A feasible model for prevention of functional decline in older home-dwelling people the GP role. A municipality-randomized intervention trial

Transcription

1 Ó The Author Published by Oxford University Press. All rights reserved. For permissions, please doi: /fampra/cmn094 Family Practice Advance Access published on 12 December 2008 A feasible model for prevention of functional decline in older home-dwelling people the GP role. A municipality-randomized intervention trial M Vass a, K Avlund b,c, V Siersma a and C Hendriksen b Vass M, Avlund K, Siersma V and Hendriksen C. A feasible model for prevention of functional decline in older home-dwelling people the GP role. A municipality-randomized intervention trial. Family Practice 2009; 26: Background. Danish municipalities are required by state law to offer two annual home visits to all non-disabled citizens >75 years. Visits are primarily carried out by district nurses. GPs are rarely directly involved. Objective. To evaluate the effects of offering an educational programme to home visitors and GPs on mortality, functional ability and nursing home admissions among home-dwelling older people. Methods. Design: Municipality pair-matched randomized trial. Setting: Danish primary care. Subject: 2863 home-dwelling 75-year-olds and 1171 home-dwelling 80-year-olds living in 34 municipalities. Intervention: Home visitors received regular education for a period of 3 years. In nine of 17 intervention municipalities, GPs participated in one small group training session during the first year. Main outcome measures: Mortality, functional ability and nursing home admission during 4½ years of follow-up. Results. Intervention was not associated with mortality. Home visitor education was associated with reduction in functional decline among home-dwelling 80-year-olds after the three intervention years in municipalities where GPs accepted and participated in small group-based training. Effects did not persist after the intervention ended. When analyses were restricted to baseline non-disabled persons, intervention was associated with beneficial effects on functional ability after three intervention years among 80-year-olds, regardless of education was given to home visitors alone or to visitors and GPs. Nursing home admission rates were lower among the 80-year-olds living in the intervention municipalities. Conclusion. A brief, practicable interdisciplinary educational programme for primary care professionals postponed functional decline in non-disabled 80-year-old home-dwelling persons. Keywords. Education, municipality intervention, older people, preventive home visits, primary care. Background The ageing demography of many welfare societies highlights the need to prevent disability in later life. GPs must realize that traditional demand-driven care does not always provide optimal management. Furthermore, important health problems of older people remain uncharted or suboptimally treated. Proactive assessment schemes give primary health care providers new opportunities to enhance active life expectancy. The acceptability of individually tailored advice has been tested in different health care systems. 1 The 75 and over health checks built into the 1990 UK contract for general practice failed to enthuse primary Received 11 February 2008; Revised 12 October 2008; Accepted 12 November a Research Unit and Section of General Practice, Department of Public Health, b Section of Social Medicine, Department of Public Health, University of Copenhagen, Copenhagen and c The Danish Aging Research Centre, University of Aarhus, Odense and Copenhagen, Denmark. Correspondence to M Vass, Research Unit and Section of General Practice, Department of Public Health, University of Copenhagen, Øster Farimagsgade 5, DK-1014 Copenhagen K, Denmark; m.vass@dadlnet.dk 56

2 A feasible model for prevention of functional decline 57 care teams and significantly impact the health of older people. 2,3 GPs reported that these initiatives were a waste of time and did not engage in the interdisciplinary possibilities of the scheme. 4 Despite the burgeoning evidence supporting a diverse array of primary care programmes for the prevention of functional decline and disability, significant barriers have precluded implementation of these programmes in most settings. 5,6 In 1996, however, The Danish Ministry of Social Affairs introduced statutory municipally organized preventive home visits among older people. The municipalities were required to offer two annual preventive home visits from the age of 75. The act did not provide specific guidelines on how to perform the visits, but aimed at supporting personnel resources, networking and social support. Ten years after the act came into force, municipalities still need more knowledge about how to best perform the visits. Visits are primarily carried out by district nurses, but other primary care professionals also participate. Visits do not comprise a health check but visitors can refer to the GP if additional medical care is found appropriate. GPs are rarely involved directly although, as from 2006, the GP contract allows GPs to offer home visits to frail older people if a functional assessment and comprehensive medication review are performed. The national scheme has thus given GPs a specifically medical role in the assessment procedure. Danish GPs are contractually organized in 14 counties (as from 2007 in five regions) and have no municipal authority. Formalized interdisciplinary education comprising GPs and municipal employees is only reported sporadically. We considered cooperation between home care and GPs to be the cornerstone for future planning of efficient primary care to older people. Scientific evidence and experience concerning how to implement new evidence-based knowledge into every day practice is known to be a challenge. 5,6 Individual professional learning skills not only must be trained and updated but also implemented to practice in different health and social care cultures and on different sector levels. We therefore designed our study intervention to be feasible in daily clinical practice and wanted to test it in multiple settings. We have previously published the main results evaluating education of preventive home visitors and GPs after 3 and 4½ years using group-based statistical methods. 7 9 The aim of this paper is to evaluate the effects of offering an educational programme to home visitors and GPs on the individualized risk of functional decline and nursing home admission after 4½ years of follow-up. Methods Design To embrace the known variations in content and organization of the national preventive home visitation scheme (i.e. municipal contextual factors influencing feasibility and acceptability), a cluster randomized design was chosen. The study was a 3-year intervention trial (January 1999 to December 2001) with randomization of educational intervention at the municipal level. Sample size and power calculations Functional ability was chosen as the unit of inference. From an earlier Danish epidemiological cohort study, we knew how functional ability declined in the surviving persons from the age of 75 to the age of More than 30% of the population either died or was censored. Only a limited number of persons (10%) improved their functional ability, 50% had unchanged functional ability either being completely independent or were in need of some help. Almost 40% deteriorated during the 5 years from the age of 75 to the age of A 5% level of statistical significance was chosen to document a population-based 10% change in functional ability with 90% power. Longitudinal follow-up had to envisage the variation among individuals and the variation between the municipalities. Two sample sizes therefore had to be calculated. To measure functional change over a 3-year period, we calculated the natural log to odds of having improved functional ability compared with having declined in functional ability among non-deceased subjects. This measure of change (D) was based on McCullagh s model for paired comparisons with categorical data. 12 Variance between municipalities was assumed larger than the variance between subjects. Since D is approximately normal distributed a model with both municipal and individual variation could be calculated. Sensitivity analyses showed that advantageously more than 100 subjects within each municipality should be sampled and that at least 15 municipal pairs (intervention and ) should participate. 13 Setting Primary care in 34 Danish municipalities. GPs needed to have contracts allowing them to collaborate in accordance with the scheme s provisions; this was possible in four counties. 7 Municipalities The recruitment process took place in autumn Municipalities to be included in the study had to offer preventive home visits as required by the law and to offer fair or good rehabilitation. 8 Consent to participation was obtained from 34 out of 50 eligible municipalities in the four counties. The main reasons for declining were to avoid being restricted to the study intervention s ways of performing home visits. No demographic differences were found between these and the remaining 16 municipalities. To avoid skewed representation after randomization

3 58 Family Practice an international journal of municipal structure and organization, a matched design was chosen. Thus, randomization was performed independently of the investigators as paired matching of intra-county municipalities, urban/rural type, size and geriatric services. 7 After randomization, there were no differences on baseline characteristics between the 17 intervention and the 17 municipalities in terms of municipality size, population density, social and health care expenses per 75+ inhabitant, study participation rates, preventive home visitor staffing and collaboration between the home care system and the local GPs. 7,8 None of municipalities discontinued participation in the study. Subjects Altogether, 5788 non-institutionalized citizens living in the 34 municipalities and born in 1918 (80 years old at baseline) or 1923/1924 (75 years old at baseline) were invited to participate in the study. Addresses were drawn from the Civil Registration Office. Written consent was obtained from 4060 persons (participation rate 70.1%). Twenty-two persons died and four were institutionalized before the intervention started (n = 4034); 3132 persons had no mobility disability at baseline while 902 persons reported that they needed help to perform at least one of the following activities: transfer, getting outdoors, walking indoors, walking outdoors in nice and poor weather and walking on stairs. A CON- SORT diagram of the derivation of the study municipalities and population is presented in Figure 1. Home visitor participation and intervention Updated geriatric and gerontological knowledge was discussed interactively at six one-day sessions during the three intervention years. Unexplained tiredness in daily activities was presented as an early sign of disability, which should alert the visitor to search for the cause of such tiredness in the health, mental or FIGURE 1 CONSORT flow chart derivation of study municipalities and participants (n)

4 A feasible model for prevention of functional decline 59 social domains. 14 Any suspicion of a health or medication problem should endeavour to contact the GP. Interdisciplinary follow-up was encouraged. Preestablished rules for referral and management plans covering frequently occurring clinical situations were discussed on the basis of case stories. Two key persons from each of the 17 intervention municipalities were entrusted with the task of promoting the training to all relevant municipality professionals. After the first educational session, 80% of the preventive home visitors stated to systematically assess functional ability at every visit, and after the second session, all participating visitors confirmed in an anonymous questionnaire to use tiredness in daily activities as a significant trigger for further action. After the third educational session, i.e. after 1 year of intervention, all intervention municipalities had mapped their communities with relation to physical activity; 15 of the 17 municipalities delivered all their written material to the research team. With very few exceptions, all municipality key persons received educational intervention at all sessions. Chosen topics were carried out as planned with acknowledged teachers. The municipality employees were not offered any education during the intervention period, but representatives from all municipalities joined a final conference together with all the teachers at the end of the intervention period. Preliminary advice on best preventive home visitor practice was in this way disseminated back to all participating municipalities. GP participation An initial invitation letter sent to all intervention GP group leaders explained the study. GPs were present in 12 of the 17 initial interdisciplinary municipal meetings. In nine municipalities, GPs agreed to and participated in a small group training session during the first study year. The session typically lasted 2 hours and introduced GPs to the future task of acting as key personnel to coordinate the care of an increasing number of older people. In 3 of 17 municipalities, GPs were neither present at the initial interdisciplinary meeting nor accepted the small group-based education. The GP education programme Updated geriatric and gerontological knowledge and skills to provide care for older people were introduced taking the acceptability and feasibility of general practice into account. Easy-to-use practical tools were presented to facilitate the assessment of older people in the complex reality of clinical practice. 15 Besides to discuss specific geriatric problems, ageism was hypothesized to be the primary key to improving and detecting remediable illness and functional conditions. Think twice before you say it s age was the take home message, but the ethics of treating and not over-treating older people was also discussed. 16 Since all visitors were instructed to refer to the GP in case of unexplained tiredness in daily activities, GPs were encouraged to take any encounter related to the home visitation programme seriously. In connection with referrals, the GPs were encouraged to explore the reason for encounter and to incorporate the 5 mnemonic D s (Table 1) into the anamnestic strategy. A comprehensive medication review was recommended. 17 More focus on how older people manage everyday life should enable the use of multidimensional interventions targeting functional abilities rather than diagnoses. Interdisciplinary approaches to common problems focusing on locally applicable solutions were introduced through concrete cases as the starting point for discussions. Individualized counselling on how to continue or initiate physical activity was a mantra, e.g. by detailed mapping of the community s physical activity services ranging from elder sports to specific rehabilitative services. 7,18 The GPs were asked to refer back to the visitors to let them take over motivational followup and ensure that such interventions were actually implemented. 19 Main outcome measures A self-administered questionnaire recorded functional ability at baseline and at 1½, 3 and 4½ years of follow-up using a validated mobility scale included: able to manage all activities without help versus need of help for one or more activities. 11 Complete data on mortality and nursing home were obtained from registers. TABLE 1 The 5 mnemonic Ds to incorporate into usual clinical practice Drugs Comprehensive ongoing review at every contact with focus on compliance/concordance Avoid nonsteriodal anti-inflammatory drugs (NSAID) and all psychotropics and try to avoid unnecessary polypharmacy Delirium Disease dehydration diabetes Exclude urinary tract infection Enough daily liquid intakes? Concentrated urine? Orientation? Dementia Delayed three-word recall Clock drawing Depression Keep in mind in connection with concomitant somatic disease and consider asking through the facade : Are you sad or depressed? Drinking Keep in mind and enquire about consumption

5 60 The main outcome measure was obtained from 2529 of the year-old survivors (98.8%) and from 957 of the year-old survivors (99.3%) at 3-year follow-up and from 2329 of the year-old survivors (98.6%) and from 843 of the year-old survivors (99.3%) at 4½-year follow-up. Statistical analysis All analyses were based on intention to treat, e.g. a person belonged to the intervention group even if the person did not have significant contacts with the health care system. Analyses were stratified by age group since functional decline is significantly different in the two age cohorts, and sample size and power calculations were done for the 75-year-olds. Mortality and nursing home admission were analysed with Cox regression. The differential risks for intervention modes were assessed by a hazard ratio (HR) compared with the municipalities in a multivariate regression additionally including sex, living alone and functional status at baseline. Adjustment for cluster sampling was performed by including information on municipality pairs as a categorical variable. Functional ability was analysed as the probability of not being in need of help at each of the four observation points and was performed using a multivariate logistic regression model with random subject effect to account for the longitudinal intrasubject correlation, death and dropout. 20 Differential probabilities of not needing help were expressed as odds ratios compared with the municipalities at each of the followup points. Other variables included in the model were sex, household and a random effect for municipal pair. SPSS version 15 and SAS PROC GLIMMIX version 9 were used. Family Practice an international journal Results The Intracluster Correlation Coefficient was 0.33 indicating only discrete clustering within municipalities of the outcome. During 4½-year follow-up, year-olds and year-olds died. The overall mortality rate was 5.2% and 8.7% in the two age cohorts, respectively. Table 2 describes the intervention HRs for deaths and admissions to nursing homes. The cumulated mortality among 75-year-olds and 80-year-olds during the 4½-year follow-up was 23% and 39%, respectively, with no differences between intervention and municipalities (Table 2). During 4½-year follow-up, year-olds in intervention municipalities moved into nursing homes compared with year-olds in municipalities. Among 80-year-olds, 38 versus 48 persons moved into nursing homes. In a Cox regression model, nursing home admission rates for the 80-year-olds were lower for participants in intervention municipalities than for participants living in the municipalities (Table 2). When analyses were restricted to non-disabled participants at baseline (n = 3,132), nursing home admission rates were markedly lower among 80-yearold participants in the intervention municipalities (HR: 0.334; 95% confidence interval: , P = 0.037). Intervention was associated with beneficial effects on functional ability after three intervention years among 80-year-olds in municipalities where intervention was given to visitors and GPs (Table 3 and Fig. 2). When analyses were restricted to non-disabled participants at baseline, intervention was associated with beneficial effects on functional ability after three intervention years among 80-year-olds, regardless of TABLE 2 Adjusted intervention HRs estimates for death and nursing home admission during 4½ years 75 years at baseline 80 years at baseline n = 2863 n = 1171 HR a 95% CI P-value HR a 95% CI P-value Mortality b Intervention Only municipal intervention Municipal and GP intervention Admission to nursing home b Intervention Only municipal intervention Municipal and GP intervention CI, confidence interval. a Cox proportional hazard regression Cox model adjusted for sex, functional status and living alone at baseline as well as cluster effect (municipality pairs). b A HR above 1 indicates a higher probability of death and admission to nursing home.

6 A feasible model for prevention of functional decline 61 TABLE 3 OR estimates for having no mobility disability Baseline 1½ years after baseline 3 years after baseline 4½ years after baseline F-test a OR 95% CI P OR 95% CI P OR 95% CI P OR 95% CI P P Unadjusted analyses Both age cohort Intervention versus years at baseline Intervention versus Municipal intervention Muni + GP intervention years at baseline Intervention versus Municipal intervention Muni + GP intervention Adjusted analyses (sex, household status) Both age cohort Intervention versus years at baseline Intervention versus Intervention versus No disability b Municipal intervention Muni + GP intervention Municipal intervention No disability b Muni + GP intervention No disability years at baseline Intervention versus Intervention versus No disability b Municipal intervention Muni + GP intervention Municipal intervention No disability b Muni + GP intervention No disability OR, odds ratio; CI, confidence interval. a F-test for the difference in the effect of the intervention over time adjusted for differences at baseline (the time intervention interaction). b Full mobility cohort: n = 2321 in the young cohort. n = 811 in the old cohort. intervention was given to visitors alone or to visitors and GPs (Table 3). Discussion The main results of this study are that among homedwelling 80-year-olds, education of home visitors and GPs was associated with reduction in functional decline after the three intervention years in municipalities where GPs accepted and participated in small group-based training. Effects did not persist after the intervention ended, but nursing home admission rates for 80-year-olds in intervention municipalities were lower at 4½-year follow-up.

7 62 Family Practice an international journal FIGURE 2 Individual probability of having no mobility disability in 75- and 80-year-olds at baseline and after 1½, 3 and 4½ years of follow-up stratified by intervention groups Explanation of results Intervention was performed as planned and may be assumed adapted into daily clinical practice in terms of enhanced gerontological and geriatric skills, as older people living in the intervention municipalities were referred more frequently to rehabilitation than those living in municipalities and as GP contacts in the intervention municipalities tended to shift from consultations to home visits during the intervention period (data not shown). It is plausible that the study education influenced several factors in the home visitors and GPs, including increased focus on early signs of disability, the multifactorial risk factors of disability in older people, interest in seeking advice from other health professionals, better communication skills, possibility for professional networking, new ideas for health promotion and prevention of disease at the home visits and in general practice and general management improvement through assessment and use of existing interdisciplinary relations. The results indicate that the home visitors in collaboration with GPs were able to transfer knowledge in a way which ultimately may be associated with improved ability to manage without help, but the beneficial effects seem to persist only in the intervention period. Beneficial results were primarily seen among the 80-year-olds, which could be due to this group s potential possibility of having received preventive home visits before the study started and a general cohort effect. Strengths and weaknesses of the study Major strengths of this study include very few structural home visitor staff or organizational differences between intervention and municipalities, 8,21 that the intervention was performed as planned, the feasibility of the intervention owing to the structured study guidelines that were easily implemented in regional education, the high number of participants, the low dropout, well-validated measures of disability 11 and that the educational intervention had a high probability of being cost neutral. 22 A study weakness is that GP participation was not randomized. Self-selection to participate may cause well-established cooperative cultures between the home care system and GPs to accept educational initiatives more willingly than others. Such cultures would be more likely to benefit. 23 Another weakness is the lack of allocation status concealment of the two occupational therapists that reminded and endeavoured participants to answer the questionnaires, but no differences were seen in participation rates, dropout or accepting and receiving the core preventive home visit service between intervention and municipalities. For obvious reasons, the home visitors were not blinded to the participants. The study was mentioned in the invitation letter and in local newspapers in order to obtain a high response rate for the questionnaire surveys. All participants knew that they participated in a study, but not whether they belonged to an intervention or a municipality. It was impossible to avoid communication between home visitors working in intervention and municipalities. During the study period, county meetings took place, where home visitors from both intervention and municipalities exchanged experiences. The final study conference with all participating municipalities also could dilute some of the intervention

8 A feasible model for prevention of functional decline 63 effects in the 4½-year follow-up. All these interventions would tend to underestimate positive effects. The generalizability of the findings must consider the inclusion criteria. All participating municipalities were motivated and had at least fair possibilities for promoting rehabilitation. They all joined a scientific study, claimed political support to act on discovered relevant needs and will to solve identified problems turning up during the visits. Differences in municipality size, population density and geriatric services in different parts of the country surely make the results generalizable to Danish/Nordic health care, but we do not know whether our findings may be generalized to other national systems. The refusal of 30% of the eligible study participants may represent a weakness. Non-participant mortality rates were not different in intervention and municipalities. 7,8,24 Mortality and nursing home admissions were higher among non-participants, but selection participation bias was of no clinical importance since subgroups of non-participants eligible to the intervention did not differ from the participants. 24 Unanswered questions and future research In a recent published systematic review and metaanalysis on how to obtain independent living in older people by complex interventions in primary care, group education and counselling seemed to have the greatest impact. 25 Time is needed to diffuse clinical research into practice. The reinforcement of relevant elements of geriatric primary care must be based on their potential for easy implementation in a busy GP agenda, and learning competences must be built into every day practice and followed up periodically to achieve permanent change. 23 One training session have not have been enough in our study to maintain beneficial effects. Prioritizing older people s functional abilities and improving cooperation with the home care system may influence the management dynamics and GP attitudes towards older people. The potential for primary care to incorporate linkage, follow-up and interdisciplinary cooperation in a qualified and efficient way is real, and GPs should be aware of the pivotal role they play in meeting the challenge comprised by an ageing demography. Acknowledgements We thank all participating municipalities and Eva Jepsen, Lisbeth Villemoes Sørensen and Annette Johannesen for following up on the questionnaires. We are indebted to Christian Cato Holm for data management and development of municipality registration software. None of the above funding sources have had any involvement with study design, data collection, data analysis, interpretation of data, draughting of the paper or the decision to submit for publication. Declaration Funding: The Danish Ministry of Social Affairs; The Danish Medical Research Council; The Research Foundation for General Practice and Primary Care; The Eastern Danish Research Forum; The County Value-Added Tax Foundation; Velux Foundation to The Danish Aging Research Centre. Ethical approval: The study complies with the Declaration of Helsinki and was approved by the relevant Regional Research Ethical Committees. Conflicts of interest: None. References 1 Elkan R, Kendrick D, Dewey M et al. Effectiveness of home based support for older people: systematic review and meta-analysis. Br Med J 2001; 323: Iliffe S, Gould MM, Wallace P. Assessment of older people in the community: lessons from Britain s 75-and-over checks. Rev Clin Gerontol 1999; 9: Fletcher A, Price GM, Ng ESW et al. Population-based multidimensional assessment of older people in UK general practice: a cluster randomised factorial trial. Lancet 2004; 364: Iliffe S, Lenihan P, Wallace P, Drennan V, Blanchard M, Harris A. Applying community-oriented primary care methods in British general practice: a case study. Br J Gen Pract 2002; 52: Gill TM. Education, prevention, and the translation of research into practice. J Am Geriatr Soc 2005; 53: Grol R, Berwick D, Wensing M. On the trail of quality and safety in health care. Br Med J 2008; 336: Vass M, Avlund K, Lauridsen J, Hendriksen C. Feasible model for prevention of functional decline in older people. Municipality- Randomized Controlled Trial. J Am Geriatr Soc 2005; 53: Vass M, Avlund K, Hendriksen C, Andersen CK, Keiding N. Preventive home visits to older people in Denmark. Aging Clin Exp Res 2002; 14: Avlund K, Vass M, Kvist K, Hendriksen C, Keiding N. Educational intervention toward preventive home visitors reduced functional decline in community-living older women. J Clin Epidemiol 2007; 60: Avlund K. Disability in old age. Longitudinal population-based studies of the disablement process. Dan Med Bull 2004; 51: Avlund K, Kreiner S, Schultz-Larsen K. Functional ability scales for the elderly. A validation study. Eur J Public Health 1996; 6: McCullagh P. A logistic model for paired comparisons with ordered categorical data. Biometrika 1977; 64: Esbjerg S. Power-calculation and Cluster-randomization. Internal Report I. Department of Biostatistics, University of Copenhagen, 1997 [in Danish, can be obtained, by request, from the authors]. 14 Avlund K, Damsgaard MT, Sakari-Rantala R, Laukkanen P, Schroll M. Tiredness in daily activities among non-disabled old people as determinant of onset of disability. J Clin Epidemiol 2002; 55:

9 64 Family Practice an international journal 15 Williams IE, Fischer G, Junius U, Sandholtzer H, Jones D, Vass M. An evidence-based approach to assessing older people in primary care. JRCollGenPractOccasPap2002; 82: Mangin D, Sweeney K, Heath I. Preventive health care in elderly people needs rethinking. Br Med J 2007; 335: Vass M, Hendriksen C. Medication for older people aspects of rational therapy from the general practitioner s point of view. Z Gerontol Geriatr 2005; 38: Dinan S, Lenihan P, Tenn T, Iliffe S. Is the promotion of physical activity in vulnerable older people feasible and effective in general practice? Br J Gen Pract 2006; 56: Barnett A, Smith B, Lord SR, Williams M, Baumand A. Community-based group exercise improves balance and reduces falls in at-risk older people: a randomised led trial. Age Ageing 2003; 32: Verbeke G, Molenberghs G. Linear Mixed Models for Longitudinal Data. Berlin: Springer-Verlag, Vass M, Holmberg R, Fiil-Nielsen H, Lauridsen J, Avlund K, Hendriksen C. Preventive home visitation programmes for older people The role of municipality organisation. Eur J Ageing 2007; 4: Andersen CK, Vass M, Lauridsen J, Avlund K. Cost effectiveness of preventive home visits in 75 and 80-year old persons: economic evaluation alongside cluster randomised led study. Eur J Health Econ 2006; 7: Grol R, Grimshaw J. From best evidence to best practice: effective implementation of change in patients care. Lancet 2003; 362: Vass M, Avlund K, Hendriksen C. Baseline and follow-up characteristics of participants and non-participants. Randomised intervention trial on preventive home visits to older people. Scand J Public Health 2007; 35: Beswick AD, Rees K, Dieppe P et al. Complex interventions to improve physical function and maintain independent living in elderly people: a systematic review and meta-analysis. Lancet 2008; 371: