This study investigates the extent to which sociodemographic, medical and psychosocial factors measured in 1967 among 1,649 American-born Israelis predict their physical and emotional health status 26 years later. In 1993, mail questionnaires were completed by 673 (40.8%) of the 1967 respondents. Evidence was obtained regarding the death of 204 (12.4%) of the 1967 respondents. The 1967 predictor variables included: sociodemographic characteristics, self-assessed health and medical risk factors, and psychosocial resources. The 1993 outcome variables included a multidimensional measure of health, physical functioning, and emotional well-being. Bivariate and multivariate statistical analyses were used to determine the predictive power of the variables measured in 1967 on the 1993 health outcomes, controlling for potential confounders. These analyses showed that the deceased in 1993 were older than the respondents; in 1967 they had reported more health problems and had more medical risk factors; they also had less formal education, were less likely to be married, and had lower adjustment scores than the respondents. Among the 1993 respondents, the addition of psychosocial variables to the explanation of health outcomes, controlling for sociodemographic and medical risk factors, contributes strongly (13%) to emotional well-being. Their contribution to general health and physical functioning is weak (2 to 3%), but statistically significant. These results show the importance of a multifactorial approach to the long-term prediction of health and illness outcomes.