In 2006, Kim Popovits, President and COO of Genomic Health, Inc., gave a talk at Stanford, in which she, among other things, mentioned that out of a hundred women that were subjected to chemotherapy for breast cancer only four would get any benefit from the treatment (15 min). The problem was that nobody knew which four, so women would take their chances and ask doctors for chemo.
After several years of work, Popovits' company, using a standard(!) diagnostics tool, developed a genetic test to identify those whom the chemotherapy had the greatest chance to help.
Key insights from the talk:
- a good example of a solution to a detection problem (genetic test) that leads to a much improved control sub-system, and therefore to a high level of system efficiency; (see also my previous post).
- application of an existing solution to a new high value problem that emerged due to the rapid growth of the treatment system; leverage a large "idle" database of tissue samples and medical histories; (another example of storage as a sign of an impending major change in system evolution);
- importance of the team, i.e. complementarity of different skill sets;
- the need to focus on the market need;
- decision criterion: "choose what's good for the patient". on one hand, creates the greatest benefits for the person; on the other hand, proliferates expensive treatments that bankrupt the society as a whole (e.g. Medicare is projected to become insolvable within the next decade);
- emergence of highly personalized drug treatments.