Product Differentiation, Consumer Learning and the Value of Me-Too Drugs

Many new drugs are similar to existing products. Consumers’ benefits from these “me-too drugs” depend on the correlation of efficacy and side effects across products within patients. We estimate a structural model of pharmaceutical demand that allows for heterogeneity in match quality between consumers and drugs and estimates the covariance in match values across products. We also incorporate consumer learning about the stochastic match qualities with specific drugs. Advertising may play an important role in consumer information and learning. We employ 2007-2008 administrative claims data. We examine patients newly diagnosed with osteoporosis who initiate treatment in 2007 or 2008. We find that naive models, without consumer heterogeneity and learning, grossly underestimate demand elasticities. Policy simulations measure consumer welfare with alternative choice sets.