Published by: INSTITUTE FOR EMPLOYMENT STUDIES Mantell Building University of Sussex Campus Brighton BN1 9RF UK

Transcription

1 Evidence review 1 Review of the Effectiveness and Cost Effectiveness of Interventions, Strategies, Programmes and Policies to reduce the number of employees who move from short-term to longterm sickness absence and to help employees on long-term sickness absence return to work Report May 2008

2

3 Evidence review 1 Review of the Effectiveness and Cost Effectiveness of Interventions, Strategies, Programmes and Policies to reduce the number of employees who move from short-term to long-term sickness absence and to help employees on long-term sickness absence return to work Jim Hillage, Dr Jo Rick Hazel Pilgrim, Nick Jagger Dr Christopher Carroll, Andrew Booth School Of Health And Related Research

4 Published by: INSTITUTE FOR EMPLOYMENT STUDIES Mantell Building University of Sussex Campus Brighton BN1 9RF UK Tel (0) Fax + 44 (0) Copyright 2008 Institute for Employment Studies No part of this publication may be reproduced or used in any form by any means graphic, electronic or mechanical including photocopying, recording, taping or information storage or retrieval systems without prior permission in writing from the Institute for Employment Studies. ID No:

5 The Institute for Employment Studies The Institute for Employment Studies is an independent, apolitical, international centre of research and consultancy in public employment policy and organisational human resource issues. IES is a not-for-profit organisation with over 60 multidisciplinary staff plus associates. IES aims to help bring about sustainable improvements in employment policy and human resource management. IES achieves this by increasing the understanding and improving the practice of key decision makers in policy bodies and employing organisations. Institute of Work Psychology, Sheffield University The Institute of Work Psychology is dedicated to conducting applied research in work settings, in both the public and private sectors. The aims of the Institute are to: advance knowledge about the causes of individual, team and organisational effectiveness at work increase understanding of the well-being of people at work advance knowledge about innovation and creativity at work disseminate this knowledge in the scientific community, in the workplace and in the wider public domain design, implement and evaluate methods of promoting effectiveness, innovation and well-being at work. The School of Health and Related Research, Sheffield University ScHARR is a large multi-disciplinary School within the Faculty of Medicine at the University of Sheffield. It employs around 200 staff, primarily behavioural and social scientists, epidemiologists, statisticians, public health specialists and health economists. Its research infrastructure and its multi-disciplinarily approach offer huge potential for collaboration with leading researchers in other fields, whilst its Information Resources Section is able to provide first class support with systematic literature searches. ScHARR staff have considerable expertise in systematic review methodology and have contributed to the development of the latest techniques for identifying, assessing and synthesising non RCT evidence in systematic review methodology. v

6 vi

7 Contents The Institute for Employment Studies Institute of Work Psychology, Sheffield University The School of Health and Related Research, Sheffield University Contents v v v vii Executive Summary Initial title and abstract sifting Full paper screening Included papers Theme 1: Exercised-based interventions for musculoskeletal disorders Theme 2: Psychological-based interventions Theme 3: Multi-disciplinary interventions Theme 4: Patient management programmes Theme 5: Other interventions Conclusions 1 Introduction Background Research objectives Structure of report 3 2 Methodology Identifying potentially relevant studies Selection of studies for inclusion Data extraction and quality appraisal Synthesis and formulation of evidence statements 33 3 Effectiveness Findings Theme 1: Exercised-based interventions for musculoskeletal disorders Theme 2: Psychological-based interventions Theme 3: Multi-disciplinary interventions 77 vii ix xi xii xii xii xiv xvi xx xxi xxi

8 3.4 Theme 4: Patient management programmes Theme 5: Other interventions Evidence tables Applicability of the evidence to the UK populations in the scope Cost Effectiveness Findings Theme 1: Exercise-based interventions for musculoskeletal disorders Theme 2: Psychological-based interventions Theme 3: Multi-disciplinary interventions Discussion Papers included in the review Effectiveness studies Cost effectiveness findings Emerging themes Papers excluded from the review 244 References 246 Appendix 1: Example Search Strategy used for Research Questions 1 to MEDLINE primary study search strategy research Questions 1 to 3 (intervention effectiveness) 252 Methodological filters 253 Appendix 2: Sifting Criteria Used 256 NICE Absence Sift Criteria: Effectiveness primary studies Questions 1 to NICE Absence Sift Criteria: Economics (primary studies and reviews) Questions 1 to NICE Absence Sift Criteria: Effectiveness reviews Questions 1 to 3 and Question Appendix 3: Full Paper Screening Checklists 262 Appendix 4: Excluded Primary Studies by Reason for Exclusion 268 Effectiveness studies: excluded studies by reason 268 Cost effectiveness: Excluded studies by reason 326 Appendix 5: Studies Pending 336 Effectiveness primary studies 336 Cost effectiveness primary studies 337 Appendix 6: Example Completed Effectiveness Data Extraction form and Quality Checklist 338 Appendix 7: Economic Data Extraction and Quality Assessment Form 353 viii

9 Executive Summary The National Institute for Health and Clinical Excellence (NICE) has been asked by the Department of Health to develop guidance for primary care services and employers on the management of long-term sickness and incapacity. The guidance will provide recommendations for good practice that are based on the best available evidence of effectiveness and cost effectiveness. This report is one in a series of reviews of the literature covering interventions, strategies, programmes and policies to reduce the number of employees who take long-term sickness absence and to help those who have been on long-term absence to return to work. As such, the report aims to inform the guidance on managing longterm sickness absence through two linked systematic reviews of the literature on the effectiveness and the cost effectiveness of interventions to reduce the length of long term sickness absence and the risk of employees moving from short-term to long-term sickness absence. Specifically, this review addresses two research questions. Research question 1 is: What work or primary care-based interventions, programmes, policies or strategies are effective and cost-effective in preventing or reducing the number of employees moving from short to long-term sickness absence? This includes activities to prevent or reduce the reoccurrence of short-term sickness absence episodes. Research question 2 focuses on long-term sickness absence: What work or primary care-based interventions, programmes, policies or strategies are effective and cost-effective in helping employees who have been on long-term sickness absence to return to work. A protocol was developed by the research team in collaboration with the NICE project team, which specified the population, interventions and outcomes of interest for the effectiveness and cost effectiveness reviews. The protocol provided the detailed inclusion and exclusion criteria that were applied to the literature retrieved via the online searches specified in the protocol. In addition to these database and website ix

10 searches, experts and the Programme Development Group (PDG) were contacted for additional studies. The citations of and references given in the included papers are also being checked. The process took into account the following included and excluded population groups. Populations covered: all adults over age 16 in full or part-time employment, both paid and unpaid all adults over age 16 who have experienced short-term sickness or who have experienced short-term sickness followed by long-term sickness or have experienced long term sickness. The descriptions may be defined as short-term sickness, long-term sickness; or sickness absence in the research all employers in the public, private and not for profit sectors. Locations to be included: developed countries belonging to the Organisation for Economic Co-operation and Development (OECD) interventions, policies, programmes or strategies delivered in a primary care setting and/or workplace setting and/or planned, designed, delivered, managed or funded in collaboration with primary care providers and/or employers. These interventions, policies, programmes or strategies can be delivered by a number of providers (such as voluntary, private, statutory sectors) and/or in various settings not just workplace or primary care settings as long as they are fully or co-planned, designed, delivered, managed and/or funded in collaboration with employers and primary care settings. Populations excluded were: self-employed individuals pregnant women who have taken sickness absence related to their pregnancy, during the course of their pregnancy unemployed individuals. The locations excluded were: studies set in developing or non OECD countries. The interventions, programmes, policies and strategies included were: any that aim to prevent or reduce the number of employees moving from short to long-term sickness or aim to prevent the re-occurrence of short-term sickness episodes (specific to research question 1) x

11 any that aim to reduce the length of long-term absences and/or support return to work (specific to research question 2. Interventions, programmes, policies and strategies excluded were any that: aim to prevent the first occurrence of sickness absence (primary prevention) target pregnant women exclusively and/or which focus on illnesses associated with pregnancy, during the course of a pregnancy tackle workplace absences which are not reported and/or recorded as sickness absence (for example, maternity leave) are delivered outside the workplace or primary care settings deal solely with the provision of clinical diagnosis, treatment for existing conditions (including pharmacological or therapeutic interventions) and management of conditions associated with short and/or long-term sickness deal solely with the effectiveness of private health insurance schemes and/or claiming statutory or occupational sick pay deal solely with preventing ill-health retirement (ie where recipient has no intention of returning to work). Based on the protocol, searches were undertaken of 19 research and specialist economics data bases and six websites by the Centre for Review and Dissemination (CRD) at York University. Five more, suggested by the PDG, were searched by the NICE project team. Initial title and abstract sifting A total of 18,101 articles were initially identified (15,297 effectiveness and 2,495 cost effectiveness and 309 reviews) as of potential interest to this review. In addition a number of suggestions have been made by experts and PDG members and further additional references have been identified from review articles and included studies and checked against the existing data base for possible inclusion. The titles and abstracts of these articles were all initially sifted against the agreed inclusion and exclusion criteria from the protocol. Papers definitely meeting the criteria were put forward for full paper screening. Those where it was unclear if a study met all the criteria were tagged as get full paper and put forward for full paper screening. Those relevant to this review from the searches for the forthcoming reviews were tagged and screened against appropriate sift criteria labelled as include or get full paper. Otherwise papers were excluded from the review. The list of includes and get full papers was re-screened prior to the papers being ordered. A total of 805 primary study papers from all sources were ordered for full paper screening for both xi

12 elements of the review, 797 have been screened and eight were not received by the agreed cut off date. Full paper screening The full paper screening involved a more thorough check of the studies suitability for inclusion in the review. This screening was undertaken using full paper screening checklist based on the agreed protocol. Given that the decisions were based on the full papers rather than simply the title and abstract, and in some cases only the title, more definitive decisions could be made. Included papers Articles passing the full paper screen were then put though a process of data extraction and quality assessment. Data extraction was performed by one reviewer and checked by another who wrote the summary. Quality assessment was undertaken by two reviewers independently and ratings of quality were then compared. Any differences were settled through discussion. A total of 45 effectiveness articles including 11 cost-effectiveness articles passed the full paper screen and are included in the review. The included papers reported evaluations of the effectiveness and costeffectiveness in reducing the number of employees who move from short-term to long-term sickness absence and helping employees on long-term sickness absence return to work under five broad themes. Theme 1: Exercised-based interventions for musculoskeletal disorders Evidence statement for Theme 1: Effectiveness studies Eight studies (none from the UK) examined the effect of interventions primarily based around physical exercise on sick leave absence related outcomes on employees suffering from musculoskeletal disorders including lower back pain. One RCT study, rated +, found manual therapy to be more effective than exercise therapy at helping long-term sick employees return to work quicker and another RCT study rated + found that patients who received specialist consultations with a physician and a physiotherapist to improve their skills to cope with their condition, returned to work quicker than those treated in primary care. ER1.1: There is evidence from one RCT study (rated + ) involving 49 patients, aged 20 to 60, 47 per cent female, on long-term sick leave (of between eight weeks and six months) with low back pain in Norway that manual therapy (involving two treatments a week for six weeks) is likely to be more effective than exercise therapy at helping employees return to work quicker (Aure et al. 2003). xii

13 ER1.2: One RCT study in Norway (rated + ) found evidence that workers, aged between 18 and 60, on long-term sick leave with lower back pain who receive consultations with a physician (specialising in physical medicine and rehabilitation) and a physiotherapist to improve skills to cope with their condition may be effective at helping workers return to work up to a year after they start sick leave than comparable people who receive were treated in primary care. In the consultation, patients received information, reassurance and encouragement to engage in physical activity as normally as possible and reports were sent to their primary care physician and local national insurance office. However, there was no significant difference between the groups in terms of return to work in the second or third year. Although the study found significant differences in the average number of sick leave days at the 12-month point between the intervention group and the control group, there was no significant difference between the groups in the proportions experiencing further sickness episodes over the three year period. (Molde Hagen et al, 2003). ER1.3: One RCT study (rated - ) found no statistically significant improvement in the speed of return to work among a group of 67 airport workers in Amsterdam, in the Netherlands, on longterm sick leave (for a mean length of six weeks) with low back pain, who engaged in a graded exercise programme compared with a similar group of 67 employees looked after through usual care (Hlobil, et al., 2005). ER1.4: One RCT study (rated - ) in Canada found no difference in the proportion of 271 workers (29 per cent female) with continuing back pain who underwent an active exercise programme (involving 40 treatment days over four weeks) who returned to work from long-term sick leave compared with a control group of 271 similar workers on usual care (Mitchell and Carmen 1994). ER1.5: One RCT study (rated - ) compared the long-term effects of combined manipulative treatment, stabilizing exercise and physician consultation among 204 employees with chronic low back pain in Finland with physician consultation alone and found no significant difference in the annual number of days sick leave taken between the two groups (Niemisto et al., 2005). ER1.6: Another non-randomised control study (rated ) found that a back school had no effect on days lost to sick leave between a group of 70 randomly selected municipal employees with back injuries from Birmingham Alabama, US who attended a back school (for up to an hour a day for six weeks) and 70 comparable employees who did not (Brown et al., 1992). ER1.7: One prospective cohort study (rated ) found that an early active exercise and education programme did not have an effect on the duration of sick leave among 1600 workers off sick with soft-tissue musculoskeletal disorders in Canada, compared with usual care (Sinclair et al, 1997). ER1.8: There is limited evidence from a longitudinal before and after comparison study (rated ) that attendance at a back school programme (for up to six hours over a period of a year) by 200 bus drivers in Holland may be effective at reducing long-term sickness absence (Versloot et al., 1992) xiii

14 Cost effectiveness evidence statement for Theme 1 ER1.45: There is evidence from one Norwegian cost benefit evaluation based on a randomised controlled trial (rated ++ ) that an examination at a primary care spine clinic by physician and physiotherapist and provision of information and individual instruction, as well as advice on how to stay active, is likely to be cost effective compared to primary care treatment in returning employees back to work following sickness absence due to low back pain (Molde Hagen et al, 2003). ER1.46: There is evidence from one Dutch cost benefit evaluation based on a randomised controlled trial (rated + ) that graded activity is likely to be cost effective from the employer s perspective compared to usual care in returning employees back to work following sickness absence due to low back pain (Hlobil et al, 2007). ER1.47: There is evidence from one economic evaluation (rated - ) that back school is cheaper than but not as effective as non-back school interventions (Brown et al., 1992). ER1.48: There is evidence from one Finnish economic evaluation (rated - ) based on a randomised controlled trial that combined manipulation, stabilising exercises, and physician consultation compared to physician consultation alone is likely to be cost effective in reducing the reoccurrence of absence due to low back pain (Niemisto et al., 2005). The authors of the evidence review believe that the results of this economic evaluation mayo have been calculated incorrectly. Given the incremental costs and incremental effects provided in the paper of the intervention in comparison to physician consultation alone, the cost per point improvement in the VAS scale has been calculated by the reviewers as 228 (2007) and the cost per point improvement in the ODI scale has been calculated by the reviewers as 913 (2007). Theme 2: Psychological-based interventions Evidence statement for Theme 2: Effectiveness studies Eleven studies (none from the UK) examined the effect of psychological-based interventions, including cognitive behavioural therapy, on sick leave absence related outcomes for employees on long-term sickness absence or at risk of moving from short-term to long-term sickness absence. Three RCTs and one longitudinal study rated + found a positive effect on long-term sick leave of a CBT-type intervention in some circumstances, while two RCT studies found CBT made no significant difference in long-terms sickness outcomes. ER1.9: One RCT study (rated + ) found that between five and seven sessions of Cognitive Behaviour Therapy (CBT), provided to 76 employed patients (51 per cent female, average age 44) on sick leave for between six and 26 weeks, with fatigue in South East Netherlands had no significant effect on sickness absence outcomes (speed of return to work or total days on sick leave) compared with a similar group of 75 employees over a period of 12 months (Huibers et al., 2004). A follow-up study (rated - ) assessed the effect on sickness absence after four years and found no significant difference between the intervention and control groups (Leone, 2006). xiv

15 ER1.10: One RCT study (rated + ) found no significant differences in the level of sick leave between 29 women on long-term sick leave who took part in a five-week multimodal cognitive behavioural programme (MMCBT) with added psychologist-led group sessions in Sweden compared with a similar group receiving conventional MMCBT (Jensen et al., 1997). ER1.11: An RCT study (rated + ) found a significant decrease in the days on short-term sick leave( ie for between two and six months) for 36 women employees in Sweden who took part in a cognitive behavioural return-to-work programme compared with a group of 36 similar women employees over a period of six months. The average age among the two groups was 46. However there was no significant effect for women on long-term sick leave. (Marhold et al., 2001). ER1.12: One RCT study (rated + ) found a significant difference in the proportion of 45 employees (27 per cent female) in the Netherlands on long-term sick leave, for up to 20 weeks, with low back pain who had returned to work after 12 months following an intervention involving behavioural graded activity and education supplemented by problem-solving therapy (for around three hours a week for 15 weeks) compared with 39 comparable employees who just received behavioural graded activity and education (van den Hout et al., 2003). ER 1.13: One RCT study (rated + ) found a significant positive difference in the proportion of 109 employees (34 per cent female) who returned to work at a Dutch post and telecommunications company three months after at least two weeks sick leave with symptoms of mental distress after undergoing a three-stage cognitive behavioural therapy (CBT) intervention compared with 83 comparable employees who received usual care (van der Klink et al., 2003). ER1.14: One RCT study (rated - ) found no significant differences in the time off sick between a group of 98 employees (approximately 60 per cent female) on long-term sickness absence owing to emotional distress or minor mental disorder in the Netherlands who received an intervention, consisting five, 50 minute sessions, by specially trained social workers to help them understand and clarify the nature of their problems and develop and implement solutions, compared with 96 comparable employees who received usual care (Brouwers et al., 2006). ER1.15: One RCT study (rated - ) found no significant difference in the length of sick leave between a group of 495 eligible patients (61 per cent female, average age 41) in Northern Norway (sick-listed for more than14 days) who received a letter and a questionnaire in the post encouraging them to consider the possibilities of work modifications and adjustments to help them return to work compared with 501 similar employees who received usual care (Fleten and Johnson, 2007). ER1.16: One RCT study (rated - ) found no significant difference in the proportion of 53 employees (76 per cent female, average age 38) in Oslo, Norway who had returned to work from long-term sickness absence for psychological or muscle skeletal disorders after attending a solution-focused group-based intervention (with eight weekly sessions, lasting three to four hours, focussing on coping strategies) compared with 50 comparable employees receiving treatment as usual (Nystuen et al., 2006). ER1.17: In a controlled before and after study (rated + ) significantly more of the 70 male and female (54 per cent of the total) employees with whiplash injuries in Canada who attended a 10- week Progressive Goal Attainment Programme (PGAP) (for an hour a week) in addition to the usual xv

16 physical therapy had returned to work four weeks after the intervention compared with a sample of 70 comparable employees who received physical therapy only (Sullivan et al., 2006). ER1.18: One non-randomised control trial (rated - ) evaluated the effect on the amount of sick leave taken over an 18-month period by 67 employees (72 per cent female, average age 39) in Sweden with chronic non-specific spinal pain who underwent an intensive multimodal cognitive behavioural programme (coupled with training on return to work for their supervisor) for eight hours a day for four weeks and found no significant difference compared with 28 comparable employees drawn from national data (Jensen and Bodin, 1998). Cost effectiveness evidence statement for Theme 2 ER1.49: There is evidence from one economic evaluation (rated + ) that an activating intervention by social workers for patients with minor mental disorders on sick leave is likely to be cost-effective. The cost-benefit of the intervention compared to the comparator is 9 (2007). The cost-utility analysis suggests that the intervention is dominating (is more effective at a lower cost) (Brouwers et al. 2006). Theme 3: Multi-disciplinary interventions Evidence statement for Theme 3: Effectiveness studies There is mixed evidence from 19 studies (including one from the UK) on the effect of multidisciplinary interventions (involving a number of separate elements) on the length of long-term absence and the movement from short-term to long-term absence. Two RCT studies (rated + ) found that a multi-component intervention had a positive effective on return to work and a further two (both rated + ) found that multi- component interventions worked in some cases (one among women and another among patients with extensive problems and a low chance of returning to work). However three further RCT studies, all rated +, found no positive effect on return to work from the multi-level interventions that they examined. Eight of the 11 studies involving some form of workplace intervention found a positive effect, including three RCTs rated +, (and three did not), while among the eight studies examining interventions which did not have a workplace-related component four had positive results (including one + rated RCT) and four did not (including three + rated RCTs). ER1.19: One RCT study (rated + ) of a group of 223 mainly office-based employees (53 per cent female) in Norway on sick leave due to low back pain for a minimum of eight weeks found that a four-week multi-modal cognitive behavioural programme (MMCBT) including physical activity and cognitive behavioural modification had no significant effect on return to work compared with a regime of usual care received by 81 similar employees over a period of ten months (Haldorsen al., 1998). ER1.20: One RCT study (rated + ) found that a behavioural medicine rehabilitation programme and its two constituent components: behaviour-oriented physiotherapy (for four weeks) and cognitive behavioural therapy (for four weeks) was effective compared to treatment as usual in xvi

17 securing faster returns to work among to 60 year old (average 43) employees (55 per cent female) long-term sick-listed for non-specific spinal pain for between one and six months in an unspecified area of Sweden for women, but not for men (Jensen et al. 2005) ER1.21: A prospective randomised control study (rated + ) found that a graded activity programme (including workplace visits, a back school and individual graded exercise for three days a week until return to work) was effective speeding up return to work among 51 car workers (23 per cent female) sick listed for eight weeks with low back pain compared with a similar group of 50 sick-listed employees in Volvo in Goteborg, Sweden (Lindstrom et al. 1992). ER1.22: One RCT study (rated + ) did not find a significant effect on return to work among 23 sick-listed patients (64 per cent female) with non-specific upper extremity musculoskeletal disorders who participated in a multidisciplinary programme (including exercise, education, sports activities, psychological sessions, delivered over 13 days) compared with a control group of 15 similar workers receiving usual care in rehabilitation centres in the Netherlands (Meijer et al. 2006). ER1.23: Three linked studies (rated + ) from an RCT involving 664 employees in Bergen Norway sick-listed for musculoskeletal pain found that a screening tool could be effectively developed to classify patients by their likelihood of returning to work. The studies show that an intensive (five, six hour sessions a week for four weeks) intervention multidisciplinary rehabilitation regime (including cognitive behavioural modification, education, exercise and occasional workplace intervention) can be effective for patients with extensive problems (and a low propensity to return to work); those with a stronger likelihood of return to work benefit just as much from usual care as from a low or high intensity intervention. The studies also show that men and women respond differently to different types of treatment (Skouen JS and Kvale A, 2006). ER1.24: A Dutch RCT study (rated + ), among 196 men and women aged between 18 and 65 who had been on sick leave for between two and six weeks due to lower back pain, found that a multistage return to work programme (involving a workplace assessment and work modifications based on participative ergonomics and counselling the employee about return to work) was effective at getting them back to work sooner than if they had just had usual care. There is also evidence that the workplace intervention was effective in reducing the total number of days taken as sick leave among the study population and that the clinical intervention (in combination with usual care or the workplace intervention) did not have a positive effect, although the clinical intervention was only adhered to by 65 per cent of cases (Steenstra et al., 2006). ER1.25: A Norwegian RCT (rated + ) of 210 patients aged between 20 and 65 on long-term sick leave with low back pain which compared medical exercise therapy (under continuous supervision by the physiotherapist) conventional physiotherapy (a combination of methods such as heat or cold massage, stretching, different forms of electrotherapy, traction, and a few exercises on the treatment table) and self exercise (patients given information and told to walk for one hour three times a week for 12 weeks) found all three interventions could be effective in reducing total reoccurring sickness absence. However no significant difference was found in the proportion of patients who returned to work during the trial period among the three groups (Torstensen et al 1998). xvii

18 ER1.26: There is evidence from an RCT study (rated - ) involving 137 workers (58 per cent female) off sick in Sweden for at least a week with musculoskeletal disorders that an early intervention involving a work rehabilitation interview and a workplace assessment can be effective at significantly reducing the number of days off sick in the subsequent year, although the generalisability of the study to the UK may be limited as the results of the study may have been influenced by the operation of the Swedish sick pay regulations (Arnetz et al., 2003) ER1.27: An RCT study (rated - ) found no significant effect on sickness absence over a period of five years from a multidisciplinary individualised rehabilitation programme lasting until return to work offered to 464 employees (62 per cent female) with non-specific musculo-skeletal pain on sick leave for at least 90 days in Goteborg, Sweden (Lindh et al., 1997). ER1.28: A Canadian RCT study (rated - ), among 104 workers who had been on sick leave for between four and 13 weeks due to lower back pain, found that a multi-stage return to work programme (involving a combination of workplace and clinical and rehabilitative interventions) was effective at speeding up their rate of return to work and in minimising the total number of days taken as sick leave. (Loisel et al., 2003). ER1.29: A pilot RCT study (rated - ) found that an eight-week work rehabilitation programme (including graded exercise, education, group work with a psychologist and workplace visits and lasting three to four hours a day, five days a week for eight weeks) had no significant effect on the ability to work among 33 employees (21 per cent female) on sick leave for at last two months with musculoskeletal disorders in Zurich in Switzerland (Meyer et al., 2005). ER1.30: One RCT study (rated - ) found that a multimodal treatment (including relaxation training, psychological support and manual therapy, provide in ten one hour sessions over two weeks) was effective at securing a return to work for 60 patients (42 per cent female) suffering from whiplash injury who were recruited within two months of sustaining a neck injury in and around Ancona in Italy (Provinciali et al., 1996) ER1.31: An RCT study (rated - ) found no evidence that a workplace intervention such as an ergonomic assessment (lasting an average of two to four hours) or a health intervention (eg physio or psycho therapy for between four and six hours) or a combination of the two was effective at improving the rate of sustained return to work among 2,845 employees (57 per cent female) off sick for between six and 26 weeks with a wide range of conditions in six areas in the UK (Purdon et al., 2006). ER1.32: There is evidence from a case control study (rated + ) conducted among 426 patients (78 per cent female), off work for at least six weeks with low back pain, in seven states in the USA that a functional restoration programme (consisting of education and work simulation exercises) is effective at increasing the proportion of patients who return to work and stay at work compared with patients who were refused entry to the program (by their physician or insurance company) (Burke et al. 1994) ER1.33: There is limited evidence from a controlled before and after study, (rated - ) that a therapeutic return to work intervention which linked graded work exposure with functional restoration therapy for 18 to 65 year old (52 per cent female) suffering from chronic low back pain and off sick for over 90 days in Quebec, Canada, compared with just functional restoration xviii

19 therapy, community services without any rehabilitation intervention or usual care (for patients denied access to the intervention by the local Compensation Board) (Durand and Loisel, 2000) ER1.34: A controlled comparison study (rated ) found that a multi-disciplinary work rehabilitation intervention (including physical conditioning, work practice instruction and rehearsal and stress management) in Rochester, USA was effective in getting 34 patients (62 per cent female) on sick leave for at least three months with chronic work-related musculo-skeletal disorders back to full-time employment (Feuerstein et al., 1993). ER1.35: There is limited evidence to indicate that a six-hour-a-day, four-week education and exercise programme is likely to be effective at improving the speed of return to work for 86 patients (52 per cent female) off work for an average of eight weeks with back pain, based on a matched comparison study (rated - ) in Sweden (Gard et al., 2000) ER1.36: There is insufficient evidence from a non-randomised control study (rated - ) amongst 97 hospital cleaners (92 per cent women) in Sweden to conclude that a broad ranging intervention (including vocational training, fitness activities and rehabilitation techniques) was effective at reducing sick leave (Landstad et al., 2002) ER1.37: There is limited evidence from a pilot before and after study (rated - ) which found that all 41 participants (80 per cent female) with chronic non-specific upper limb musculoskeletal disorders had returned to work after undergoing a multidisciplinary programme (including graded activity and help with developing strategies to reduce stress and return to work sessions, provided over a period of ten days) in an occupational setting in Netherlands (Schakenraad et al., 2004). Cost effectiveness evidence statement for Theme 3 ER1.50: There is evidence from two economic evaluations (1 Norwegian, 1 Swedish, both rated + ) that multidisciplinary treatment is likely to be cost effective in improving return to work and reducing sickness absence for people with low back pain. The net present value of productivity gains is equal to 352,953 (2007) for light and extensive multidisciplinary treatment (results not provided individually within the paper) and the cost-benefit results of behaviour-oriented physiotherapy, cognitive behavioural therapy and the combination of these is 62,294, 98,197 and 154,475 respectively for females. The interventions are not considered cost-effective for males individually; however combined the cost-benefit of behaviour-oriented physiotherapy and CBT for males is 71,024 (Haldorsen et al., 2002; Jensen et al., 2005). ER1.51: One Dutch economic evaluation (rated + ) also concludes that multidisciplinary treatment affects individuals positively, but shows no significant difference in cost-effectiveness on the societal level as compared to usual care (Meijer et al., 2006). ER1.52: There is evidence from one Dutch economic (randomised controlled trial) evaluation (rated + ) that a multi-stage return to work programme (involving usual care plus a workplace assessment and work modifications based on participative ergonomics and counselling the employee about return to work) is likely to be cost effective in reducing the re-occurrence of absence due to low back pain when measured against usual care as outlined by Dutch Occupational Physician guidelines for lower back pain. The cost per return to work day gained is estimated to be 17 and the cost per quality adjusted life year (QALY) gained is estimated to be xix

20 dominating (- 1,294) in comparison to usual care. However, based on the analysis, it is unlikely that physiotherapy based on operant behavioural principles provided following eight weeks of other ineffective treatment in terms of return to work is cost-effective in comparison to the provision of Dutch usual care for the same indication (Steenstra et al., 2006). ER1.53: There is evidence from one Norwegian cost benefit (randomised controlled trial) evaluation (rated + ) that both medical exercise therapy (under continuous supervision by the physiotherapist) and conventional physiotherapy (a combination of methods such as heat or cold massage, stretching, different forms of electrotherapy, traction, and a few exercises on the treatment table) are likely to be cost effective in comparison to self exercise (patients given information and told to walk for one hour three times a week for 12 weeks)in helping people on long-term sick leave return to work (Torstensen et al 1998). ER1.54: There is evidence from one Canadian cost benefit and cost-effectiveness analysis (rated + ) based on a randomised controlled trial that the clinical intervention, the occupational intervention and the Sherbrooke model (a combination of clinical and occupational interventions) is likely to be cost-effective in comparison to standard care for back pain. The analysis suggests that the combination of the clinical and occupational interventions (the Sherbrooke model) is likely to better value for money than the two interventions individually. (Loisel et al. 2002) Theme 4: Patient management programmes Evidence statement for Theme 4: Effectiveness studies Six studies (including one from the UK) looked at interventions which changed the way patients on sick leave were directly or indirectly managed by their physicians or others. One RCT rated + found no positive effects on long-term sickness absence from changing sick leave arrangement and a non-randomised control file (rated + ) found no effect from training general practitioners and occupational physicians to work closer together. However one econometric analysis (rated + ) did find a positive effect from the introduction of a case management approach: ER1.38: One RCT study (rated + ) found no positive effects from interventions to increase the use of active sick leave (ASL) (which encourages absentees to return to work on modified duties on the equivalent of full pay) among patients with low back pain (LBP) in Norway is a study based on 6,179 National Insurance records (50 per cent of which were for females) (Scheel et al. 2002) ER1.39: One RCT study (rated - ) found a positive but not significant effect on 120 health and university employees (67per cent female)in the Netherlands on sick lave with low back pain for more than 10 days having an early appointment with an occupational physician as compared with management and supervision alone (Verbeek et al. 2002) ER1.40: There is evidence from econometric secondary data analysis (rated + ) of survey and administrative data from 1,685 sick-listed (for three to 12 weeks) employees (56 of whom were female) randomly drawn from across Denmark that a case management approach (in which sicklisted employees are interviewed by a person or team who can direct health and occupational xx

21 services to help the interviewee back to work) is effective at helping people return to work. (Hogelund and Holm, 2006). ER1.41: A non-randomised case controlled study (rated + ) in Flanders in the north of Belgium, found no effect from the introduction of better information exchange, in the shape of a standardised communication form between social insurance physicians and occupational physicians, in helping facilitate work resumption among 505 employees, aged between 18 and 50 with sub-acute (more than one month) sickness absence (Mortelmans et al., 2006). ER1.42: A non-randomised control study (rated - ) found no positive effect from a trial involving training General Practitioners (GPs) and Occupational Health Physicians (OHPs) to work together in the treatment of low back pain on the amount of sick leave taken by 56 patients with a new episode of low back pain compared with a similar control group (56 people) in an industrial region in Netherlands (Faber et al., 2005). ER1.43: One before and after study (rated - ) in a council in northern England found a positive effect on facilitating a quicker return to work from a change in procedures to shorten the periods at which employees off sick were referred to occupational health (OH) department (Malcolm et al. 1993). Theme 5: Other interventions Evidence statement for Theme 5: Effectiveness studies ER1.44: There is insufficient evidence from an RCT study (rated - ) to assess the impact of a follow-up body awareness group training intervention on the sickness absence of 19 employees in Norway who had previously received multi-modal treatment having been off sick with chronic widespread pain (Anderson et al., 2007) Conclusions The studies covered by this review are heterogeneous and no two studies measure the same population, intervention and outcome, meaning that analysis of evidence is restricted to descriptive synthesis. There are a number of themes across the studies which would benefit from further investigation, but the evidence presented here is not sufficient to draw conclusions about their importance in determining the success or otherwise of specific interventions. The themes are outlined below: Early interventions or studies which examine populations with a similar, specified length of absence are more likely to report positive results (although this is not a consistent finding across all such studies) than those with later interventions or more mixed populations. xxi

22 On balance, studies of multi-disciplinary approaches included in this review are more likely to report positive results in terms of effectiveness and cost-effectiveness than studies of single modality interventions. Studies with a workplace component (including vocational counselling) were also more likely to report successful outcomes than those that did not include such an element. Additionally one study explicitly measured return to work prognosis at the start of the intervention and this was found to be important in determining the level of intervention required to achieve return to work. This coincides with recent findings in other literature on individual differences and the impact of work characteristics and if confirmed by further research could have considerable implications for the effectiveness of return to work interventions, suggesting that efficacy is enhanced if interventions are designed to take into account individual differences in return to work prognoses. Overall, the majority of studies included in the review considered interventions for return to work (as opposed to preventing the move from short to long term absence ) or preventing re-occurrence of short term absence), and little is available from current evidence base on interventions to tackle short term absence. This perhaps reflects an interest in the higher costs associated with long term absence. It is also noticeable that the majority of studies look at musculo-skeletal conditions. Whilst this remains the single greatest cause of long term absence in the UK, other significant causes such as mild to moderate mental health problems do not appear to have commanded the same research interest.. xxii

23 IES, IWP, ScHARR 1 1 Introduction The National Institute for Health and Clinical Excellence (NICE) has been asked by the Department of Health to develop guidance for primary care services and employers on the management of long-term sickness and incapacity. The guidance will provide recommendations for good practice that are based on the best available evidence of effectiveness and cost effectiveness. The Institute for Employment Studies (IES) and the Institute of Work Psychology (IWP) and School of Health and Related Research (ScHARR) both at Sheffield University were contracted to undertake a series of three effectiveness and economic evidence reviews of primary randomised control or longitudinal studies (covering four research questions) and an economic analysis of the evidence in those studies to support the production of this guidance. This report is the third effectiveness and economic review to be delivered to the Programme Development Group (PDG). The first, discussed in December 2007, examined interventions aimed at moving people from Incapacity Benefit to employment (research question 4). The second, discussed in February 2008, looked at interventions to reduce the number of employees who take long-term sickness absence on a recurring basis (research question 3). This final review covers cover the remaining two research questions: interventions, strategies, programmes and policies to reduce employees moving from short-term to long-term absence or the reoccurrence of short-term absences (research question 1) and to get long-term absentees back to work (research question 2) and was initially discussed at the PDG meeting in April This review has been revised in the light of the comments received from the PDG and any further evidence received through the search and sifting process since the first draft was completed. An economic analysis, including economic modelling, will be presented at the PDG meeting in May This will cover a selection of topics identified in the reviews which have been chosen by the PDG and where there are sufficient data to make modelling feasible. 1

24 2 Interventions to reduce move from short-term to long-term and long term sickness absence 1.1 Background In 2006, UK employees were absent for an average 3.5 per cent of the time they were due to spend working. An estimated 40 million working days are lost each year in Britain due to ill health and injury. Sickness absence costs the British economy an estimated 13 billion each year (CBI 2006), although the quality and accuracy of available data on absence and sickness absence is variable (Barham and Leonard 2002; Barham and Begum 2005). One-fifth of absences are classified as long term (ie four weeks/20 working days or longer) and on average, in 2006, one per cent of the UK workforce was absent from work due to long-term sickness. It is estimated that 12 per cent of employees on longterm leave are covered by the Disability Discrimination Act. Back pain, musculoskeletal injuries, acute medical conditions, mental ill health and stress are the most common causes. In the public sector, mental ill health and stress were identified as the main causes of long-term sickness absence for non-manual workers; musculo-skeletal injuries and back pain most affected manual workers (CIPD 2006). It is against this background that the Department of Health has asked NICE to provide guidance to primary care services and employers on the management of long term sickness absence. The guidance is intended to be used by professionals and managers who have public health as part of their remit working in the NHS, local authorities and the wider public, private, voluntary and community sectors. 1.1 Research objectives This review addresses the following two research questions which are referred to as research questions 1 and 2 throughout the report. Research question 1 is: What work or primary care-based interventions, programmes, policies or strategies are effective and cost-effective in preventing or reducing the number of employees moving from short to long-term sickness absence? This includes activities to prevent or reduce the reoccurrence of short-term sickness absence episodes. Research question 2 focuses on long-term sickness absence: What work or primary care-based interventions, programmes, policies or strategies are effective and cost-effective in helping employees who have been on long-term sickness absence to return to work. The following secondary research questions were developed to interrogate the data further (data permitting): What is the frequency, content, length and duration of an effective intervention, programme, policy or strategy? 2