Health care costs for volunteers and non‐volunteers in an intervention for people with osteoarthritis

OBJECTIVE: To determine whether older people who volunteered for a health intervention study used the health care system differently from a randomly selected group of eligible non-volunteers. METHODS: Three hundred sixty-three volunteers and 92 non-volunteers were compared; both groups were members of a large health maintenance organization (HMO). RESULTS: Volunteers were more educated and had been members of the HMO longer. Volunteers used the health care system significantly more at all time periods, but their cost per contact was significantly lower than that for non-volunteers. Non-volunteers were more likely to have a comorbid condition and had fewer arthritis-related health care contacts. CONCLUSIONS: There will always be problems in generalizing results of studies with volunteers, but volunteers are, nevertheless, the most appropriate controls if the experimental participants are also volunteers. Designs should ideally include both volunteer and non-volunteer controls.