apy drug regimen increased from $127 in 1993 to $36,300 in 2005. There has been substantial innovation in the treatment of colorectal cancer patients over the last decade.3 Five new drugs were approved by the Food and Drug Administration (FDA) for the treatment of colorectal cancer between 1996 and 2004, and these drugs collectively had an 86 percent market share by the third quarter of 2005.4 We examine whether the substantial increase in spending associated with phar- maceutical innovation in the treatment of colorectal cancer has been worth it. Specifically, we estimate a price index for colorectal cancer drugs for each quarter between 1993 and the first half of 2005 that takes into consideration the quality (i.e., the efficacy and side effects as reported in clinical trials) of each drug on the market and the value that oncologists attach to drug quality. We estimate a quality-adjusted price index using methods developed by Berry (1994), Berry et al. (1995), Berry and Pakes (2007) and Nevo (2003). These tech- niques have been used to estimate the welfare effects of new automobiles (Petrin (2002), computers (Pakes (2003), Song (2007), Song (2005)), and breakfast cereal (Nevo (2003))), but few contributions apply them to medical markets. Trajten- berg (1990) focused on the value of innovation on medical devices, specifically CT scanners, and provided various price indices. Cleanthous (2004) and Conti (2006) study the value of innovation for anti-depressants, however, neither of them constructs a price index and their estimation methods do not address the implications of their functional form assumptions in the error term. The first step in constructing the price index is to estimate oncologists’ demand for colon cancer regimens, which we argue is a function of the observed and unobserved quality of each regimen, as well as the price a physician must pay to acquire the regimen. An observation in this estimation is the market share of each regimen for each quarter. The second step is to calculate the equivalent 313.2 percent of drugs entering Phase 1 trials between 1998 and 2004 were targeting cancer according to
the PharmaProjects data base. 4Market share data are from IntrinsiQ.