BACKGROUND It has previously been shown that the increased use of therapeutic intervention may not reduce patient fatality if there is a simultaneous increase in case severity. The present study was designed to extend the relationship between case severity and therapeutic interventions to long-term survival in the same study population. OBJECTIVE To compare five-year survival of patients discharged after acute myocardial infarction from 1984 to 1988 and from 1989 to 1993, and to evaluate possible reasons for survival differences. METHODS The present study was population-based. Survival time was determined by record linkage into the Canadian Mortality Database. Association of five-year survival with patient characteristics, in-hospital treatment and discharge medications was assessed by logistical regression analysis. Case severity was calculated as the probability of death within five years, given the patient profile and excluding any interventions. RESULTS Between the two study periods, most patient characteristics and treatment intensity changed, but case severity for the study population remained constant. Five-year survival improved from 74.8% to 79.2% (P(chi2)=0.001). The improvement was adequately described by the combination of changes in patient profile and treatment without residual period effect (P(goodness-of-fit)=0.752). The treatments significantly associated with five-year survival were coronary artery bypass graft surgery (OR 2.74; 95% CI 1.86 to 4.05), percutaneous coronary intervention (OR 2.63; 95% CI 1.67 to 4.14) and thrombolysis (OR 1.98; 95% CI 1.50 to 2.62) during admission, as well as acetylsalicylic acid (OR 1.39; 95% CI 1.15 to 1.68) or beta-blocker (OR 1.60; 95% CI 1.34 to 1.92) prescription at discharge. CONCLUSIONS Changes in patient profile did not affect long-term prognosis; instead, treatment modalities accounted for the observed improvement in five-year survival.