OBJECTIVE:To examine 30 day survival after acute myocardial infarction as an outcome indicator, and explore the effects of adjusting for available prognostic factors such as age, sex, co-morbidity, deprivation, and deaths outside hospital. DESIGN:Cohort study. SETTING:The Scottish Record Linkage System was analysed. This national data-base links inpatient data to death certificate information for a population of 5.1 million. SUBJECTS:All 40,371 admissions to hospital with a principal diagnosis of acute myocardial infarction, plus all 18,452 deaths outside hospital with a principal cause of death registered as acute myocardial infarction (ICD9 code 410) during 1988-1991. MAIN OUTCOME MEASURES:The outcome event was death from any cause, within hospital or elsewhere, within 30 days of admission. RESULTS:During 1988-1991, 30 day survival after acute myocardial infarction was 77% in 40,371 hospital admissions, but only 53% when 18,452 acute myocardial infarction deaths in the community were included (a population-based outcome indicator with many advantages). Using logistic regression at an individual patient level, the odds of dying within 30 days effectively doubled for each decade of age (odds ratio compared with patients aged under 55: 2.3 aged 55-64, 4.4 aged 65-74, 8.2 aged 75-84, 12.0 aged 85 plus); were marginally higher in females than in males (odds ratio 1.07); were almost doubled in patients with a history of previous infarction, coronary heart disease, or other heart disease, and were also significantly increased in patients with circulatory disease, respiratory disease, neoplasm, or diabetes. Socioeconomic deprivation had no significant effect. Marked variations in survival between different hospitals and health board areas persisted, even after adjusting for the above prognostic factors. CONCLUSION:One month survival after acute myocardial infarction could be a useful means of measuring outcome of hospital care. There was important geographical variation in one month survival. These differences could be accounted for by variations in referral, admission, diagnosis, definition, and coding. These variables merit further research and local clinical audit before one month survival after acute myocardial infarction can be reliably used for detecting differences in quality of care. In addition, it would be essential to take account of infarct severity.