Joel Kallich of Big Health Data wrote me:
Over the last 30 years, the government has been reluctant to release identifiable information, first on an institutional level, then physician level, and at the patient level. These restrictions have slowly been peeled back and the light of day has been shined on this information. But now it seems we’ve taken a step backwards.
About this, I would like to point out two salient facts:
- There has never been a case, to my knowledge, in which a researcher has identified a patient publically or provided identifying information to any organization that did not already have access to it. And Nicholas Bagley pointed out that to me that, when regulations on disclosure were drafted in 1976, it was noted that there had been “no instances of abuse on the part of persons acquiring patient identifying information” [40 Fed. Reg. 20537].
- The provision of ‘identifiable data’ — including physician data, patient treatment date data, unique patient identification numbers — has generated significant and important research that has in turn led to many initiatives to improve patient health, and the overall efficiency of health care delivery in the US.
My prime example of the later is the discovery of Medicare patient readmission issues. Anderson and Steinberg’s 1984 NEJM publication reporting a 22% Medicare rehospitalization rate would not have been possible without a data set that provided unique identification numbers for institutions and patients and discharge and admission dates. In fact, the ability of MedPAC researchers and others to use and refine this analysis led to changes in the Medicare program included in the ACA, accounting for an estimated $14B of preventable readmissions. None of this would have been possible without detailed, patient level information provided to independent health services researchers.