The goal of a standard clinical study is to assess the mean or median effectiveness and side effects of drugs through randomized clinical trial experiment. In this paper we argue that, it is important to understand how physicians and patients evaluate the effectiveness and side effects. Second, it is also important to understand the extent of heterogeneity in these outcomes across patients, especially when physicians and patients are risk averse in choosing treatments. Finally, using data from a market environment has advantages over clinical trials since in the former patients and physicians are faced with alternative drugs they can choose from while exposed to marketing efforts from pharmaceutical companies.In this study we use a physician panel data in the Erectyle Desfunction (ED) category to examine how the mean effectiveness and side effects as well as the heterogeneity in these treatment outcomes affect prescription decisions. To separately identify the patient evaluation of effectiveness and side effects, we augment the observed prescription choices with a unique data on self-reported reasons for switching in our estimation. Two sources of uncertainty in the evaluation of the effectivness and side effects of drugs are accounted for in our model. First, effectiveness and side effect of a drug can vary across patients and second even the mean effectiveness and side effects may be unknown to physicians and patients, especially for those drugs that are new to the market. We allow physicians to learn about the effectiveness and side effects of new drugs based on patient feedbacks and detailing visits from pharmaceutical companies. Results show that though the two new drugs, Levitra and Cialis, have higher mean effectiveness and lower mean side effects than the exiting drug, Viagra, their competitive weakness is the larger heterogeneity in effectiveness and side effects, respectively. Detailing visit is effective in reducing the uncertainty of effectiveness but much less effective in reducing the uncertainty of side effects. We also find that for both new drugs the informative role of detailing visit is more important than its persuasive role in influencing the prescription utility.