Search tips
Search criteria 


Logo of blackwellopenThis ArticleFor AuthorsLearn MoreSubmit
Arthritis and Rheumatism
Arthritis Rheum. 2007 October 15; 57(7): 1220–1229.
PMCID: PMC2675012

Economic evaluation of a rehabilitation program integrating exercise, self-management, and active coping strategies for chronic knee pain



To conduct an economic evaluation of the Enabling Self-Management and Coping with Arthritic Knee Pain through Exercise (ESCAPE-knee pain) program.


Alongside a clinical trial, we estimated the costs of usual primary care and participation in ESCAPE-knee pain delivered to individuals (Indiv-rehab) or groups of 8 participants (Grp-rehab). Information on resource use and informal care received was collected during face-to-face interviews. Cost-effectiveness and cost-utility were assessed from between-group differences in costs, function (primary clinical outcome), and quality-adjusted life years (QALYs). Cost-effectiveness acceptability curves were constructed to represent uncertainty around cost-effectiveness.


Rehabilitation (regardless of whether Indiv-rehab or Grp-rehab) cost £224 (95% confidence interval [95% CI] £184, £262) more per person than usual primary care. The probability of rehabilitation being more cost-effective than usual primary care was 90% if decision makers were willing to pay £1,900 for improvements in functioning. Indiv-rehab cost £314/person and Grp-rehab £125/person. Indiv-rehab cost £189 (95% CI £168, £208) more per person than Grp-rehab. The probability of Indiv-rehab being more cost-effective than Grp-rehab increased as willingness to pay (WTP) increased, reaching 50% probability at WTP £5,500. The lack of differences in QALYs across the arms led to lower probabilities of cost-effectiveness based on this outcome.


Provision of ESCAPE-knee pain had small cost implications, but it was more likely to be cost-effective in improving function than usual primary care. Group rehabilitation reduces costs without compromising clinical effectiveness, increasing probability of cost-effectiveness.

Keywords: Economic evaluation, Rehabilitation, Knee pain


Chronic knee pain is a major cause of disability (16). Health and social care expenditure (3,79) on this condition will increase as more people live longer (10), and may exceed predictions based on actual information as patterns of consultation change (11) and obesity and sedentary lifestyles increase (12,13).

Knee pain is usually managed with analgesia, despite concerns about safety (9,1416), efficacy, and direct and indirect costs of medication (17). Functional impairment resulting from knee pain is frequently overlooked despite the disability this causes and its role in the etiology of chronic ill health, e.g., diabetes, cardiovascular disease, obesity, depression (18). Functioning can be improved through self-management interventions (19,20) and exercise (2123). However, complex, expensive rehabilitation programs benefit few individuals. Effective and affordable interventions are needed that can be administered to large numbers of people.

We have demonstrated that a rehabilitation program integrating exercise, patient education, and self-management strategies (Enabling Self-Management and Coping with Arthritic Knee Pain Through Exercise [ESCAPE-knee pain]) improves functioning (24,25). However, when planning health care for local communities, commissioners also need cost and cost-effectiveness information. This report details the economic evaluation of ESCAPE-knee pain.


Design, setting, and participants

A pragmatic, cluster randomized controlled trial was carried out and analyzed in accordance with the prespecified protocol. Inclusion and exclusion criteria, randomization, and clinical outcomes are described elsewhere (25). Briefly, participants had to be age 50 years or older and have attended their primary care practice reporting mild, moderate, or severe knee pain of more than 6 months' duration. Exclusion criteria were as follows: lower limb arthroplasty, physiotherapy for knee pain during the preceding 12 months, intraarticular injections during the preceding 6 months, unstable medical conditions, inability/unwillingness to exercise, wheelchair bound, and inability to understand English. The study was approved by the local ethics committee and participants gave informed consent.


The units of randomization were primary care practices, which were allocated to 1) usual primary care, 2) usual primary care plus the rehabilitation program administered to individual participants (Indiv-rehab), or 3) usual primary care plus the rehabilitation program delivered to groups of 8 participants (Grp-rehab). The content and format of the rehabilitation program were similar for Indiv-rehab and Grp-rehab, and consisted of 12 supervised sessions (twice weekly for 6 weeks). For 10–15 minutes of each session, the supervising physiotherapist facilitated a discussion on a specific topic, advising and suggesting simple coping strategies. Then for 30–45 minutes, participants performed simple exercises to improve functioning; these exercises were personalized and progressed according to each participant's ability. (For details of the program see We hypothesized that rehabilitation (disregarding whether costs were incurred during Indiv-rehab or Grp-rehab) would cost more than usual primary care, and that Indiv-rehab would cost more than Grp-rehab.


The a priori primary outcome was a clinically meaningful improvement (15% from baseline level) in self-reported functioning, assessed using the physical function subscore of the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC-func) (26), 6 months after completion of rehabilitation (7.5 months after baseline assessment when 6-week intervention or usual primary care control period was included). Quality of life was assessed using the EuroQoL (EQ-5D) (27). Quality-adjusted life years (QALYs) were calculated by applying UK general population utility weights (28) to the 5-dimensional health state descriptions for each participant.

Resource use

Management of all participants' knee problems continued at the primary care physician's discretion, and was documented at assessment. Data on resource use were collected retrospectively at baseline and followup using the Client Services Receipt Inventory (CSRI) (29), which was customized to ensure it was relevant to individuals with knee pain. The CSRI was administered as a face-to-face clinical interview at the baseline assessment covering the 6 months prior to baseline assessment, and at the followup assessment covering the period between baseline assessment and 6 months after completion of the rehabilitation program (7.5 months from baseline). Data collected included use of health and voluntary resources, use of rehabilitation facilities, patient costs (e.g., travel, medication, over-the-counter drugs), time off work, social security benefits, and informal care. The resources used during the Indiv-rehab and Grp-rehab programs were recorded separately to estimate the costs of rehabilitation (disregarding whether costs were incurred during Indiv-rehab or Grp-rehab).

Unit costs

Costs were calculated for each participant by multiplying resource use data by unit costs based on national statistics (Appendix A, available at the Arthritis Care & Research Web site at Informal care costs were estimated using the replacement cost approach by applying the unit cost of a social services home help worker.

Unit costs of rehabilitation were estimated from a health care perspective. Because resource inputs for a session of Indiv-rehab and Grp-rehab did not vary between sessions or participants, unit costs for each intervention were estimated as an average cost per person per session. Intervention unit cost estimations (described below) were then applied to individual-level data, according to the number of sessions attended, to derive a total intervention cost for each participant. Calculation of these unit costs involved first identifying all resource inputs normally associated with running one session of each program (e.g., contact and noncontact time with the therapist, capital costs, overhead costs, exercise equipment, materials/photocopying). Second, the cost of each of these components was estimated. A research associate (hourly rate £18.78 [$30]) supervised all rehabilitation sessions. Average face-to-face contact time was 45 minutes for each Indiv-rehab session and 60 minutes for each Grp-rehab session; these times were adjusted for group size, then multiplied by the hourly cost to obtain per-person face-to-face contact time costs of £14.09 ($23) and £2.35 ($4) for Indiv-rehab and Grp-rehab sessions, respectively. Average noncontact time (prepare session, write up notes) associated with each person for each session was 30 minutes for Indiv-rehab and Grp-rehab. This was multiplied by the hourly cost to obtain per-person non–face-to-face contact time costs of £9.39 ($15). Third, the costs of all resource components were summed to obtain a total cost per session for each intervention. Finally, the total cost for Grp-rehab was divided by the average group size (n = 8) to estimate its cost per person per session.

All costs were standardized to 2003/2004 prices, using the UK National Health Service Executive hospital and community health services or personal social services inflation indices (30). Discounting costs was unnecessary because of the short assessment period (6 months). Costs are presented in British pounds sterling (£) and converted to US dollars ($) based on a 2003 purchasing power parity (conversion rate: £1 = $1.613) that equalizes the purchasing power of the 2 currencies (31).

Statistical analyses

Because the CSRI was administered by interview, missing data were minimal. However, if use of a particular service was reported but the quantity was not specified, within-group mean costs for that item for participants who did have data at the same assessment point were used to impute missing quantities. Where information about the use of a particular service was missing, we assumed that service had not been used and a zero cost was applied. For missing EQ-5D values at 6-month followup, values were imputed using the last value carried forward. Values could not be imputed for missing EQ-5D scores at baseline.

QALYs were calculated as gains since baseline. This calculation incorporated values from a 6-week measurement. Six-week and 6-month health state measurements were assumed to represent the time since the last assessment. Proportional QALY gains for these 2 time points were summed to represent a total 6-month gain since baseline.

Intent-to-treat analyses were conducted. Mean costs per group were based on participant-level costs, unadjusted for clustering. However, a cluster adjustment was included in the examination of differences in cost between groups. Differences were compared using linear regression with the cluster adjustment procedure in Stata v8.2 (StataCorp, College Station, TX). Each cost category was adjusted for the baseline value of the same category. To allow for the non-normal distribution commonly associated with cost data, 1,000 nonparametric bootstrap replications were performed to obtain estimates of mean differences and 95% confidence intervals (95% CIs). The primary perspective was that of health and social care, the secondary perspective that of society. Analyses were performed using SPSS 12.0.1 (SPSS, Chicago, IL) and Stata 8.2 (StataCorp).

Cost-effectiveness analyses

The main cost-effectiveness analyses focused on between-group differences in total costs and WOMAC-func (the primary outcome) at 6-month followup. In addition, a cost-utility analysis was conducted by combining cost and QALY data.

Uncertainty concerning cost-effectiveness results was represented by cost-effectiveness acceptability curves (CEACs) using the net benefit approach (32,33). CEACs show the probability that one intervention is cost–effective compared with alternative(s), given a range of willingness-to-pay values (WTP) that a health care commissioner may be prepared to accept for improvements in outcome.

For CEACs based on WOMAC-func, we compared Indiv-rehab with Grp-rehab, and rehabilitation with usual care; these comparisons were made from both the health and social care perspective and the societal perspective. To make the cost-effectiveness analysis meaningful to health care commissioners, WOMAC-func was examined as the proportion of participants showing clinically meaningful improvements, i.e., at least 15% from baseline (34). A total of 1,000 bootstrap samples of costs and effects data were used to generate distributions of the mean costs and effects for each arm to deal with the dichotomous nature of the outcome measure.

For the CEACs based on QALYs, each intervention was compared with the alternatives simultaneously, and analyses were only conducted from a health and social care perspective in line with the current decision-making framework used by the UK National Institute for Health and Clinical Excellence. CEACs based on QALYs were based on observed data. Both sets of CEACs were constructed using Excel (Microsoft, Redmond, WA) program macros created by Elizabeth Fenwick (University of York, UK).

Sensitivity analyses

Two sets of sensitivity analyses were performed. First, we tested the effect of alternative assumptions for the calculation of unit costs of rehabilitation. Second, we tested the impact of alternative assumptions for the calculation of total health/social care costs and total societal costs. We examined rehabilitation unit costs under 6 alternative assumptions: supervision by a junior therapist; reducing groups to 6 participants; increasing groups to 10 participants; widening the cost difference between the interventions by 50%; narrowing the cost difference by 50%; taking a societal perspective that includes participant's travel expenses, travel time to and from, and participation in rehabilitation sessions.

For the examination of total costs, sensitivity analyses were performed to examine alternative assumptions concerning medications and informal care. There were no data on the duration of medication use, so we assumed that all reported medications were taken for the full period of assessment. Because this may have overestimated medication costs, costs were reestimated assuming medications were taken for half the assessment period (i.e., medication costs reduced by 50%). The impact of this was tested in the economic evaluation from the health/social care perspective (sensitivity analysis 1) and from the societal perspective (sensitivity analysis 2). Participants also had difficulty estimating the duration of informal care inputs and, specifically, identifying the inputs necessitated by having knee pain, so we excluded these from total societal costs (sensitivity analysis 3).


Participant characteristics

A total of 418 individuals (294 women) were recruited. Baseline anthropometric, clinical, and sociodemographic characteristics were balanced across trial arms; baseline mean age was 66 years (range 50–91 years), mean height 1.64 meters (range 1.30–1.97), body mass 81 kg (range 47–139), body mass index 30 kg/m2 (range 18–51), WOMAC-func 27.2 points (95% CI 25.7, 28.6), and median duration of symptoms was 6 years (interquartile range 3–12). The majority of participants received some state benefits, predominantly a pension. There were some between-group imbalances in receipt of social security benefits at baseline (usual care: 118 [84%] of 140; Indiv-rehab: 108 [74%] of 146; Grp-rehab: 120 [91%] of 132). At 6-month followup, economic data were available for 338 (81%) participants and completion rates were balanced between trial arms (usual care: 110 [79%] of 140; Indiv-rehab: 121 [83%] of 146; Grp-rehab: 107 [81%] of 132).

From the health/social care (although not societal) perspective, participants in the usual care arm who remained in the study at the 6-month assessment had lower total baseline costs and higher (better) baseline utility values than participants who withdrew. Because these differences were not present in the intervention arms, the data of usual care participants remaining in the study were biased toward participants with better quality of life and/or lower costs, causing the cost-effectiveness of rehabilitation to be underestimated.

Outcome results

At baseline there were no differences between the groups in costs or utilities (Table 1). There were some changes in costs and QALYs at the 6-month assessment, but no overall differences between any of the trial arms (Table 2). WOMAC-func improved in participants in Indiv-rehab and Grp-rehab, and the proportion of participants with clinically meaningful improvement in functioning was greater following rehabilitation than usual primary care (121 of 226 versus 47 of 113, respectively; χ2 = 4.301, P = 0.038) (Table 2)

Table 1
Mean and mean differences in costs (£, 2003/2004) and quality-adjusted life years at baseline*
Table 2
Mean and mean differences in costs (£, 2003/2004) and QALYs, and differences in proportion of individuals improving on WOMAC function score at 6 months*

Used resources

Resource use was low, with outpatient and general practitioner visits the most commonly used services. Mean visits among users of general practitioner services were 2.6 for usual care, 1.7 for Indiv-rehab, and 1.9 for Grp-rehab. Mean visits among users of outpatient services were 2.3 for usual care, 1.3 for Indiv-rehab, and 2.2 for Grp-rehab. Although approximately two-thirds of participants received some level of informal care, it was at relatively low levels (Table 3).

Table 3
Number of participants and average weekly duration of informal care inputs at 6-month followup*

Rehabilitation costs

Mean costs

Usual care incurred no rehabilitation costs. Indiv-rehab cost £30.40 ($49) per session per person (Table 4). Accounting for attendance rates, mean costs for Indiv-rehab were £314 ($507) per person (Tables 2 and Tables 4). Grp-rehab cost £14.33 ($23) per session per person (a total cost of £114.64 [$185] per session based on an average group size of 8 participants) (Table 4). Mean cost of Grp-rehab was £125 ($202) per person (Tables 2 and and44)

Table 4
Unit costs of intervention, mean ± SD intervention costs, and mean differences in costs (95% CI)*

Differences in mean costs

Participation in rehabilitation cost £224 (95% CI £184, £262 [$361; 95% CI $297, $423]) more than usual care, with differences of £171–£359 ($276–$579) in the alternative scenarios (Tables 2 and and4).4). Indiv-rehab cost £189 (95% CI £168, £208 [$305; 95% CI $271, $336]) more per person than Grp-rehab, and was more expensive than Grp-rehab under all alternative unit cost assumptions with differences of £128–£344 ($207–$555) (Tables 2 and and44).

Total costs

There were no differences in total costs between Indiv-rehab and Grp-rehab (Table (Table2).2). Although rehabilitation did not have significantly greater costs than usual care from a health/social care perspective, it cost £584 more from a societal perspective and based on 2 sensitivity analysis scenarios (Table 2).


Only participants with both cost and relevant outcome data were included in the cost-effectiveness analyses. For the WOMAC-based analyses from both perspectives, the sample sizes were as follows: 110 (79%) of 140 in usual care, 118 (81%) of 146 in Indiv-rehab, and 107 (81%) of 132 in Grp-rehab. For the QALY-based analyses, the sample sizes were 109 (78%) of 140 in usual care, 121 (83%) of 146 in Indiv-rehab, and 106 (80%) of 132 in Grp-rehab.

If decision makers were unwilling to pay anything for an increased proportion of individuals with improved functioning, the probability that rehabilitation would be cost-effective compared with usual care was 9% (Figure 1). If health care commissioners were willing to pay ~£800 ($1,291) for rehabilitation, there was 50% probability that supplementing usual primary care with rehabilitation was more cost-effective than usual primary care alone. This increased to >90% at WTP values ≥£1,900 (≥$3,065) (Figure 1)

Figure 1
Cost-effectiveness acceptability curves: probability (given as percentage) that 1) rehabilitation (individual or group) is cost-effective compared with usual primary care and 2) Indiv-rehab is cost-effective compared with Grp-rehab, for a range of values ...

Compared with Grp-rehab, Indiv-rehab had higher costs but only marginally better outcomes. If decision makers were unwilling to pay anything for an increase in the proportion of individuals improving their functioning, there was almost zero probability Indiv-rehab would be more cost-effective than Grp-rehab (Figure 1). The probability of Indiv-rehab being more cost-effective than Grp-rehab increased as commissioners' WTP increased, reaching 50% probability at WTP values >£6,000 (>$9,678).

Because there was very little difference in QALYs across the arms (Table 2), CEACs based on this outcome measure showed relatively low probabilities of cost-effectiveness for both Indiv-rehab and Grp-rehab (Figure 2). It was not until threshold values were ≥£19,200 that either Indiv-rehab or Grp-rehab showed higher probabilities of cost-effectiveness than usual care. At this threshold and above, the probability that Indiv-rehab would be cost-effective compared with Grp-rehab and usual care exceeded probabilities of cost-effectiveness for the other 2 treatment strategies. However, the probability did not exceed 38% in the range of thresholds examined.

Figure 2
Cost-effectiveness acceptability curves: probability that each treatment strategy is cost-effective compared with the other 2, for a range of values of decision makers' willingness to pay for an additional quality-adjusted life year (QALY), from a health/social ...


This is the first study evaluating the costs and cost-effectiveness of an integrated exercise and self-management rehabilitation program designed to improve functioning in persons with chronic knee pain. Rehabilitation had cost implications, but at modest levels of investment was more likely to be cost-effective than usual primary care: investing £1,900 (or more) provided a 90% (or greater) chance of rehabilitation being more cost-effective than usual primary care. Administering ESCAPE-knee pain to small groups of individuals reduced its costs without compromising clinical effectiveness, increasing the probability of cost-effectiveness.

The strengths of this trial and intervention are its rigour, clinical relevance, and generalizability (35). Performing prospective interview-based economic evaluation alongside the clinical trial minimized missing data. Constructing CEACs is more informative than incremental cost-effectiveness ratios because they indicate the uncertainty associated with the findings in the absence of knowledge of how much health care commissioners would be willing to pay for meaningful improvements in functioning.

Inaccurate or biased recall due to memory and cognitive impairments is a concern in elderly individuals and may have affected the results. However, none of our participants had overt memory problems, and all were light users of health care services, had a slowly progressive condition, and lived independently, therefore we have no reason to believe that short-term recall of resource utilization was very inaccurate or biased toward over-or underestimation of resource use. Like many population-based economic evaluations with a broad perspective, it was not possible to validate self-reported resource use against all relevant service provider records. There is no better alternative; medical records are reasonably valid for conditions requiring medical diagnosis and monitoring (e.g., diabetes, hypertension) but are less valid for conditions that have important effects on individuals but do not require continued supervision, i.e., chronic knee pain (36).

Assessment of the amount of informal care individuals receive is problematic but important (37). Participants found it difficult to estimate informal care in terms of time. Frequently what we defined as informal care (housework, shopping, gardening, emotional support) were activities that partners, family, and friends normally performed as part of a relationship that participants played down. Some may have been reluctant to admit needing and accepting help. Many coped by adapting their environment and activities (showering rather than bathing) with minimal help. Many ignored nonessential tasks (e.g., decorating) and refrained from or reduced participation in social activities. This might underestimate the amount of informal care needed.

Although the CSRI asked about impacts related to knee pain, this is difficult to disentangle from the impact of other comorbidities common in elderly persons. Consequently, the impact of knee pain may have been overstated inadvertently. Conversely, because knee pain is a major determinant of functional limitation and functional limitation is a risk factor for chronic ill health, the indirect health care costs for chronic conditions due to functional limitations caused by knee pain may be underestimated.

Because most participants were retired, lost productivity and social welfare benefits were small. Although all the participants reported chronic knee pain, health care costs were low, mostly repeat medication. A large proportion of individuals were taking over-the-counter remedies (e.g., fish oil, glucosamine), but few were receiving therapies often sought for chronic pain (e.g., acupuncture, private therapies). Individuals accepted and coped with functioning limitations by decreasing or altering their activities. This enabled them to maximize their independence and minimize the burden on others. Assistance was most frequently necessary for essential arduous tasks (e.g., housework, shopping).

Overall, health professionals provide, and people seem to expect and accept, a low level of care for chronic knee pain. Chronic knee pain is regarded as a benign, inevitable consequence of aging, for which little can be done beyond prolonged medication, regardless of its costs (17) and risks (9,1416). However, because of the prevalence of knee pain, health and social care expenditure is very great (3,79) and likely to escalate as the working life extends (increasing lost productivity, earnings, and social security payments), the population ages, and consultations for musculoskeletal disorders increase (11).

Usual primary care was the cheapest option but least cost-effective. Supplementing this with rehabilitation incurred extra costs, but improved functioning. Group rehabilitation achieved comparable improvement in functioning at half the cost of individual rehabilitation, and therefore was most cost-effective. This finding could be used to justify withdrawing provision of individual therapy. However, because individual sessions can be arranged at convenient times and missed sessions can be re-arranged, removing this flexible management option might deny some people (the employed) effective treatment. It may be preferable to retain this management option but reduce its costs and maximize efficiency by finding the optimal number of individual sessions without compromising effectiveness.

We found little difference in QALYs between the arms. ESCAPE-knee pain was a brief intervention designed to improve functioning and be deliverable to many individuals. It was not expected or designed to enable individuals to function normally, completely alleviate pain, or substantially affect all factors that contribute to quality of life in an elderly, inner-city population (e.g., comorbidity, family and social issues, anxiety, depression, socioeconomic situation). Therefore, quality of life will remain compromised. Moreover, the EQ-5D may be insensitive to quality-of-life change in this patient group (38,39) and the trial was not powered to detect meaningful change in this outcome.

Although intervention costs could be reduced by using junior therapists and/or larger groups, safety issues might necessitate additional personal support, which could negate potential savings. We considered the health and social care perspective to be the most appropriate to health commissioners considering productivity, because lost earnings, sickness benefit, and travel time are not major issues in this retired patient population. Including these issues in a societal perspective increased the costs in all the trial arms, but the broad inferences from the study remain unaltered.

This is the first study integrating exercise and self-management for persons with chronic knee pain. There is evidence that self-management is cost-effective for chronic knee pain (40,41). There is less evaluation of the costs of exercise, but recent studies of similar patient populations demonstrated that prolonged exercise regimens resulted in benefits immediately following completion (42,43) and were cost-effective (4446). The important differences in this study are that the improved functioning attained during ESCAPE-knee pain was sustained for 6 months, the number needed to treat was smaller (better) (43,45), and the intervention was less costly (44,45).

Interventions for chronic pain conditions will benefit few individuals if they require participation in complex, prolonged interventions that are impractical and too expensive to implement widely. As more individuals live and work longer with chronic ill health, safe, affordable interventions will be needed to maximize functioning, independence, and efficient use of health and societal resources. ESCAPE-knee pain was designed to be a brief, affordable, and cost-effective intervention to improve functioning for many people. We have previously shown it to be safe and effective (24,25). This economic evaluation demonstrates that ESCAPE-knee pain is cost-effective if value is placed on meaningful improvements in functioning.


Dr. Hurley had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Hurley, Walsh, Pimm, Jones, Reeves, Dieppe, Patel.

Acquisition of data. Hurley, Walsh, Mitchell.

Analysis and interpretation of data. Hurley, Walsh, Pimm, Williamson, Jones, Reeves, Dieppe, Patel.

Manuscript preparation. Hurley, Walsh, Mitchell, Pimm, Williamson, Jones, Reeves, Dieppe, Patel.

Statistical analysis. Hurley, Williamson, Patel.

Trial Steering Group members. Hurley, Walsh, Mitchell, Pimm, Williamson, Jones, Reeves, Dieppe, Patel.


The authors thank all the participants who gave their time and effort during this study, and the primary care practices who agreed to participate. The authors are grateful for the constructive comments of 2 anonymous reviewers on an earlier version of this article.

Supplemental material


1. Badley EM, Rasooly I, Webster GK. Relative importance of musculoskeletal disorders as a cause of chronic health problems, disability, and health care utilization: findings from the 1990 Ontario Health Survey. J Rheumatol. 1994;21:505–14. [PubMed]
2. Badley EM, Ibanez D. Socioeconomic risk factors and musculoskeletal disability. J Rheumatol. 1994;21:515–22. [PubMed]
3. Peat G, McCarney R, Croft P. Knee pain and osteoarthritis in older adults: a review of community burden and current use of primary health care. Ann Rheum Dis. 2001;60:91–7. [PMC free article] [PubMed]
4. Dawson J, Linsell L, Zondervan K, Rose P, Randall T, Carr A, et al. Epidemiology of hip and knee pain and its impact on overall health status in older adults. Rheumatology (Oxford) 2004;43:497–504. [PubMed]
5. Alonso J, Ferrer M, Gandek B, Ware JE, Aaronson NK, Mosconi P, IQOLA Project Group et al. Health-related quality of life associated with chronic conditions in eight countries: results from the International Quality of Life Assessment (IQOLA) Project. Qual Life Res. 2004;13:283–98. [PubMed]
6. Cooper JK, Kohlmann T. Factors associated with health status of older Americans. Age Ageing. 2001;30:495–501. [PubMed]
7. Lanes SF, Lanza LL, Radensky PW, Yood RA, Meenan RF, Walker AM, et al. Resource utilization and cost of care for rheumatoid arthritis and osteoarthritis in the managed care setting: the importance of drug and surgery costs. Arthritis Rheum. 1997;40:1475–81. [PubMed]
8. Gabriel SE, Crowson CS, Campion ME, O'Fallon WM. Indirect and nonmedical costs among people with rheumatoid arthritis and osteoarthritis compared with nonarthritic controls. J Rheumatol. 1997;24:43–8. [PubMed]
9. Leardini G, Salaffi F, Caporali R, Rovati L, Montanelli R, Italian Group for the Study of the Costs of Arthritis Direct and indirect costs of osteoarthritis of the knee. Clin Exp Rheumatol. 2004;22:699–706. [PubMed]
10. Hootman JM, Helmick CG. Projections of US prevalence of arthritis and associated activity limitations. Arthritis Rheum. 2006;54:226–9. [PubMed]
11. Perruccio AV, Power JD, Badley EM. Revisiting arthritis prevalence projections: it's more than just the aging of the population. J Rheumatol. 2006;33:1856–62. [PubMed]
12. Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Org. 2003;81:646–56. [PubMed]
13. World Health Organisation. The burden of musculoskeletal disease at the start of the millennium: technical report 919. Geneva: WHO; 2003. [PubMed]
14. Juni P, Nartey L, Reichenbach S, Sterchi R, Dieppe P, Egger M. Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. Lancet. 2004;364:2021. [PubMed]
15. Hippisley-Cox J, Coupland C. Risk of myocardial infarction in patients taking cyclo-oxygenase-2 inhibitors or conventional non-steroidal anti-inflammatory drugs: population based nested case-control analysis. BMJ. 2005;330:1366–9. [PMC free article] [PubMed]
16. Hippisley-Cox J, Coupland C, Logan R. Risk of adverse gastrointestinal outcomes in patients taking cyclo-oxygenase-2 inhibitors or conventional non-steroidal anti-inflammatory drugs: population based nested case-control analysis. BMJ. 2005;331:1310–6. [PMC free article] [PubMed]
17. Solomon DH, Glynn RJ, Bohn R, Levin R, Avorn J. The hidden cost of nonselective nonsteroidal antiinflammatory drugs in older patients. J Rheumatol. 2003;30:792–8. [PubMed]
18. Ettinger WH, Afable RF. Physical disability from knee osteoarthritis: the role of exercise as an intervention. Med Sci Sports Exerc. 1994;26:1435–40. [PubMed]
19. Newman PS, Steed L, Mulligan K. Self-management interventions for chronic illness. Lancet. 2004;364:1523. [PubMed]
20. Warsi A, LaValley MP, Wang PS, Avorn J, Solomon DH. Arthritis self-management education programs: a meta-analysis of the effect on pain and disability. Arthritis Rheum. 2003;48:2207–13. [PubMed]
21. Van Baar ME, Assendelft WJ, Dekker J, Oostendorp RA, Bijlsma JW. Effectiveness of exercise therapy in patients with osteoarthritis of the hip and knee: a systematic review of randomised clinical trials. Arthritis Rheum. 1999;42:1361–9. [PubMed]
22. Fransen M, McConnell S, Bell M. Therapeutic exercise for people with OA of the hip and knee: a systematic review. J Rheumatol. 2002;29:1737–45. [PubMed]
23. Roddy E, Zhang W, Doherty M. Aerobic walking or strengthening exercise for osteoarthritis of the knee? A systematic review. Ann Rheum Dis. 2005;64:544–8. [PMC free article] [PubMed]
24. Hurley MV, Walsh NE, Mitchell H, Williamson E. Enabling Self-Management and Coping with Arthritic Knee Pain through Exercise (Escape-Knee Pain): a cluster randomised trial of an integrated rehabilitation programme. Arthritis Rheum. 2005;52(Suppl 9):S505. [abstract]
25. Hurley MV, Walsh NE, Mitchell HL, Pimm TJ, Patel A, Williamson E, et al. Clinical effectiveness of a rehabilitation program integrating exercise, self-management, and active coping strategies for chronic knee paa cluster-randomized trial. Arthritis Rheum. 2007;57:1211–9. [PMC free article] [PubMed]
26. Bellamy N, Buchanan W, Goldsmith C, Campbell J, Stitt L. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol. 1988;15:1833–40. [PubMed]
27. The EuroQol Group. EuroQol: a new facility for the measurement of health related quality of life. Health Policy. 1990;16:199–208. [PubMed]
28. Dolan P, Gudex C, Kind P, Williams A. Working papers. York: Centre for Health Economics, University of York; 1995. A social tariff for EuroQol: results from a UK general population survey.
29. Knapp M. The economic evaluation of mental health care. Aldershot (UK): Arena; 1995.
30. Curtis L, Netten A, editors. Unit costs of health and social care 2004. Kent (UK): Personal Social Services Research Unit, University of Kent; 2004.
31. Organisation for Economic Co-operation and Development. URL Purchasing power parities, comparative price levels.
32. Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost-effectiveness acceptability curves. Health Econ. 2001;10:779–87. [PubMed]
33. Briggs AH, O'Brien BJ, Blackhouse G. Thinking outside the box: recent advances in the analysis and presentation of uncertainty in cost-effectiveness studies. Annu Rev Public Health. 2002;23:377–401. [PubMed]
34. Angst F, Aeschlimann A, Stucki G. Smallest detectable and minimal clinically important differences of rehabilitation intervention with their implications for required sample sizes using WOMAC and SF-36 quality of life measurement instruments in patients with osteoarthritis of the lower extremities. Arthritis Rheum. 2001;45:384–91. [PubMed]
35. Glasgow RE, Magid DJ, Beck A, Ritzwoller D, Estabrooks PA. Practical clinical trials for translating research to practice: design and measurement recommendations. Med Care. 2005;43:551–7. [PubMed]
36. Beckett M, Weinstein M, Goldman N, Yu-Hsuan L. Do health interview surveys yield reliable data on chronic illness among older respondents? Am J Epidemiol. 2000;151:315–23. [PubMed]
37. Van den Berg B, Brouwer WB, Koopmanschap MA. Economic valuation of informal care: an overview of methods and applications. Eur J Health Econ. 2004;5:36–45. [PubMed]
38. Fransen M, Edmonds J. Reliability and validity of the EUROQoL in patients with osteoarthritis of the knee. Rheumatology (Oxford) 1999;38:807–13. [PubMed]
39. Brazier JE, Harper R, Munro J, Walters SJ, Snaith ML. Generic and condition-specific outcome measures for people with osteoarthritis of the knee. Rheumatology (Oxford) 1999;38:870–7. [PubMed]
40. Lorig KR, Sobel DS, Stewart AL, Brown BW, Bandura A, Ritter P, et al. Evidence suggesting that a chronic disease self-management program can improve health status while reducing hospitalisation: a randomized trial. Med Care. 1999;37:5–14. [PubMed]
41. Mazzuca SA, Brandt KD, Katz BP, Hanna MP, Melfi CA. Reduced utilization and cost of primary care visits resulting from self-care education for patients with osteoarthritis of the knee. Arthritis Rheum. 1999;42:1267–73. [PubMed]
42. McCarthy CJ, Mills PM, Pullen R, Roberts C, Silman A, Oldham JA. Supplementing a home exercise programme with a class-based exercise programme is more effective than home exercise alone in the treatment of knee osteoarthritis. Rheumatology (Oxford) 2004;43:880–6. [PubMed]
43. Thomas KS, Muir KR, Docherty M, Jones AC, O'Reilly SC, Bassey EJ. Home based exercise programme for knee pain and knee osteoarthritis: randomise controlled trial. BMJ. 2002;325:752–6. [PMC free article] [PubMed]
44. Richardson G, Hawkins N, McCarthy CJ, Mill PM, Pullen R, Roberts C, et al. Cost-effectiveness of a supplementary class-based exercise programme in the treatment of knee osteoarthritis. Int J Technol Assess Health Care. 2006;22:84–9. [PubMed]
45. Thomas KS, Miller P, Doherty M, Muir KR, Jones AC, O'Reilly SC. Cost effectiveness of a two-year home exercise program for the treatment of knee pain. Arthritis Rheum. 2005;53:388–94. [PubMed]
46. Coupe VM, Veenhof C, van Tulder MW, Dekker J, Bijlsma JW, van den Ende CH. The cost-effectiveness of behavioural graded activity in patients with osteoarthritis of hip and/or knee. Ann Rheum Dis. 2007;66:215–21. E-pub 2006 July 31. [PMC free article] [PubMed]

Articles from Wiley-Blackwell Online Open are provided here courtesy of Wiley-Blackwell, John Wiley & Sons