To the Editor.
—Dr McClellan and colleagues1 unfortunately misinterpret the results from the Health Care Financing Administration administrative hospital discharge database used for the evaluation of invasive coronary diagnostic and therapeutic procedures. The analytic techniques they have applied cannot determine the efficacy or effectiveness of cardiac catheterization or whether they are overused or underused. The problems with their methods of data analysis were discussed in a classic article by Sheps in 1958.2 In that article, Sheps described the problem of evaluating the efficacy of polio vaccine. For example, consider 100 000 individuals who were given polio vaccine and 100 000 controls. The subsequent rate of clinical polio was 57/100 000 in the controls and 16/100 000 in the intervention group. The difference of 41/100 000, or 0.04%, was so small that policymakers today would probably be advised by outcomes research to reconsider vaccination programs. Now, if we take
Kuller LH, Detre K. Catheterization and Mortality in Elderly Patients With Acute Myocardial Infarction. JAMA. 1995;273(17):1331-1332. doi:10.1001/jama.1995.03520410025013