Extraordinary claims require extraordinary evidence.

We thank Light and Warburton (2005) [hereafter LW] for providing a context for us tofurther discuss and clarify our studies on the cost of new drug development (DiMasi et al.,2003[hereafterDHG];DiMasietal.,1991b[hereafterDHGL];Hansen,1979).Wedisagree,however, with their analyses, comparators, and conclusions. Although LW suggest, throughtheinvocationofselectedreportedcostfiguresandotherstatistics,aswellasquestionsaboutwhat may or may not have been included in our analyses, that new drug R&D costs arereally substantially lower than are our estimates, their central criticism is that some of theunderlying data are proprietary, and they claim, unvalidated. However, they do not address,or even acknowledge, the numerous validations of our results obtained from alternative datasourcesandanalysesthatwerereportedinDHGL(pp.131–132,137–140),inUSCongress,Office of Technology Assessment (OTA) (1993, pp. 54–67), and in DHG (pp. 176–180).We cross-checked our data against data from a variety of sources, including govern-ment healthcare statistics, other investigators’ analyses, data from pharmaceutical firms ondrugdevelopmentactivitiescompiledbythird-partyvendorsforuseasbusinessintelligencetools,internalchecksagainstourownlong-standingdatabasesofdrugdevelopmentmetrics,published industry-level trade association data, and audited company financial statements.The validations vary from checks on the comparability of key characteristics of the investi-gational compounds of survey and non-survey firms to data collected and reported by others