Thursday, July 20, 2006

Russ Whitehead doesn't know Daniel

I would have preferred to label this entry, "Russ Whitehead doesn't know Jack," but we'll have to stick with Daniel because Daniel Levy and his fellow investigators are currently running the continuing Framingham Heart Study, originally started in 1948. The current obsession in Washington education circles is scientifically based research, because it's mentioned in the No Child Left Behind Act as a priority. Russ Whitehead has interpreted that to mean that randomly-controlled experiments (RCTs) are the gold standard of research.

Please tell that to my doctor. When I go in for my occasional checkups, he takes my vitals and plugs them into a well-known formula to predict 10-year risks of heart attacks. This formula comes from the Framingham Heart Study, which has contributed thousands of person-years of exposure to various heart risks from over 10,000 participants. Dr. Whitehead, may I introduce you to the term epidemiology? It's not randomly controlled, but it happens to be the best way to identify patterns of risks and consequences in a real, living population.

RCTs are great to test initial therapies in medicine. And I wish that we could invest more money into the type of validation for educational reforms that therapeutic drugs have to go through, not only before being approved for use in human beings but also in identifying side effects and following up on the consequences of broader uses. But RCTs only test narrow questions: does A work better than B in a certain context? They are only a part of medical research, and one could make a pretty good argument that they have been less important to increasing life expectancy than good public-health research. (Reality-check question: Who was the first person to single-handedly stop a cluster of cholera cases, and how?)

Of everyone invited to a 2002 Ed Department seminar on what scientifically based research is, only Stephen Raudenbusch acknowledged the silliness of saying only RCTs are science or even a gold standard. Or, if RCTs are a gold standard in medicine, epidemiology is the platinum standard. Shame on everyone who pretends otherwise.

No comments: