By Mary Chitty
October 28, 2008 | CAMBRIDGE, MASS—Drug safety has been much in the news recently, and it was the topic of a panel discussion at MIT's Center for Biomedical Innovation held a two day stakeholder meeting Oct. 23-24 in Cambridge, Mass.
Moderated by Garry Neil, corporate VP, Corporate Office of Science & Technology, Johnson and Johnson, the discussion identified some of the conflicting factors surrounding drug safety and looked to other industries for advice.
Panelists included Rachel Behrman, FDA, associate commissioner for Clinical Programs and Mark McClellan, Director of the Engelberg Center for Health Care Reform, Brookings Institution, and commissioner of the FDA 2002-2004 participating by videoconference from Washington D.C. Glenn Forbes, CEO Mayo Clinic; Richard Platt, Harvard Medical School and Harvard Pilgrim Healthcare; Jean Slutsky, Director Center for Outcomes and Evidence, Agency for Healthcare Research and Quality; Richard Larson, Director, Center for Engineering Systems Fundamental MIT; Nancy Leveson, Professor of Aeronautics and Astronautics, MIT; and Deborah Nightingale, Professor of the Practice, Aeronautics and Astronautics and Engineering Systems Division, MIT.
Several panelists mentioned Peter Drucker, often called the “father of management,” one noting his description of health care delivery as one of the most complex organizations in human history. One of the chief challenges is to use routinely collected data to learn what we don't know. There are many ways for drugs to be unsafe.
Clinical data has the chance to provide useful cohorts for better modeling of patient populations of diseases and disease progression. Systematic data on 100,000 people is possible within several years. There should be a much greater capacity to adjust case mixes and truly compare like with like, to be better able to understand outcomes.
But confounding factors abound. Data integrity is a huge issue. Database interoperability is challenging, while unmanaged expectations are an even bigger problem. Standards are hard to come by, and even more difficult to change. The very ways in which we think about doing research are changing. Health care has not been configured for systematic investigation; methodological challenges are huge. Which are the truly strategic metrics and incentives?
Several panelists hailed from MIT’s engineering and aeronautics divisions, and they seemed to have many useful lessons for health care. Panelists spoke of the importance of fixing systems not just symptoms. Sophisticated statistical analyses can be transferable from other disciplines. As airline accidents have decreased, greater attention has been paid to careful examination and definition of near misses. Perhaps health care can adopt a similar practice.
The greatest challenges may revolve around sharing, trust, and transparency. Starting to talk about these issues and to figure out ways to maintain the conversation seems a very important beginning.