Osteoarthritis, Patient Centered Outcomes and Complementary and Alternative Therapy (CAM)

June 15, 2010 (historical)


Principal Investigator: Charles b. Eaton, MD
Director, CPCP
Memorial Hospital of Rhode Island
111 Brewster Street
Pawtucket, RI 01860

Project Abstract

Osteoarthritis (OA), the most common form of arthritis, is a slowly progressing disease characterized clinically by pain, deformity, and loss of function. OA is a significant contributor to disability and loss of independence among the elderly. Knee OA is associated with a progressive reduction in function, including difficulty in changing from the sitting to the standing position, and a decrease in mobility and in the ability to carry out activities of daily living. While recent scientific advances have yielded highly effective disease-modifying therapies for rheumatoid arthritis, no such therapies exist for osteoarthritis. Because of the chronic nature of the disease and variable clinical outcomes, a better understanding of the relationship between various measures of clinically relevant or patient-centered outcomes — pain, stiffness, reduced function, use of analgesics, disability, quality of life and composite scores accounting for these and other clinically relevant outcomes are needed. Due to its chronic nature and the lack of effective disease modifying therapies a range of biobehavioral, dietary, pharmacologic and complementary and alternative (CAM) therapies are used by patients to reduce OA symptoms, pain, and improve function and quality of life. Despite their wide use, relatively few CAM therapies have been tested adequately with appropriate study designs to assess their effectiveness in improving patient-centered outcomes. The Osteoarthritis Initiative is large epidemiologic study with multiple patient-centered outcomes assessed yearly in subjects at high risk or having already developed knee OA as defined by fixed flexion radiographs at baseline. In addition at baseline, a large inventory of biobehavioral, dietary, pharmacologic and CAM therapies was assessed and most of these "treatments" are assessed either yearly or every two years throughout 8 years of projected follow-up. When studying the effectiveness of biobehavioral, dietary, pharmacologic and CAM therapies on improving patient centered outcomes related to knee osteoarthritis several analytic issues arise when evaluating epidemiological data. First no validated composite outcome measure has been developed that accounts for the full impact of knee osteoarthritis on patient-centered outcomes. While Western Ontario and McMaster Universities Osteoarthritis index (WOMAC) and Knee Injury and Osteoarthritis Outcomes Score (KOOS) summary scales have been developed they rely exclusively on subjective outcomes, and are highly inter-correlated making a summary score of questionable value. Second, participant s at baseline with worse patient-centered outcomes and knee osteoarthritis are more likely to utilize CAM therapies in an attempt to alleviate symptoms. This "confounding by indication" biases analyses of the potential beneficial of CAM therapies effects towards the null, making it difficult to discern whether they are effective for improvement of knee OA pain and symptoms. State-of-the art statistical methods have been developed that deal reasonably effectively with these analytic issues and will be employed in this project.

Specific Aims

The specific aims are:

Aim 1. Characterize the biobehavioral, dietary, pharmacologic and CAM therapies utilized by OAI participants both at baseline and /longitudinally. Characterize patterns of use, predictors of use at baseline and patterns of use and predictors of continued use, intermittent use or decreased use over time.

Hypothesis 1a: CAM therapy at baseline will be greater in those participants with more severe symptoms (frequency and intensity) and severity of radiographic osteoarthritis as well as other socio-demographic factors (gender, race).

Hypothesis 1b: The predictors of the use of multiple CAM therapies will differ than those that us just one or two therapies.

Hypothesis 1c: There will be a pattern of use of CAM therapies such that certain therapies such as use of NSAID, glucosamine and chondroitin will cluster together more than others such as use of chiropractor, herbs and copper bracelets.


Aim 2. Characterize patient-centered outcomes (pain, loss of function, disability, quality of life) in OAI including developing a clinical severity score using principal component analysis (PCA) to define an overall clinical severity score both at baseline and changes longitudinally over six years of follow-up.

Hypothesis 2a: A composite clinical severity score with demonstrate greater responsiveness to change (standardized response means) than individual measures such as WOMAC pain, stiffness, disability, and KOOS symptoms, sports and recreation function, quality of life and participant global assessments.

Hypothesis 2b: Participants with greater levels of symptoms, pain, functional limitations and severity of radiographic OA at baseline will demonstrate more rapid changes in patient centered outcomes over 4-6 years of follow-up.


Aim 3. Analyze the effectiveness of different biobehavioral, dietary, and CAM therapies using propensity score matching analysis, instrumental variable analysis and marginal structural models on patient-centered outcomes defined in specific aim 2.

Hypothesis 3a: Participants, who are not obese and maintain a stable weight or those who are obese and lose weight at a threshold level, will have less progression of poor patient centered outcomes (pain symptoms, loss of function, worse quality of life).

Hypothesis 3b: Participants, who are not sedentary and maintain an active lifestyle or those who are sedentary and become physical active at a threshold level, will have less progression of poor patient centered outcomes (pain, symptoms, loss of function, worse quality of life).

Hypothesis 3c: Participants who have diets rich in fish, antioxidants, vitamin D, Vitamin K, including a Mediterranean diet will have less progression of poor patient centered outcomes (pain, symptoms, loss of function and worse quality of life) compared to participants with diet low in these food, vitamins, and dietary patterns.

Hypothesis 3d: Participants who utilize CAM therapies on a long term basis will have better patient centered outcomes than those who stop use or who never use CAM therapies after adjusting for confounding by indication.

Hypothesis 3e: Participants who utilize glucosamine on a long term basis compared to non-steroidal medication therapy will have better quality of life and health utilities over four to six years of follow-up.