Generic and disease-specific health status instruments are commonly used to assess patients' outcomes. The hypothesis that they measure distinct but complementary aspects of patients' quality of life was tested using a sample of patients aged 67 to 99 years who had undergone knee replacement surgery 2 to 7 years previously. Patients' scores on a generic health-related quality-of-life (HRQOL) measure, the SF-36, were compared to those of the Western Ontario and McMaster Universities (WOMAC) Osteoarthritis Index; the WOMAC was developed specifically for patients with lower extremity arthritis, whereas the SF-36 is aimed at all conditions. A stratified sample of 1,750 Medicare beneficiaries was surveyed and an overall response rate of 80.3% achieved, resulting in 1,193 usable surveys (after adjustment for ineligible, incapacitated, and deceased individuals). The distribution of scores on the three dimensions common to both instruments (i.e., pain, physical function, and overall score) showed consistently higher scores on the WOMAC, on a scale of 0 (worst) to 100 (best), than on the SF-36, indicating less disability from arthritis than from other conditions after knee surgery in this elderly population. Statistically significant differences in the number of people with perfectly healthy scores were detected between the instruments; with regard to pain, 32.2% of the sample reported no pain due to arthritis on the WOMAC, compared with only 13.6% reporting no pain due to any conditions on the SF-36. The figures for physical function and overall score were 9.6% versus 1.4%, and 6.9% versus 0.2%, respectively. Examination for discriminant validity showed that the scores on the two scales followed hypothesized patterns: the WOMAC discriminated better among subjects with varying severity of knee problems, whereas the SF-36 discriminated better among subjects with varying levels of self-reported health status and comorbidity. The results of this study support the inclusion of both a generic and a disease-specific HRQOL measure to assess patient outcomes fully.