PURPOSE This article examines two competing hypotheses for the impact of disability and age on health service utilisation in Canada: the double jeopardy and age-as-leveller hypotheses. METHOD The study uses a retrospective cohort design to examine the effect of age and disability on four aspects of health service utilisation: family doctor, medical specialist, hospital and homecare. The cohort was assembled from the longitudinal component of the National Population Health Survey. The effective sample size for this analysis was 1629. RESULTS This study showed that disability is a stronger predictor of doctor and hospital utilisation than age. No significant relationship was found between age and specialist use, and there were only small to moderate increases in the use of family doctors and hospitals with each 5-year increment of age over 65. There is a strong association between the use of home care and both age and disability. Results support the age-as-leveller hypothesis, in that negative interaction effects were found between age and disability for use of both family physicians and medical specialists. In other words, age and disability together have an effect that is less than would be expected, given the main effects of each. CONCLUSION The results of this study support the importance of disability as an indicator of health service utilisation. Rehabilitation practitioners are encouraged to continue to sensitise other members of the health care team to the importance of disability as a way of understanding health and health service use.