Portfolio Decisions in Early Development
- First Online:
- Cite this article as:
- Herschel, M. Pharm Med (2012) 26: 77. doi:10.1007/BF03256895
Portfolio decisions in early development happen under much uncertainty. Particularly when a novel target or treatment method is being explored, early negative clinical results may lead to the discontinuation of compound development, or even of an entire branch of discovery research. Although we can find examples where ‘failed’ concepts have been picked up later and shown to work, there is no way to estimate how large the percentage of false negative discontinuations may be as this type of information is not routinely added to publicly accessible databases; any analysis therefore is limited to specific articles and personal experience. Factors that favour false negative trials are known, and some of them can be influenced. In many cases, intelligent programme and clinical trial design can prevent false negative results. Well designed dose-finding studies and comparisons against both a placebo and a gold-standard treatment arm play a major role in achieving reliable data. Apart from following a more intelligent way of performing trials, the inherent mechanisms of human decision making need to be understood. Many of those decisions are not purely rational. Therefore, the findings of Kahneman and Tversky are very much applicable to early termination decisions. Thus, there are not only ways to prevent false negative clinical trials, but also clues to the psychology/mental frameworks that lead to erroneous conclusions in development. Awareness of such thinking may reduce the inclination to abandon an entire area of research based on only a few experimental data. The development of diagnostic tests can be used as an analogue to the decisions to be made in early clinical research, as such decisions can be classified according to false positives or false negatives, and so on, depending on the quality of the available information. It is hoped that the analysis of these early portfolio decisions will be based on data from public domain databases such as http://www.clinicaltrials.gov in the future.