Charles M. Beasley, Jr and Roy Tamura: What We Know and Do Not Know by Conventional Statistical Standards About Whether a Drug Does or Does Not Cause a Specific Side Effect

Barry Blackwell’s comment

          Charles Beasley and Roy Tamura have produced 30 pages of statistical wizardry to demonstrate that the pharmaceutical industry, if it was so motivated, might be able to demonstrate whether or not a new drug does or does not cause a specific side effect or Adverse Drug Reaction (ADR), distinguished from an Adverse Event (AE), that is coincidental with a study but not due to the drug.

          Their purpose is to illustrate the large sample sizes and elaborate statistical techniques to accomplish this task. Over the course of five months (November 2018 to March 2019) they published an outline of their thesis in seven different episodes including an introduction, comments, potential sampling errors, proof of or absence of presence of ADR, the real incidence of AE and the requirements to undertake massive and costly measures that might be imposed by regulatory agencies.

          Lacking the statistical weaponry to digest, critique or refute such an impressive body of work, now an e-book, I would normally not attempt to do so were it not for the fact that authors view their enterprise as a reposte to comments made in and subsequent to my essay on corporate corruption in the pharmaceutical industry (Blackwell 2016).

          My rebuttal rests more on logic rather than statistical wisdom. To begin with their “definition of terms” states that whether or not an AE can be distinguished form an ADR depends on whether there is “reasonable evidence.” But they note: “To the best of our knowledge this has never been operationally defined or even qualified by any regulatory entity or drug safety organization.” They cite two international organizations and the American FDA.

          As noted in my essay this is hardly surprising. The FDA derives 50% of its budget from industry payments for approval of a new drug application (NDA), a requirement imposed by the Republican Reagan administration (1980-1988). This creates a massive conflict of interest, encouraging the FDA to turn a blind eye to imposing costly sample sizes and elaborate statistical techniques that might and has discouraged industry innovation in the contemporary me-too era.

          So, where is the evidence that any pharmaceutical company has attempted to accomplish a costly endeavor it is not obligated to initiate? There are no contemporary examples cited. It is less expensive to cynically build the cost of potential class action lawsuits into the price of a drug before a missed ADR makes its presence known and then have court settlements impose silence on the victims.

          Charles Beasley has spent a distinguished and blameless career in industry. This body of work embellishes his statistical ingenuity and if his former employers were to use it they might salvage a glimpse of the integrity their greed has consumed.

Reference:

Blackwell B. Corporate Corruption in the Pharmaceutical Industry. inhn.org.controversies. March 23, 2016.

July 25, 2019