The aim of the paper is to disentangle the roles that patients, physicians and technology can have on patient health outcomes using a large and detailed dataset of Italian patients collected by the Italian College of General Practitioners (SIMG) over the period 2001–2006. As our data show the existence of heterogeneity in the time needed to reach an optimal level of health stock, we study this measure of health outcome rather than simply the level of health stock. Limiting our analysis to patients suffering from hypercholesterolemia, the empirical work is based on two different analyses. We first explore whether patients recovering faster exhibit lower hospitalization rates and then we study the determinants of the speed of recovery to a good health status. The results confirm that a 10% increase in the speed of recovery can reduce hospitalization rates by about 1.0%. Furthermore, we show that recovering to a good health status is a multifaceted phenomenon, with technology explaining at best 62% of the combined effect, while patient and physician behaviors share the residual effect. These results are then discussed in terms of policy.