1 This atter quote suggests that success is about more than ski and abiity. Aison was abe to iustrate how negative thinking can often ead to an unsuccessfu outcome. Reinforcement from others suggests that another participant is not enough on its own: Everybody kept saying Oh, you are doing so we but it wasn t enough for me (Heen 8.2), suggesting that positive reinforcement does not compensate for intrinsic worth. The data signify that participants confidence eves have been affected by having a stroke to varying degrees. In seeking to expore the meaning of confidence after stroke, connections and interreations between themes were identified, and simiarities and differences in the data and between participants were disentanged, confirming that the phenomenon of confidence is mutifaceted and compex. Summary Loss of confidence after stroke is a common experience. The impact of stroke is truy reaised ony when a stroke survivor begins to estabish routines, and continues with his/her ife. Loss of former roes and a ack of competence when trying to engage in previousy famiiar tasks is prevaent in the eary stages of stroke recovery, often inked to questioning sef-identity. Graduay regaining skis and re-estabishing identity appeared to increase confidence over time. Fear was identified in this study as a huge barrier to being confident to do the things participants wanted or needed to do. Avoidance behaviours were evident and described in the data, owering participant s sef-efficacy beiefs, further imiting the opportunity to become more competent and confident in activities and tasks. Socia confidence, fear of socia interactions and stigma, was aso embedded within the study s findings. Gracey et a. 67 argued, if these issues fai to be resoved, a poor psychosocia outcome ong term after stroke is reaised. Strategies that enabe stroke survivors to resove oss of confidence in socia situations shoud be considered a necessary treatment component in stroke rehabiitation. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 55

84 QUALITATIVE STUDY: METHODS AND RESULTS Confidence in team and/or partner was evident in the data. Descriptions of encouragement, patience and positive reinforcement are exampes of factors that heped improve confidence and sef-efficacy beiefs. Prevention of opportunities for independence and choice of activities resuted in oss of confidence, contro and disengagement during recovery. A mode to increase confidence in sport, which uses a mutivaried framework encompassing sef-confidence, roe confidence, partner confidence, cohort confidence, team confidence, coach confidence and organisationa confidence 68 might be suitabe for stroke patients when trying to improve outdoor mobiity. Highighted in the iterature is the concept that physica appearance often impacts on sef-confidence; this was not prevaent in these study findings. 56 NIHR Journas Library

85 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 6 Discussion Key findings Stroke patients can become housebound, miserabe and in poorer heath because they cannot participate in outdoor mobiity. 14 This report presents the findings of a muticentre RCT to evauate an outdoor mobiity rehabiitation intervention for peope with stroke. The intervention under evauation was a compex mix of goa setting, practising outdoor mobiity (which incuded waking outside), psychoogica support, and provision of information and sef-monitoring of outdoor mobiity through daiy trave caendars. These caendars were used as an outcome measure. It was deivered by NHS therapists based in the primary care setting. The aim of the study was to repicate a singe-centre study that found significant improvements in outdoor mobiity participation when stroke patients were given the intervention by one highy skied therapist. It aimed to evauate whether the resuts were generaisabe to NHS therapists in different UK ocations. In addition, heath-reated quaity of ife and cost-effectiveness were measured. The muticentre study repicated the methodoogy used in the singe-site study. It reached its recruitment target on time. There was a high rate of foow-up at both the 6-month foow-up (to measure for an immediate effect of the intervention) and the 12-month foow-up (to assess whether it had a onger-asting effect), and 70% of a trave diaries were returned. Stroke survivors and therapists were keen to take part in the study and athough no forma quaitative study was competed to assess processes, it woud appear from the proportion of patients who competed the trave diaries that getting out of the house is a major issue, even many years after stroke. A tota of 568 participants from 15 UK-based sites were recruited over 18 months. The intervention was deivered, according to the protoco, to 287 participants, by 29 therapists, and was simiar to that deivered in the singe-centre study. Outcomes were coected by posta questionnaire and participants were supported, if needed, in their competion by RAs bind to aocation. The resuts were anaysed using an intention-to-treat anaysis, with the effect of different therapists at different sites being taken into consideration. A sma quaitative study expored how a reduction in confidence might have affected peope in the intervention group. The primary outcome measure used to cacuate the sampe size was heath-reated quaity of ife, as assessed by the Socia Function domain of the SF-36v2. In addition, outcome measures used in the singe-centre study SWOM, trave journeys, activities of daiy iving abiity and psychoogica we-being were competed. A fu economic evauation was undertaken. This chapter wi discuss the intervention, resuts, strengths and imitations of the study. Intervention deivery The intervention in this tria was deveoped over a number of years (2000 8), a treatment manua was produced and the CI taught therapists how to deiver it by a 2-hour training session. The intervention was a mixture of physica, psychoogica and panned preparation to achieve getting out of the house. Therapists were instructed to provide treatments as needed up to a maximum of 12 sessions over a 4-month period. They reviewed and panned their treatments foowing oca protocos. We consider that the intervention was deivered as per protoco to the majority of participants. We make this assertion because the intervention was deivered by occupationa therapists and physiotherapists or by supervised rehabiitation assistants 100% of the time. These therapists had been trained to use the outdoor mobiity manua by the study CI. It may have been that the manua was too prescriptive and did not aow therapists the possibiity of providing more treatment sessions, but, owing to research methodoogy and financia restraints, therapists had to stop treatment after 12 sessions. The fideity of treatment checking recorded that 100% of sampe indicated that the therapists were deivering the correct Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 57

86 DISCUSSION treatment. We have to acknowedge that the fideity of treatment checkist was not a standardised assessment owing to a ack of pubished measurements and that the ast question which gave us this 100% rate had not undergone psychometric testing. The participants received a median of seven rehabiitation sessions with a duration of 3.5 hours. This is simiar to that deivered in the singe-centre study (six sessions, duration 4 hours). However, when the range of the intervention sessions was expored, it was found that approximatey equa numbers (20 had one session, 20 had two sessions..., 20 had 11 sessions) but 60 participants received 12 sessions. It woud appear that some patients may have needed > 12 sessions to achieve their goa. This coud be the reason that neary one-third did not compete their rehabiitation to the satisfaction of the therapist. Studies of upper imb rehabiitation suggest that > 3 hours per day for 30 days is needed to measure an improvement. 69 Athough the number of intervention sessions was restricted to repicate what might be provided by the NHS, the content of the intervention had been deveoped to be person centred and therapists coud be fexibe in how it was deivered. Therapists who provided the intervention and participants provided informa feedback and said they fet it was an appropriate and cinicay reevant treatment technique. The research team have received over 50 requests for the intervention manua and training over the ast 3 years. This indicates a rea need for an evidence-based outdoor mobiity programme. A more recent study deveoped a person-centred rehabiitation intervention by asking 132 participants to prioritise their main goas. The goas stated were mainy reated to active recreation, househod and community management, mobiity and sociaisation, information on stroke and prevention of new strokes, outdoor mobiity and transportation. 70 Effect of the intervention on heath-reated quaity of ife The primary outcome measure of heath-reated quaity of ife found no significant differences between the contro and intervention group at either the short- or ong-term foow-up. The Socia Function domain of the SF-36v2 was used to assess heath-reated quaity of ife in this popuation because the SF-36v2, as a whoe, has been subjected to psychometric testing, used in numerous descriptive and intervention studies, and can be competed in < 10 minutes by sef, proxy, interviewer or teephone. 71 This aows comparisons to be made with other heath-care treatments. It incudes a range of domains, such as oss of roe, participation away from the home and socia incusion, which are areas that therapists hope to improve through the type of rehabiitation under evauation. The Socia Function domain of the SF-36v2 consists of two questions and was competed by most (500) participants out of the origina sampe (568 participants). The two questions were: 1. During the past 4 weeks, to what extent has your physica heath or emotiona probems interfered with your norma socia activities with famiy, friends, neighbours or groups? i. (Pease tick one.) Not at a/sighty/moderatey/quite a bit/extremey 2. During the past 4 weeks how much of the time have your physica heath or emotiona probems interfered with your socia activities (ike visiting friends, reatives, etc.)? i. (Pease tick one.) A of the time/most of the time/some of the time/a itte of the time/none of the time This measure was not used in the singe-centre study and other trias of rehabiitation using this type of measure have found simiar neutra findings Therapists and participants were concerned about the use of this measure, as they fet that the intervention aimed to improve outdoor mobiity participation and the expectation that it woud impact on socia participation was overambitious, especiay as the intervention was imited to 12 sessions over 58 NIHR Journas Library

87 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO months. In addition, they stated that participants wanted to undertake mobiity activities for a number of reasons: getting to the doctors, dentists, shopping and work, which might not be cassified as an improvement in quaity of ife as measured by these two questions. Improvement in outdoor mobiity therefore might not have an impact on heath-reated quaity of ife as measured or it might take onger for the benefits to be reaised. We concude that the resuts were reiaby coected and that the outdoor mobiity intervention deivered in this study did not effectivey improve stroke patient s quaity of ife as measured by the Socia Function domain of the SF-36v2. We suggest, however, that this is not a reason to stop outdoor mobiity rehabiitation beginning offered to stroke patients. We beieve that this targeted intervention has potentia to increase outdoor mobiity participation, which we consider an important part of activities of daiy iving. In addition, we have to consider whether the contro group received a treatment that may have a roe to pay with patients for whom there have been many years since their stroke. We present our reasoning beow. Effect of the intervention on outdoor mobiity participation The main finding of the singe-site study 24 was that the intervention group were significanty more ikey to participate in outdoor mobiity. This positive outcome was considered an important cinica benefit. Stroke patients have compained about becoming housebound, isoated and miserabe. This singe-centre study demonstrated that, with a reativey short period of rehabiitation, patients coud be taught how to increase their outdoor mobiity. This incuded waking, using buses and eectric pavement scooters. The outcome was assessed using two measures: SWOM participation and number of journeys. In the muticentre study, the intervention participants were no more satisfied with their outdoor mobiity participation than the contro participants at both 6 and 12 months but they were more ikey to make journeys than those participants in the contro group. However, this difference in number of journeys was apparent ony when the data were adjusted to take the therapist effect into consideration. It appears that some therapists are abe to produce these significant resuts, whereas others are not. We do not know whether it was the way the intervention was deivered, the characteristics of the participant or the skis of the therapist that produced this significant resut. This outcome was measured using participant-reported journeys, which incuded outdoor waking, by trave diaries coected each month by post. Participants were taught how to use the caendars at the baseine visit by the RA and were encouraged to send them back using pre-paid enveopes. We fee that this was an objective measure as participants had to record each time they eft the house and went somewhere. A participants competed the diaries and, as 70% of a trave diaries were returned, we fee this is a reiabe outcome. The resuts repicate those found in the singe-site study, for which intervention participants took significanty more journeys at both 4 and 10 months. We need further research to expore the therapist and site effects in more detai. We consider that an increase in the number of outdoor journeys is an outcome that patients and heath-care providers wi fee is worthy of investment. It may be that an increase in outdoor mobiity is heping peope get to the doctors or heath centre, they maybe going to work, or heping famiy with chidcare. We know that peope wish to get out of the house just for the sake of it or to get fresh air or to enjoy just moving around. 23 In addition, there is evidence from a systematic review of exercise after stroke, which found improvements in heath-reated quaity of ife in the short term. 75 It is possibe that the outdoor mobiity intervention provided in our study, which was mosty mobiity training with the aim of waking > 100 m, was providing an exercise programme. Satisfaction with outdoor mobiity was measured using a singe question: Do you get out of the house as much as you woud ike? Athough this measure was not designed as a patient-reported outcome measure (PROM) to assess heath-reated quaity of ife, it has many of the eements. It measures how peope perceive the impact of a heath intervention, is pertinent to the intervention being deivered and is easy to compete. Prior to this muticentre study, the measure had been subjected to some reiabiity testing 76 and we have no reason to beieve that the participants in the muticentre study found the question ambiguous. When we compared the groups we found no significant difference in outcome for this measure. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 59

88 DISCUSSION Effect of the contro on outdoor mobiity There was very strong evidence that the contro group improved markedy over time. At baseine, 259 out of 281 (92.2%) participants were dissatisfied with outdoor mobiity but at the 6-month assessment this had reduced to 78% (160/205), a 15% reduction. The corresponding reduction in the intervention group was sighty greater (18%) with 268 out of 287 (93.4%) expressing dissatisfaction with outdoor mobiity at baseine and 171 out of 227 (75.5%) expressing this at the 6-month assessment. Participants in the contro group must have been provided with a treatment that affected a change. It woud be expected that peope 3.5 years after stroke woud remain stabe over the next 6 months. This was a pragmatic tria and the therapists were tod to deiver the baseine contro as they woud in the NHS. This was to repicate that given in the singe-centre study. However, the combination of a face-to-face baseine visit presenting taiored (and participant focused) oca transport and mobiity information and use of daiy trave diaries ed to an effect in both the contro and intervention groups. The baseine visit incuded a discussion about potentia goas, which coud have focused the participant s thoughts towards improving outdoor mobiity. We state previousy that passive provision of eafets and information is ineffective. However, we fee, with the motivated peope who took part, that the contro (targeted provision of eafets and information) has acted in a different way. In addition, it is estabished that providing patients with daiy diaries (e.g. food diaries) can ead to behavioura changes and an effect on the outcome being recorded. They provide a reference point for refection on past and current behaviours, and they aow peope to track changes over time and bring their behaviour to the front of their mind. In essence, they are a sef-monitoring too. Diaries have been used in behaviour modification programmes for a range of probems from weight oss and exercise programmes through to managing incontinence, tinnitus and migraines/headaches. Greenhagh 77 who investigated the feasibiity of using diaries to measure the daiy impact of mutipe scerosis on participants found that by competing a dairy participants became more aware of their symptoms, and this resuted in increased reporting. This impact may be further improved by the participants being aware that each month we woud request the information they had recorded. Effect of the outcome on functiona abiity and psychoogica distress The intervention did not ead to greater eves of activity or reduce participant or carer psychoogica distress. The assessments used to coect the outcomes were we known and standardised measures used in many stroke studies We have to concude that the intervention, as deivered, does not impact on these areas. This resut is in ine with previous studies, which has found that a targeted rehabiitation intervention deivered over approximatey six sessions can improve the specific domain being addressed but does not affect other domains. 82 Was the intervention cost-effective? In the base-case anaysis the mean incrementa cost (tota NHS and PSS cost) was estimated to be (95% CI to ) with an incrementa QALY gain of (95% CI to 0.007). This woud suggest that the intervention was dominated compared with the contro. The sensitivity anayses were broady in ine with the resuts of the base case. Effect of the intervention on fas Fas occurred in a age groups of stroke patients, with a median of three fas per year compared with one in four peope from an aged-matched, non-stroke popuation faing each year. 83 It may be that we need 60 NIHR Journas Library

89 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 to expore specific fas-prevention interventions for peope with stroke. There were no differences in fas between the intervention group and the contro group, providing evidence that the intervention did not cause peope to fa over more often. What did we earn about confidence after stroke? Strongy emerging themes indicated that barriers to high eves of confidence incuded oss of former sef, negative reinforcement from others incuding heath professionas, oss of confidence in roes and fear of everyday activities. Confidence eves were found to change over time, with participants having good and bad days, but, generay, the participants spoke of improved confidence as time since stroke increased. The factors that heped stroke patients regain their confidence were positive reinforcement from others, successfu ski mastery and sef-defined positive changes in roes. The study demonstrated that stroke survivors in this study were experts of the confidence phenomenon by identifying a variety of factors that have both enabed them to achieve and succeed, resuting in increased confidence eves, and aso identifying barriers that prevent them from regaining confidence. These findings suggest there is a need for therapists to understand the impact that ow confidence might have on outdoor mobiity and to aim to provide interventions to overcome this barrier. Adverse events We had a sma number of adverse events [20 from 17 (17/287; 5.9%) participants] within the intervention group during the deivery of the intervention. There was no evidence to suggest that the intervention ed to an increase in fas that required the assistance of a heath-care professiona. Methodoogica issues Comparison with other studies This Getting out of the House Study evauated a compex, but targeted, outdoor mobiity intervention for community-based peope who had experienced a stroke many years previousy. Owing to the number of unique methodoogica features there is ony one study to which we can compare the resuts. This was the singe-centre study on which this muticentre study was based. 24 We aimed to repicate the recruitment and intervention deivery but a major difference between the studies was the time from stroke. In the singe-centre study, participants were recruited months after stroke, whereas in the muticentre study they were recruited months after stroke. By competing a pragmatic muticentre tria and recruiting participants many years after stroke, we may have found a popuation more adapted to their situation and ess ikey to respond to the intervention. This coud have contributed to the neutra effects found in the muticentre study compared with the positive effects found in the singe-centre study. This highights the difficuty of competing a pragmatic study over an expanatory study where incusion criteria are more rigid. It woud appear by comparing the two studies that the popuation most ikey to benefit from the outdoor mobiity training are those 1 year after their stroke. The other major difference between the studies is that in the singe centre ony one therapist deivered the intervention, whereas in the muticentre 29 therapists provided the intervention. The resuts from both studies provide some evidence that outdoor mobiity rehabiitation is more effective when provided by specific therapists. However, we are not sure what characteristics these therapists need to have to eicit the benefit. Again, owing to the muticentre study being pragmatic and using routine NHS therapists, we consider that the resuts are more ikey to be those seen if the intervention was provided to a stroke patients. Generaisabiity of findings The intervention deivered in this study was deveoped through university academics working with NHS and internationa partners. It was found to be feasibe for deivery in the NHS by both occupationa Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 61

90 DISCUSSION therapists and physiotherapists in a parts of the UK and suitabe for patients, both men and women, many years post stroke. The age of the popuation was simiar to the stroke popuation generay, so the intervention coud be deivered to routine patients. The resuts cannot be generaised to patients eary after stroke, who are most ikey receiving Eary Supported Discharge services or Community Rehabiitation. The evidence from the singe-centre study is most appicabe to this popuation and provides evidence that outdoor mobiity participation can be increased by receiving a targeted intervention. Strengths and imitations The main strengths of the study are that it was conducted and reported foowing the CONSORT recommendations. 84 The Nottingham CTU managed the data coection, keeping the group aocation ocked unti the anaysis had been competed. The study was overseen by an independent TSDC who provided rigorous governance checks. The patient and carer outcomes were coected by post to reduce bias and three statisticians independenty checked the anaysis. This study was designed to be as pragmatic as possibe by recruiting participants from a variety of settings and using NHS staff to deiver the intervention in 15 different sites across Scotand, Waes and Engand. The participants were randomised using a secure and independent service and the baseine characteristics of the two groups are baanced, providing evidence that the randomisation process worked. Key areas of strengths and weakness are discussed beow. Economic evauation The study was strengthened by having the economic data coection and evauation being nested within the tria. The anaysis was predefined and competed by a research team at a separate university. A weakness was that a number of assumptions were required in order to estimate some of the cost variabes, for exampe it was uncear (for a few participants) whether they were reporting average times per carer/home hep visit or tota times for the week. That said, the resuts are broady consistent, for exampe for the different cost perspectives, so it does not seem that this has any great effect on the resuts of the cost-effectiveness anaysis. Quaitative study The incusion of a sma quaitative interview study that expored confidence eves in participants who received the intervention provided some indications that confident peope with stroke are very keen to get out of their houses and are often restricted by a reduction in confidence as much as a physica impairments. A major imitation of this study was that a fu quaitative study to expore processes was not undertaken. This may have reinforced the main findings, provided vauabe impementation information and et us understand how the therapist effect was infuencing the trave journeys. In future studies of this kind it is recommended that participants, staff, commissioners and managers are interviewed. Choice of outcome measures The study was strengthened by measuring outcomes at the functiona and socia participation eve, using standardised patient-reported outcomes but weakened by not having avaiabe measures that have been used in peope ate after stroke. The main outcome measure has been discussed above. The rest of the measures have been used successfuy with peope in other stroke rehabiitation trias 8 up to a year after stroke. There are very few outcome measures designed for changes in peope with ong-term neuroogica conditions. The participants woud most ikey have a number of comorbidities that may have affected their outcomes, such as arthritis. Therapists who provided the intervention were concerned that the cinica changes they observed were not measured. They fet that a goa attainment scae might have been a better primary outcome measure as it woud have refected the diverse needs of the patients. Some were aiming to wak 20 m and some were aiming to use the bus. Participant identification and randomisation Originay, we cacuated that seven sites woud be needed to reach the sampe size over the period of the study. It was obvious that recruitment was sower than anticipated and the sites were doubed, and then the recruitment period was engthened by 4 months. This improved the generaisabiity, as the sites were 62 NIHR Journas Library

91 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 more representative of different ocations rura, cities, suburban and participants were recruited in a seasons. Invitations from patients identified from stroke registers ed to a higher proportion of randomised participants than identification from the GP database, with anecdota evidence of invovement of the stroke survivor s origina rehabiitation team eading to enhanced response rates. This aso indicated that more use of stroke rehabiitation services in stroke research woud be beneficia for participant identification and recruitment purposes, as we as these services acting as hosts to future research. Whether there is a correation between participant perception of stroke care from the sender of the invitation and their wiingness to participate is uncear and deserves further investigation. GP approach yieded a poor response rate; however, this is not unusua among other research studies from a range of discipines. The majority of participants who were not randomised did not meet eigibiity criteria, with a sma number wishing to enter the study ony if they coud receive the intervention. Overa, there was a high eve of acceptance and enthusiasm for the study and the intervention in particuar. Whether the acceptance and enthusiasm was due to existing services not meeting the participants needs and the opportunity to receive any sort of therapy was appeaing is not cear, but again deserves further investigation. Anecdota evidence from participants showed that they were very appreciative of receiving the intervention and found it of great benefit; however, conversey, there were high eves of dissatisfaction and frustration from participants in the contro group, indicating that maybe there is a strong desire from participants to receive ongoing therapy to address ongoing issues with their recovery from stroke. The eigibiity criteria was very broad, with an aim to incude as many potentia peope as possibe, so peope iving in care homes, incuding those in wheechairs who needed hoisting to transfer. Athough this makes the findings very generaisabe it often caused concerns to the treating therapists, as they fet that 12 sessions was not enough to achieve the goas in severey impaired patients. An additiona concern from the sites, but again with no rea evidence, was that deivery of the intervention to the quaity expected by NHS staff was sometimes difficut despite the great dea of enthusiasm and support from research networks to recruit participants to the study. Coection of foow-up data The dua approach of coection of foow-up questionnaire by either posta approach or RA approach proved very effective and ogisticay was easiy managed. Response rates for trave diaries were surprisingy strong considering that ony a singe posta attempt per month was impemented. These diaries were the source of a vast number of data and, potentiay, for future trias, coud be adapted to coect additiona information, for exampe socia activity and resource use (GP appointments, carers). Issues of unbinding Athough there was a high eve of unbinding either prior to or during 6-month visits by the RA, it was tempered by the fact that the outcomes were sti participant reported and the RAs were not assessing a subjective outcome measure for either primary or secondary outcomes. However, there was evidence from persona communication with RAs that they carified certain questions that participants (in the opinion of the RA) had a tendency to misunderstand or misinterpret. Any impact of this bias wi be minima in the overa interpretation of the data, as this incidence is ikey to be equay distributed across both treatment groups and occurred independenty of participant aocation or RA unbinding status. There was a cear risk of unbinding when coecting outcome data in a compex intervention in stroke patients via RA-assisted questionnaires, so there is a need for prevention methods and aternative approaches to be evauated, whie baancing that with the potentia bias to outcome data from unbinding. Differentia foow-up A common issue with studies that have a participant-perceived differentia benefit, as apparent as in the Getting out of the House Study, is that a proportion of participants wi be incined to fee that continued Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 63

92 DISCUSSION participation in the contro group is of no overa benefit. This benefit is reated to the participant themseves, whereas others fee that it is of no benefit to study data overa. Athough the former may be true, of course, at the outset of the study there was ony preiminary evidence that the intervention was effective and the atter is definitey not the case. Despite efforts to maintain engagement with participants wishing to withdraw, of which the majority were in the contro group, there was sti differentia foow-up between the two groups. However, this does not affect the overa study resuts, as we have 6-month foow-up data from both groups within predefined attrition rates. Attention contro One of the biggest imitations of this study was the input given to the contro participants. The exporatory data impy that contro-group participants were given a treatment that caused an improvement, which was more than that expected from routine care. This is a methodoogica weakness and one that is difficut to overcome in rehabiitation research. We considered that the information given to a participants was simiar to that provided routiney but coaborating therapists disagreed. They woud have preferred in retrospect for contro participants to receive nothing, as they beieve this is what routine care is at present for peope ate after stroke. The addition of the monthy trave diaries seems to have changed the behaviour of participants whom therapists woud have expected to be stabe. As with other rehabiitation research projects it is impossibe to compete a doube-bind tria and contro participants may have been eager to use the trave information and the monthy diaries to change their ives. Therapist and site effect A strength of the anaysis undertaken in this study was that the statistica anaysis pan was predefined and ocked prior to the start of the anaysis (see Appendix 1). The two patient groups were compared using an adjusted modeing method, which took into consideration the fact that many therapists were providing the intervention and that some patients woud be treated by a number of different therapists. This technique has been recommended for studies of this kind, 85,86 as there is the potentia for custering of outcomes (owing to which site a patient is at and which therapist treated them). However, this adjusted modeing method has not been readiy used in rehabiitation studies and is therefore a point of discussion. Before the adjusted modeing method coud be used, a weighting for each therapist providing the intervention needed to be cacuated using an approach that considered the number of therapists treating each patient and the number of times that each patient was treated. This weighting was then appied to the outcome resuts whie the groups were compared using the adjusted method. The adjusted resuts and the unadjusted resuts can be seen in Tabes Ony one of the outcomes was affected by the therapist and site adjustments, and that was the number of journeys made. The resut went from a neutra one to a significant positive one. A concern, though, is that the descriptive data in the 6- and 12-month tabes suggests that the contro group are perhaps sighty more ikey to do more journeys. Another concern for the number of journeys outcome is that it was not possibe to cacuate the sizes of the obviousy infuentia site and therapy effects (owing to the modeing process used). This meant that it was therefore impossibe to determine which of these effects had the most weight on the adjusted resut. However, even if it was possibe to cacuate the site and therapy effects for the number of journeys outcome, as was done for the other outcomes of this study, it woud be very difficut to determine how these effects specificay atered the outcome. For instance, it woud be hard to discover if any of the individua sites or therapists had a strong (or weak) infuence on an outcome. These resuts have been checked by three medica statisticians. This therapist effect can be expained by understanding that the intervention is a combination of the therapist and the techniques prescribed in the manua. It is very reiant on the skis of the therapist and the wiingness of the participant. This type of anaysis and understanding of the therapist effect needs further research. 64 NIHR Journas Library

93 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Chapter 7 Concusions What the study found This study provides robust evidence that getting out of the house is a rea and substantive imitation among stroke survivors even after many years. It demonstrates that rehabiitation interventions that aim to improve outdoor mobiity are appropriate for deivery by NHS therapists and, therefore, the resuts provide definitive and generaisabe answers to cinicay important questions. The intervention has the potentia to increase outdoor mobiity participation in stroke patients as measured by journeys and SWOM but it has to be deivered in a specified way. In the singe-centre study that found simiar resuts, the patients were treated by the same therapist who was experienced and skied in stroke care. This muticentre study has highighted the issues of impementing an intervention that was deveoped for a singe-centre study across the NHS. The intervention had no measureabe effect on quaity of ife (socia function), activities of daiy iving or mood so we cannot recommend that this intervention is used to improve imitations in these areas. Unfortunatey, a major imitation of this study is that contro participants who were expected to remain stabe over the period of the study improved in their SWOM over the first 6 months. We suggest that the information and verba advice pus monthy trave diaries may have acted as a treatment that may have affected the resuts of the primary anaysis. The intervention was more costy than the contro and did not improve quaity of ife compared with contro. Impications for cinica practice This study repicated the resuts found in the singe-centre study. An outdoor mobiity intervention can increase outdoor mobiity participation but this change does not appear to have a significant effect on quaity of ife (socia function) compared with a contro. So is this increase in journeys a cinicay reevant finding that NHS commissioners wi be paying for? To impement this intervention into cinica care, it woud appear from the singe-centre study that the intervention needs to be provided by one experienced and stroke-trained therapist, who woud provide a of the intervention sessions therefore becoming an expert in outdoor mobiity. To impement the contro, outdoor mobiity information and verba advice needs to be personaised and provided face to face by an expert in community services. This needs to be supported by a sef-reporting process that aows patients to et the therapists know how they are progressing. Recommendations for research 1. Further research is needed to find new and innovative techniques and interventions that might hep stroke patients get out of their houses as much as they woud ike. The resuts of our muticentre study indicated that we shoud be ooking at the contro, where we targeted the motivated participants, gave peope decent information, and provided a structured behavioura change programme that incuded charting the behaviour changes. 2. Research is needed to expore aternative outcome measures that are cinicay reevant but that can be used to compare heath-care techniques. In this muticentre study we used a heath-reated quaity of Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 65

94 CONCLUSIONS ife (socia function) measure but the intervention was targeted at outdoor mobiity participation. Stroke patient s vaue increased outdoor mobiity in its own right. Further quaitative research is needed to expore how increased journeys may impact on quaity of ife. 3. We recommend that research is competed to expore the divergence in resuts between singe-site and mutisite studies. This successfuy competed muticentre tria foowed from a positive singe-centre study. However, the resuts were different. We considered that we had repicated the methodoogy used in the singe-centre study as far as possibe when competing a pragmatic study, but we reaise that this was a compex intervention deivered in a compex way to compex patients. 4. Research is needed to deveop stroke-specific fas prevention interventions and psychoogica interventions to improve confidence. In this study fas occurred in a age groups of stroke patients and more often than in an age-matched group without stroke. Fear of faing and reduced confidence have been shown to reduce outdoor mobiity. 5. Data from this study need to be shared with other research groups to aow discussion and comparison of mutipe modeing anaysis. 66 NIHR Journas Library

95 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Acknowedgements Many thanks to a of the PIs and oca network cinica researchers and coordinators. Without their hep this study woud have not been possibe. The study was managed by a TMG (Mat Leighton, Sarah Armstrong, Pip Logan, Tony Avery, Hywe Wiiams, Marion Waker, Simon Leach, Katheen O Nei, John Gadman, Tracey Sach, Shirey Smith, Ossie Newe, James Scott, Kathryn Brown) and a TSDC (Tom Robinson, Norma Fenton, Cath Sackey, Nadina Lincon, Heen McCoughry, Angea Shone, Sarah Armstrong, Pip Logan, Mat Leighton). We woud especiay ike to thank Dr Annie McCuskey from the University of Sydney, Sydney, Austraia, for heping to deveop the intervention manua. A tria manager (Mat Leighton, NCTU) was in charge of the day-to-day running of the tria, a tria administrator (Pat Morris, NCTU) was responsibe for sending and receiving questionnaires, both supported with dedicated time from the CI (Pip Logan). Mrs Gai Arnod heped with tria administration and organisation of the therapist events. The tria sponsor is the University of Nottingham. Fu ethica committee and R&D director approva were granted for each site (09/H0403/55). The 15 sites were: 1. NHS Nottingham City: PI, Pip Logan; RAs Jane Horne and Amy Moody; therapists, Jane Horne, Lorraine Lancaster and Janet Darby. 2. NHS Nottinghamshire County: PI, Pip Logan; RAs Janet Darby and Amy Moody; therapists, Jane Horne, Lorraine Lancaster and Janet Darby. 3. NHS Linconshire/United Linconshire Hospitas NHS Trust: PI, Simon Leach, RA Debbie McRobbie; therapist, Aison Read. 4. Gateshead PCT/Gateshead Heath NHS Foundation Trust: PI Dave Barer; RAs, Eie Morre, Linda Waker and Maria Bokhari; therapists, Katheen O Nei and Charotte Cainan. 5. NHS Lanarkshire: PI, Caire Ritchie; RAs, Derek Esson, Stephen Kirk and Eaine Feey; therapist, Meanie Campbe. 6. NHS North Somerset: PI, Aiie Turton; RA, Sarah Dunn; therapist, Louise Biffin. 7. Woverhampton City PCT: PI and RA, Jane Bisiker; therapists, Michee Corr and Heen Jones. 8. NHS Norfok/Norfok Community Heath and Care NHS Trust: PI, Ingrid Watmough; RA Sue Aen; therapists, Charmaine Chander and Jo Scrivens. 9. East Kent Community Services: PI, David Smithard; RAs Linda Cowie and Laura Brockway; therapist, Brian Macnay. 10. Tower Hamets Primary Care Trust: PI, Tess Baird; RAs, Stephanie Pohman, Caire Pentecost and Seina Gann; therapists, Emma Beton and Suong Nguyen. 11. NHS South East Essex/Southend University Hospita NHS Foundation Trust: PI, Dr Pau Guyer; RAs, Sharon Tysoe and Ajithra Praveen; therapist, Janusz Mendyka. 12. NHS Grampian: PI, Therese Jackson; RAs, Emma Wison and Christine Daas; therapist, Mae Ong. 13. Cwm Taf Loca Heath Board: PI, Janet Ivey; RAs Matthew Wiiams and Caire Nott; therapists, Juie Thomas and Karu Kodi. 14. Cardiff and Vae University Heath Board: PI, Maggie Webster; RAs, Matthew Wiiams and Caire Nott; therapist, Rache Smyth. 15. NHS Bristo: PI, Aiie Turton; RA, Sarah Dunn; therapist, Jo Corr. Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 67

96 ACKNOWLEDGEMENTS Contributions of authors Phiippa A Logan conceived the study, was the grant hoder and CI for the study, and wrote this summary and the origina protoco. Mat P Leighton, based in the NCTU, was responsibe for day-to-day running of the tria and data coection/management, and was responsibe for most revisions of the protoco, incorporating suggestions and comments from Phiippa A Logan, Marion F Waker, John RF Gadman, Tony J Avery, Hywe C Wiiams, Ossie Newe, Sarah Armstrong, Tracey H Sach, Katheen O Nei and Garry R Barton. Marion F Waker, John RF Gadman, Tony J Avery, Hywe C Wiiams, Ossie Newe, Sarah Armstrong, Tracey H Sach, Katheen O Nei, Simon Leach, David Barer, Jane Horne, Janet Darby, Garry R Barton and Nadina B Lincon have contributed to the writing of this summary and heped to write the origina fu protoco. Hywe C Wiiams, when director of the NCTU, heped to design the study and secure funding. In addition, Sarah Armstrong was the study statistician and has particuar responsibiity for the anaysis. Samir Mehta and Lisa J Woodhouse were study statisticians and anaysed the data. In addition, Tracey H Sach and Garry R Barton were the study heath economists and were responsibe for economic anaysis. A authors have read and approved the fina manuscript. Pubication Logan P, Leighton M, Waker M, Armstrong S, Gadman J, Sach T, et a. A muti-centre randomized controed tria of rehabiitation aimed at improving outdoor mobiity for peope after stroke: study protoco for a randomized controed tria. Trias 2012;13: NIHR Journas Library

97 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 References 1. Feigin VL, Lawes CMM, Bennett DA, Anderson CS. Stroke epidemioogy: a review of popuation-based studies of incidence, prevaence, and case-fataity in the ate 20th century. Lancet Neuro 2003;2: Logan P, Gadman J, Radford K. Use of transport by stroke patients. Br J Occup Ther 2001;64: Nationa Audit Office. Reducing Brain Damage: Faster Access to Better Stroke Care. London: May C. The hard work of being i. Chronic In 2006;2: X Langhorne P, Legg L. Evidence behind stroke rehabiitation. J Neuro Neurosurg Psychiatry 2003;74(Supp. 4):iv18 iv Eary Supported Discharge Triaists. Services for reducing duration of hospita care for acute stroke patients. Cochrane Database Syst Rev 2005;2:CD Legg L, Langhorne P, Andersen HE, Corr S, Drummond A, Duncan P, et a. Rehabiitation therapy services for stroke patients iving at home: systematic review of randomised trias. Lancet 2004;363: Waker MF, Leonardi-Bee J, Bath P, Langhorne P, Dewey M, Corr S, et a. An individua patient meta-anaysis of randomised controed trias of community occupationa therapy for stroke patients. Stroke 2004;35: Aziz N, Leonardi-Bee J, Phiips M, Gadman J, Legg L, Waker M. Therapy-based rehabiitation services for patients iving at home more than one year after stroke. Cochrane Database Syst Rev 2008;2:CD Hackett ML, Anderson CS. Frequency, management, and predictors of abnorma mood after stroke: the Auckand Regiona Community Stroke (ARCOS) study, 2002 to Stroke 2006;37: Parker CJ, Gadman JR, Drummond AE. The roe of eisure in stroke rehabiitation. Disabi Rehabi 1997;19: Hestrom K, Nisson L, Fug Meyer AR. Reationship of confidence in task performance with baance and motor function after stroke. Physiother Theory Pract 2001;17: org/ / Department of Heath (DH). Nationa Stroke Strategy. London; URL: Pubicationsandstatistics/Pubications/PubicationsPoicyAndGuidance/DH_ (accessed 22 Apri 2008). 14. Gihooy M, Hamiton K, O Nei M, Gow J, Webster J, Pike F. Transport and Ageing: Extending Quaity of Life for Oder Peope via Pubic and Private Transport. Sheffied: ESRC Growing Oder Programme; Robertson K, Logan P, Ward M, Poard J, Gordon A, Wiiams W, et a. Thinking fas-taking action: a fas prevention too for care homes. Br J Community Nurs 2012;17: Department of Transport. Oder Peope: Their Transport Needs and Requirements. London: Department of Transport, Pubications; Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 69

98 REFERENCES 17. Oxey P, Aexander J. Disabiity and Mobiity in London. A Foow-up to the London Trave Survey. London: Transport Research Laboratory; Rabbitt P, Carmichea A, Jones S, Hoand C. When and Why Oder Drivers Give Up Driving. Cheade: AA Foundation for Road Safety Research; Virgii G, Rubin G. Orientation and mobiity training for aduts with ow vision. Cochrane Database Syst Rev 2003;4:CD Roya Coege of Physicians Intercoegiate Stroke Working Party. Nationa Cinica Guideine for Stroke. London: Roya Coege of Physicians; Brazzei M, Saunders DH, Greig CA, Mead GE. Physica fitness training for stroke patients. Cochrane Database Syst Rev 2011;11:CD CD pub4 22. Forster A, Brown L, Smith J, House A, Knapp P, Wright JJ, et a. Information provision for stroke patients and their caregivers. Cochrane Database Syst Rev 2012;11:CD Logan P, Dyas J, Gadman J. Using an interview study of transport use by peope who have had a stroke to inform rehabiitation. Cin Rehabi 2004;18: cr742oa 24. Logan PA, Gadman JRF, Avery AJ, Waker MF, Groom L, Dyas J. Randomised controed tria of an occupationa therapy intervention to increase outdoor mobiity after stroke. BMJ 2004;329: Brazier J, Roberts J, Deveri M. The estimation of a preference-based measure of heath from the SF-36. J Heath Econ 2002;21: Nouri F, Lincon N. An extended activities of daiy iving scae for stroke patients. Cin Rehabi 1987;1: Gadman JRF, Lincon NB, Adams SA. Use of the Extended ADL Scae with stroke patients. Age Ageing 1993;22: Coen FM, Wade DT, Robb GF, Bradshaw CM. The Rivermead Mobiity Index: a further deveopment of the Rivermead Motor Assessment. Int Disabi Stud 1991;13: Godberg DP, Wiiams P. A Users Guide to the Genera Heath Questionnaire. Windsor: NFER Neson; Brooks R. EuroQo EQ-5D: EuroQo: the current state of pay. Heath Poicy 1996;37: Brazier J, Waters SJ, Nicho JP, Koher B. Using the SF-36 and the Euroqo on an edery popuation. Qua Life Res 1996;5: Wyrwich K, Buinger M, Aaronson N, Hays R, Patrick D, Symonds T, et a. Estimating cinicay significant differences in quaity of ife outcomes. Qua Life Res 2005;14: Britte N, Brown M, Mant J, McManus R, Riddoch J, Sackey C. Short term effects on mobiity, activities of daiy iving and heath-reated quaity of ife of a Conductive Education programme for aduts with mutipe scerosis, Parkinson s disease and stroke. Cin Rehabi 2008;22: Deming WE, editor. Samping new materia. In Sampe Design in Business Research. New York, NY: Wiey & Sons; Great Britain. Menta Capacity Act London: The Stationery Office; NIHR Journas Library

99 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO Medica Research Counci (MRC). A Framework for Deveopment and Evauation of RCTs for compex Interventions to Improve Heath. London: MRC: Logan P, Waker M, Gadman J. Description of an occupationa therapy intervention aimed at improving outdoor mobiity. Br J Occup Ther 2006;69: Sakind NJ, editor. Encycopedia of Research Design. London: SAGE; Rubin DB, editor. Mutipe Imputation for Nonresponse in Surveys. New York, NY: Wiey and Sons; Agenda for Change Impementation Team. The NHS Knowedge and Skis Framework (NHS KSF) and the Deveopment Review Process. London: DH; Briggs A, Cark T, Wostenhome J, Carke P. Missing... presumed at random: cost-anaysis of incompete data. Heath Econ 2003;12: Curtis L. Unit Costs of Heath and Socia Care Canterbury: PSSRU; Department for Transport Statistics. Free Fow Vehice Speeds on Buit-up Roads in Great Britain, Annua from 2006 to London: Department for Transport; Nationa Institute for Heath and Care Exceence (NICE). Guide to the Methods of Technoogy Appraisa. London: NICE; Department of Heath (DH). NHS Reference Costs London: DH; Forder J, Fernandez J-L. Length of Stay in Care Homes. Canterbury: Bupa Care Services; Statistica Buetin Annua Survey of Hours and Earnings. Newport; URL: gov.uk/ons/re/ashe/annua-survey-of-hours-and-earnings/2010-revised-resuts/index.htm (accessed 21 May 2013). 48. Drummond M, Scupher M, Torrance G, O Brien B, Stoddart G. Methods for the Economic Evauation of Heath Care Programmes. 3rd edn. New York, NY: Oxford University Press; Sach TH, Whynes DK. Measuring indirect costs: is there a probem? App Heath Econ Heath Poicy 2003;2: Curtis L. Unit Costs of Heath and Socia Care Canterbury: PSSRU; Brooks R. EuroQo: the current state of pay. Heath Poicy 1996;37: / (96) Doan P. Modeing vauations for EuroQo heath states. Med Care 1997;35: Sach T, Barton GR, Jenkinson C, Doherty M, Avery A, Muir KR. Comparing cost-utiity estimates: does the choice of EQ-5D or SF-6D matter? Med Care 2009;47: Ware JE Jr, Sherbourne CD. The MOS 36-Item Short-Form Heath Survey (SF-36): I. Conceptua Framework and Item Seection. Med Care 1992;30: Kharroubi SA, Brazier JE, Roberts J, O Hagan A. Modeing SF-6D heath state preference data using a nonparametric Bayesian method. J Heath Econ 2007;26: Grieve R, Grishchenko M, Cairns J. SF-6D versus EQ-5D: reasons for differences in utiity scores and impact on reported cost-utiity. Eur J Heath Econ 2009;10: s Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 71

100 REFERENCES 57. Manca A, Hawkins N, Scupher MJ. Estimating mean QALYs in tria-based cost-effectiveness anaysis: the importance of controing for baseine utiity. Heath Econ 2005;14: Briggs AH, O Brien BJ, Backhouse G. Thinking outside the box: recent advances in the anaysis and presentation of uncertainty in cost-effectiveness studies. Annu Rev Pubic Heath 2002;23: Heyse JF, Cook JR, Carides GW, editors. Statistica Considerations in Anaysing Heath Care Resource Utiization and Cost Data. Oxford: Oxford University Press; Stinnett AA, Muahy J. The negative side of cost-effectiveness anaysis. JAMA 1997;277: Barton GR, Briggs AH, Fenwick EA. Optima cost-effectiveness decisions: the roe of the cost-effectiveness acceptabiity curve (CEAC), the cost-effectiveness acceptabiity frontier (CEAF), and the expected vaue of perfection information (EVPI). Vaue Heath 2008;11: Office for Nationa Statistics (ONS). Statistica Buetin: 2010 Annua Survey of Hours and Earnings. Newport: ONS; URL: (accessed 21 May 2013). 63. Rankin J. Cerebra vascuar accidents in patients over the age of 60. Scott Med J 1957;2: Van Swieten JC, Koubstaa PJ, Visser MC, Schouten HJ, van Gijn J. Interobserver agreement for the assessment of handicap in stroke patients. Stroke 1988;19: STR Smith J, Fowers P, Larkin M. Interpretative Phenomenoogica Anaysis: Theory, Method and Research. London: Sage; Henschen K, Newton M. Buiding Confidence in Sport. Miton, QLD: John Wiey & Sons; Gracey F, Pamer S, Maey D, Keohane C, Cooper J, Prince L, et a. Cases iustrating a Y-shaped mode of identity and participation change processes in hoistic rehabiitation of brain injury. Brain Impair 2008;9: Veaey R. Confidence in Sport. Brewer BW, editor. Chichester: Wiey Backwe; Han C, Wang Q, Meng PP, Qi MZ. Effects of intensity of arm training on hemipegic upper extremity motor recovery in stroke patients: a randomized controed tria. Cin Rehabi 2013;27: Lund A, Micheet M, Kjeken I, Wyer TB, Sveen U. Deveopment of a person-centred ifestye intervention for oder aduts foowing a stroke or transient ischaemic attack. Scand J Occup Ther 2012;19: Geyh S, Cieza A, Koerits B, Grimby G. Content comparison of heath reated quaity of ife measures used in stroke based on the internationa cassification of functioning, disabiity and heath (ICF): a systematic review. Qua Life Res 2007;16: s Forster A, Young J, Nixon J, Kara L, Smithard D, Pate A, et a. A custer randomized controed tria of a structured training programme for caregivers of inpatients after stroke (TRACS). Int J Stroke 2012;7: Tiing K, Cosha C, McKevitt C, Daneski K, Wofe C. A famiy support organiser for stroke patients and their carers: a randomised controed tria. Cerebrovasc Dis 2005;20: NIHR Journas Library

101 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO Eis G, Mant J, Langhorne P, Dennis M, Winner S. Stroke iaison workers for stroke patients and carers: an individua patient data meta-anaysis. Cochrane Database Syst Rev 2010;5:CD Chen M, Rimmer J. Effects of exercise on quaity of ife in stroke survivors: a meta-anaysis. Stroke 2011;42: Logan PA. Outdoor Mobiity After Stroke. Nottingham: University of Nottingham; Greenhagh J. An assessment of the feasibiity and utiity of the MS Symptom and Impact Diary (MSSID). Qua Life Res 2005;14: Waker MF, Sunderand A, Fetcher-Smith J, Drummond A, Logan P, Edmans JA, et a. The DRESS tria: a feasibiity randomized controed tria of a neuropsychoogica approach to dressing therapy for stroke inpatients. Cin Rehabi 2012;26: Legg L, Drummond A, Leonardi-Bee J, Gadman JR, Corr S, Donkervoort M, et a. Occupationa therapy for patients with probems in persona activities of daiy iving after stroke: systematic review of randomised trias. BMJ 2007;335: Corr S, Phiips CJ, Waker M. Evauation of a piot service designed to provide support foowing stroke: a randomized cross-over design study. Cin Rehabi 2004;18: Gibertson L, Langhorne P, Waker A, Aen A, Murray G. Domiciary occupationa therapy for stroke patients discharged from hospita: a randomised controed tria. BMJ 2000;320: Legg L, Drummond A, Langhorne P. Outpatient Service Triaists: occupationa therapy for patients with probems in activities of daiy iving after stroke. Cochrane Database Syst Rev 2006;4:CD Sayer AA, Sydda HE, Martin HJ, Dennison EM, Anderson FH, Cooper C. Fas, sarcopenia, and growth in eary Life: Findings from the Hertfordshire Cohort Study. Am J Epidemio 2006;164: Zwarenstein M, Treweek S, Gagnier J, Atman D, Tunis S, Haynes B, et a. Improving the reporting of pragmatic trias: an extension of the CONSORT statement. BMJ 2008;337:a Lee KJ, Thompson SG. Custering by heath professiona in individuay randomised trias. BMJ 2005;330: Waters SJ. Therapist effects in randomised controed trias: what to do about them. J Cin Nurs 2010;19: Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 73

102

103 DOI: /hta18290 HEALTH TECHNOLOGY ASSESSMENT 2014 VOL. 18 NO. 29 Appendix 1 Statistica anaysis pan TOMAS Getting out of the House Study Statistica anaysis pan The foowing peope have reviewed the Fina Statistica Anaysis Pan and are in agreement with the contents Name Roe Signature Date Author Statistica Reviewer Chief investigator Abbreviations Abbreviation AE CI CRF DMC EOT GCP ICF NHS NHS IC ONS OT PI PIS RA Description Adverse Event Chief investigator Case Report Form Data Monitoring Committee End of tria Good Cinica Practice Informed consent form Nationa Heath Service NHS Information Centre Office of Nationa Statistics Occupationa Therapist Principe Investigator at a oca centre Participant information sheet Research assistant Queen s Printer and Controer of HMSO This work was produced by Logan et a. under the terms of a commissioning contract issued by the Secretary of State for Heath. This issue may be freey reproduced for the purposes of private research and study and extracts (or indeed, the fu report) may be incuded in professiona journas provided that suitabe acknowedgement is made and the reproduction is not associated with any form of advertising. Appications for commercia reproduction shoud be addressed to: NIHR Journas Library, Nationa Institute for Heath Research, Evauation, Trias and Studies Coordinating Centre, Apha House, University of Southampton Science Park, Southampton SO16 7NS, UK. 75