A potential use of the National Death Index for postmarketing drug surveillance.

A particularly difficult problem for both the Food and Drug Administration and the pharmaceutical manufacturer is evaluation of the importance of spontaneous reports of suspected drug-associated fatalities. These reports are rare, and usually no accurate denominator data on drug use exist. This article proposes that the National Death Index be used to calculate mortality rates for selected drugs as part of the postmarketing surveillance efforts of the government and manufacturers. When hypotheses are generated from spontaneous reports and/or the mortality data, additional studies can be conducted on the cohorts that were identified for mortality follow-up.