The support for scientific investigation in biomedicine depends in part on the adoption of new knowledge into medical practice. We investigate how a technological advance, in the form of a large and influential 2010 randomized controlled study, changed physician practice in statin (a medication used to manage high cholesterol levels) prescribing. We analyze data representative of the Italian population for the period 2003-2014. Our analysis accounts for possible nonrandom sorting of patients into treatment. We show that both doctors and patients responded promptly to this technological shock, changing the mix of patients who received therapy, drug dosing, and frequency of testing for side effects, as well as patient adherence to therapy. The results show that investments in scientific knowledge can rapidly diffuse into practice in professions where continuing education is the norm.