INTRODUCTION The care of the elderly presents serious challenges to general practice. In 1979, the first author took over the care of a general practice in Scotland where 21% of registered patients were elderly. This resulted in a high workload and prompted research into how this might be mitigated. AIM To measure serial tests of T-cell function in these individuals in order to identify those whose immune response was impaired and assess the effect of this in a long term follow up. METHODS This research comprised two phases. In the assessment phase (1979-82), patients were invited to have a 3-monthly visit from a research nurse where clinical measurements were made and blood taken for immunological tests of lymphocyte proliferation after culture with phytohaemagglutinin (PHA). For each patient, all records were surveyed and problems identified. In the follow-up phase (post 1982), all deaths were assessed with complete life-long follow up. RESULTS Of 405 people originally invited to participate in this research, 314 (78%) agreed and 246 (153 female, 93 male) entered the follow-up phase and were followed for 36.5 years. Factors significantly associated with lower survival were age, male sex, diastolic blood pressure, current smoking and poor immune function, as demonstrated by the percentage of negative responses in at least six PHA tests. Considered in four groups by percentage of failing tests, the lowest group had a life span 4 years shorter than the highest (P<0.01). The four groups did not differ significantly in general practitioner workload, diagnosed problems or causes of death. DISCUSSION Poor cellular immune function was associated with poor survival over lifetime follow up of >30 years. A sensitive, specific and longitudinally consistent measure of T-cell function is required to predict who may be at risk of poorer survival within our practices.