Many researchers have reported gender differences in levels of reported symptoms, morbidity, mortality and medical care utilization, but the debate continues about the underlying causes of these differences. Some have argued that women use more medical services because they are more sensitive to symptoms and interested in health, while others believe that women's greater service utilization arises from the fact that women experience more morbidities than do men. To date, these questions have not been studied prospectively. Using data from a household interview survey carried out in 1970-1971 and linked to 22 years of health services utilization records, we explored the effects of gender, self-reported health status, mental and physical symptom levels, health knowledge, illness behaviors and health concerns and interest on the long-term use of health services. After controlling for the aforementioned factors, female gender remained an independent predictor of higher utilization over the 22-year period studied, and psychosocial and health factors measured at the initial interview predicted service use even 19-22 years later. Controlling for factors identified as likely causes of gender-related differences in healthcare utilization, gender remains an important predictor of medical care use before and after removing sex-specific utilization. In addition, the consistent predictive ability of attitudinal and behavioral factors, combined with the finding that health knowledge did not predict utilization, indicates that efforts to help patients assess their service needs should target the attitudinal and behavioral factors that vary with gender, rather than health-related knowledge alone.