Trusheim and Berndt provide an excellent overview of the economics of stratified medicine. I have reviewed the topic before and specifically in this post. An ideal companion diagnostic perfectly identifies and distinguishes treatment responders from those who will not.Unfortunately, in practice no diagnostic performs ideally. All diagnostics experience some level of error. In the case of a companion diagnostic, some patients will receive false positive results, scores indicating they will respond, but will not when treated. Other patients will receive false negative results, scores indicating they will not respond, but if treated they would.
So what? The key issue is about the setting of the cut-off that will define the size of the market:
Under competition, three essentially identical drugs may receive dramatically different labels, incremental cost-effectiveness ratio (ICER) justified pricing, and market positioning depending on their stratification approach. It appears superior to use an imperfect biomarker to none at all. It is less obvious whether patients, payers and firms prefer the same cut-off values for the companion diagnostic, or even whether each stakeholder a priori prefers the high, low or perhaps some other CDx cut-off value.The implications for regulating and financing of companion diagnostics are immediate. In Europe current legislation is as old as of 1998, the year that Herceptin entered into the market and the begining of stratified medicine. New regulation is still pending and there are no perspectives about a unified approach to lab tests assessment. European regulator is still on vacation.
The competing development teams may face a version of the game theory ‘prisoner’s dilemma’ in which the optimal result for patients and all firms would be to select a low or mid companion diagnostic cut-off value but the advantages of a potentially differentiating high efficacy claim may drive developers to select a high cut-off value. If all choose this approach, overall value may be reduced with many patients excluded from treatment. But the potential advantage of a higher cut-off value may prove too alluring, or the fear of a competitor selecting one, may drive all to do so. Each situation will depend on the specific facts of the indication, therapeutic, companion diagnostic and competitors
