We study the impact of social health insurance on mortality. Using the introduction of compulsory health insurance in the German Empire in 1884 as a natural experiment, we estimate flexible difference-in-differences models exploiting variation in eligibility for insurance across occupations. Our findings suggest that Bismarck's health insurance generated a significant mortality reduction. Despite the absence of antibiotics and most vaccines, we find the results to be largely driven by a decline of deaths from infectious diseases. We present evidence suggesting that the decline is associated with access to health services but not sick pay. This finding may be explained by insurance fund physicians transmitting new knowledge on infectious disease prevention.