Determinants of Arthritis Disability and Joint Replacement

In 1999, we received further funding from CIHR to continue our cohort study in order to investigate the predictors of time to receipt of joint replacement, and to describe patterns of change in arthritis-related symptoms and disability over time and their associated economic impact.

The overall objective of this study was to reduce the burden of illness in the population due to arthritis disability through development of strategies for improving the provision of health care to individuals with arthritis.

The specific objectives were:

  1. to investigate the determinants of arthritis-related disability in a community-based cohort of individuals with arthritis;
  2. to determine the patient and provider characteristics which best predict time to total joint replacement (primary and revision); and
  3. to describe changes in arthritis-related symptoms and disability, and their associated economic impact (health care utilization, prescription drug costs) in a population-based cohort.


All participants who had been included in the Study of Arthritis in Your Community, and were alive and well, completed annual telephone interviews over a five-year period. These interviews assessed their living circumstances, general health status, arthritis care and their arthritis severity. We also asked about direct and indirect costs related to arthritis, such as equipment purchased, paid help with chores, time lost from employment and household chores, and unpaid help provided by family and friends. Arthritis costs to the health care system were estimated by linking data from the annual questionnaires to provincial databases of physician use, hospital visits, home care, long-term care and use of prescription drugs.


Data collection ended in the spring of 2004. Data analysis and manuscript preparation are ongoing.

Patterns of OA Over Time

We developed and utilized a novel statistical approach (factor followed by cluster analysis) to identify distinct patterns of change in OA symptoms and disability over time using the cohort data. Four distinct clusters were identified, including one cluster of individuals (approximately ¼ of the cohort) who experienced unexpected and significant self-reported improvement in their OA symptoms and disability. Prior studies had generally examined mean changes over time for groups, and thus provided no information either about changes within individuals, or patterns of change from start to finish. The approach used in our paper was extremely innovative and documented for the first time variable and distinct patterns of disease “progression” in OA. This initial statistical work has lead to further analyses to identify predictors of the different patterns of change, including predictors of worsening versus improvement over time.

  • Leffondre K, Abrahamowicz M, Regeasse A, Hawker GA, Badley EM, McCusker J, Belzile E. Statistical measures were proposed for identifying longitudinal patterns of change in quantitative health indicators. J Clin Epidemiol 2004; 57: 1049-1062.
  • Leffondre K, Abrahamowicz M, Hawker GA, Badley EM, Regeasse A, McCusker J, Belisle E. Longitudinal patterns of change in osteoarthritis. Arthritis & Rheumatism 2003; 48(Suppl.9): 1736.

Costs Associated with Living with Arthritis

Informal care plays a major role in the total care provided to people with chronic diseases like OA. With increasing efforts by governments to contain health care expenditures by minimizing lengths of in-hospital stay for joint replacement surgery, and shifting post-operative rehabilitation from an in-patient setting to home, the need for informal care is expected to increase.

This study demonstrated that the costs of OA are substantial, and due mainly to indirect costs. Sixty percent of participants reported out-of-pocket costs and/or time costs related to their arthritis. The average cost was $12,200 CDN per year. Most of these costs (80%) were associated with time lost from employment and leisure activities and for unpaid caregivers, such as family and friends, providing care or help with chores.

The more severe their arthritis was, or the poorer their general health was, the more likely they were to have higher out-of-pocket and time costs. Men were half as likely to report having costs than women (maybe because they are more likely to be living with someone). But when men did report costs, their expenses were 25% higher than those reported by women. Further, we showed that indirect costs are split between those attributable to the individual with OA and those attributable to their caregivers, with the value of caregiver time accounting, on average, for 40% of total indirect cost.

We feel this study clearly illustrates that failure to incorporate caregiver costs undervalues the cost of illness, and more importantly, prevents an honest evaluation of the tradeoffs in choosing between various delivery and treatment options.

  • Wickremaarachi S, Hawker GA, Laporte A, Croxford R, Coyte PC. The Economic Burden of Disabling Hip and Knee Osteoarthritis (OA) from the Perspective of Individuals Living with this Condition. (Rheumatology, accepted with revision).
  • Hawker GA, Coyte P, Wickremaarachi S, Croxford R, Laporte A, Badley EM. Costs of living with osteoarthritis from the patients’ perspective. Arthritis & Rheumatism 2004; 50(Suppl. 9):S505.

Costs to the Health Care System

The study revealed that over a 1-year period, the average health care costs due to arthritis was $6,150 CDN per person (ranging from $0 CDN - $114,330 CDN). Physician services and lab tests accounted for about half of the costs to the health care system. One quarter of these costs were related to prescription medications. The rest of the costs were due to hospitalizations, long-term and home care, and for Emergency Department visits. Costs were higher for older individuals, individuals with greater arthritis disability and those with poorer general health. Health care costs were also higher in the urban region (East York) than in the rural region (Oxford County), possibly reflecting differences in access to arthritis care between the two regions. Arthritis related costs may be even higher than our estimates because the cost of non-prescription medications and services such as physiotherapy or chiropractic care were not included in our estimate of health care costs.

  • Hawker GA, Badley EM, Guan J, Croxford R, Coyte P. Health system costs associated with living with osteoarthritis. Arthritis & Rheumatism 2004; 50(Suppl. 9):S505.

Cost effectiveness of joint replacement

Health system costs were higher for people on the waiting list for a total knee and hip replacement. We looked to see if health care costs (excluding the costs related to the surgery itself) were lower after hip or knee joint replacement. We found that arthritis-related health care costs decreased following a joint replacement, providing further proof that this surgery is cost-effective.

  • Hawker GA, Guan J, Croxford R, Coyte P, Badley EM. Total joint arthroplasty is cost-saving at a population level. Arthritis & Rheumatism 2004; 50(Suppl. 9):S676.

Time to TJR

The factors that predicted how long someone would wait to undergo a total joint replacement were examined. Time to TJR was shorter amongst: those with more severe arthritis, those with fewer other health problems, those who were not too young and too old and those who were willing to consider a total joint replacement.

  • Hawker GA, Badley EM, Guan J, Croxford R. Predictors of time to total joint replacement of the hip and knee in a community cohort. Arthritis & Rheumatism 2004; 50(Suppl. 9):S322.


Gillian Hawker (PI), Elizabeth Badley, Peter Coyte, Richard Glazier, Jack Williams, James Wright


Determinants of Arthritis Disability and Joint Replacement


Medical Research Council of Canada