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 (Adverse Drug Reaction)Overview
Barry Blackwell’s comments on Charles Beasley’s reply to Edward Shorter’s comment
Olanzapine and diabetes mellitus; evolving data illustrating the difficulties in identification of adverse drug reactions
Background
When Tom Ban, at the suggestion of Ned Shorter, asked me to reply to the recent substantial INHN posting of Beasley and Tamura’s statistical methodology for tackling the difficulty of identifying adverse drug reactions, initially in general but at Shorter’s suggestion focusing on Olanzapine, (marketed by Eli Lilly as Zyprexa for schizophrenia), I entertained serious reservations.
It is 10 years since I retired, for the third time, and although I am a trained psychopharmacologist I am neither a statistician nor an endocrinologist although I held academic professorships in Psychiatry, Pharmacology and Medicine.
But I decided to accept in the context of perspectivism; OED: “The philosophical theory that knowledge of a subject is inevitably partial and limited by the individual perspective from which it is viewed.”
This perspective includes an historical point of view derived from more than 10 years working with Tom Ban, first on the Oral History of Neuropsychopharmacology: The First Fifty Years (2011), and since 2013 as editor of Biographies and Controversies on the INHN website.
In that context I posted a critical essay on “Corporate Corruption in the Pharmaceutical Industry” in which Charles Beasley spent a distinguished career as the lead biostatistician for Eli Lilly. Charles was working on Olanzapine and its side effects until 2002 when higher authority ordered him to cease. His recent postings on INHN are the product of his work post retirement.
The Historical Context
Beasley’s work on INHN includes 10 separate postings between November 29, 2018, and April 4, 2019, (Pages 1-31 of the Ebook) which summarize the approach he developed at Eli Lilly intended to identify adverse drug reactions. On April 25, 2019, Edward (Ned) Shorter made a brief comment congratulating the authors on their “commitment to the high ground of science” but challenging them to provide a specific demonstration of the method with regard to Olanzapine and the risk of diabetes. Charles Beasley responded with a 56-page posting (pages 4-60 of the Ebook) which is the topic of my comments.
Significant Historical Findings
For retired folks who no longer keep up with academic sources of wisdom Google is an appealing contemporary substitute. Type in Olanzapine and 3,000,000 postings are recorded. Limit the search to Olanzapine and Diabetes and the number falls to just under a million, 986,000. Looking for a cogent publication the most relevant and frequently cited (516 times) was a study published in the British Medical Journal titled “Assessment of the independent effect of Olanzapine and Risperidone on the risk of Diabetes among patients with schizophrenia; population based nested case control study” (Koro, Fedder, L'Italien et al. 2002).
This study was derived from the UK General Practice Research Data Base. It involved 19,637 patients diagnosed and treated for schizophrenia, compared to 2,696 controls. The conclusions were, “Olanzapine is associated with a clinically important increased risk of diabetes. Patients taking Olanzapine had a significantly higher risk of developing diabetes than both non-users of antipsychotics and Risperdal (risperidone). Taking Risperdal had a non-significant risk compared to both non users and those taking conventional antipsychotics.”
There is little doubt this was a seminal study in a refereed international medical journal that had an impact on regulatory and labelling practices. On October 17, 2003, the FDA issued a warning letter to manufacturers about the risk of diabetes followed the next year by practice guidelines from the American Diabetes Association and the American Psychiatric Association. These required Weight and BMI measures at every visit and a fasting blood sugar and blood lipid level at week 12 and annually. In March 2004 Eli Lilly added a warning statement to the labelling of Olanzapine describing the increased risk of hyperglycemia and diabetes.
The impact on research and regulatory procedures.
The timing of the BMJ study in 2002 coincides with top management at Eli Lilly requiring Charles Beasley to cease work on Olanzapine although that study is not cited in the 34 references which include 22 published during or after 2002. I speculate that the timing of that publication served to render moot any further need for costly in-house statistical analysis. In Section 5 of the statistical approach Charles outlines “A hypothetical study to resolve the uncertainties.” But he concedes that search for co-morbid variables influencing risk of diabetes pale into insignificance now that the severity and prevalence of the disorder itself is known. He speculates that a manufacturer would be unwilling to devote profits to such a large and expensive enterprise.
Working for Industry; a personal perspective
In 1968 at the age of 34 after completing my psychiatric training at the Maudsley Hospital in London I migrated to America (still the land of opportunity) to accept the position as Director of Psychotropic Drug Development at the Merrell pharmaceutical company in Cincinnati. They had recently emerged from the fiasco of marketing thalidomide to pregnant women as a safe hypnotic. This triggered Congress to enact legislation to empower the FDA to develop regulatory procedures for the development of new drugs. Merrell, like other pharmaceutical companies hope to benefit from this new lucrative field of research.
Early on in the evolution of modern psychopharmacology I soon realized I was better suited to research then marketing so, after two years, I returned to academic life as a Professor of Psychiatry and Pharmacology at the University of Cincinnati.
I believe that Charles Beasley was a person of similar temperament and capability who in his mid to late career became snared in the 1980’s industry transition from credible research on innovative compounds to the ingenious marketing of me-too compounds coupled with a steady erosion of ethical standards in top management focused on profit who established tight controls over talented scientists like Charles Beasley. I believe that the statistical methodology he describes had the capacity to fulfill his personal and ethical goals as well as his desire to see it posted on INHN post retirement which does credit to his distinguished career in industry.
References:
Beasley C, Tamura R. 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 (Adverse Drug Reaction). inhn.org.ebooks. November 21, 2019.
Koro CE, Fedder DO, L'Italien GJ, Weiss SS, Magder LS, Kreyenbuhl J, Revicki DA, Buchanan RW. Assessment of independent effect of olanzapine and risperidone on risk of diabetes among patients with schizophrenia: population based nested case-control study. BMJ. 2002; 325(7358):243.
November 28, 2019