Abstract
OBJECTIVES: Alternative definitions of arthritis in community surveys provide very different estimates of arthritis prevalence among older Americans. This telephone interview study examines prevalence estimates based on the current Behavioral Risk Factor Surveillance System (BRFSS) arthritis case definition. METHODS: Interviews were conducted with 851 Chicago residents age 45 and older. Logistic regression was used to compare the age and sex controlled prevalence of poor health, restricted activity, and arthritis risk factors among those with a previous arthritis diagnosis from a health professional, those with undiagnosed chronic joint symptoms, and those who were joint symptom free and without a previous arthritis diagnosis. RESULTS: BRFSS-defined arthritis prevalence was 47% of older residents, including 33% reporting a previous arthritis diagnosis and 14% chronic undiagnosed joint symptoms. Only 25% of these respondents reported current arthritis treatment by a doctor. After controlling for age and sex, respondents with a previous arthritis diagnosis and those with undiagnosed chronic symptoms had significantly worse health and functioning, and more prevalent arthritis risk factors, than respondents without joint symptoms. CONCLUSIONS: BRFSS-defined arthritis included almost half the area population over age 45. Both diagnosed and undiagnosed chronic joint symptoms are associated with major functional limitations and arthritis risk after controlling for age and sex. The inclusion of previously undiagnosed chronic joint symptoms in the BRFSS arthritis definition is appropriate and indicates that previous arthritis prevalence estimates may be too low.