Date of Award

January 2020

Document Type


Degree Name

Medical Doctor (MD)



First Advisor

Joseph S. Ross


The U.S. Food & Drug Administration (FDA) is the government agency tasked with protecting public health by supporting safety and efficacy of drugs prescribed to patients. One of FDA’s key tools to ensure drug safety is through updating product labeling with a boxed warning (also known as "black box warning (BBW)"), its strongest action short of a drug recall. Despite the central role that BBWs play in supporting safe use of pharmaceuticals, no study has comprehensively characterized BBWs issued for drugs after initial regulatory approval. Additionally, several studies have shown that the impact of these warnings on prescribing has been inconsistent, and one understudied reason may relate to insurer formulary coverage.

To address these gaps in knowledge, we conducted two studies. The goal of Study 1 was to evaluate the characteristics of post-market BBWs and the drugs for which they were issued. We identified all post-marketing BBWs from January 2008 through June 2015 listed on FDA’s MedWatch and Drug Safety Communications websites, and used each drug’s prescribing information to characterize these BBWs with respect to pre-specified BBW-specific and drug-specific features. We found 111 BBWs issued to drugs on the US market, of which 29% (n=32) were new BBWs, 32% (n=35) were major updates, and 40% (n=44) were minor updates. New BBWs and major updates were most commonly issued for death (51%) and cardiovascular risk (27%). The new BBWs and major updates impacted 200 drug formulations over the study period, of which 64% were expected to be used chronically and 58% had available alternatives without a BBW.

In Study 2, we aimed to examine Medicare prescription plan drug formulary changes after post-market BBWs, and to characterize BBW and drug characteristics that may affect these changes. In this cohort study, one control drug was matched to each drug from Study 1 that was expected to be reimbursed through Medicare Part D. Of 101 new BBWs and major updates to pre-existing warnings impacting 68 unique drug formulations, the mean percentage of formularies providing unrestrictive coverage was 65.4% (95% confidence interval [CI], 59.6% to 71.2%) pre-warning, 62.6% (95% CI, 56.3% to 68.9%) at least one year post-warning, and 61.9% (95% CI, 55.4% to 68.5%) at least two years post-warning. Compared to matched controls, there was no significant change in unrestrictive coverage at least one year (-3.1%, 95% CI, -14.6% to 8.4%) or two years (-4.6%, 95% CI, -16.3% to 7.0%) after these warnings. No BBW or drug characteristic was associated with a significant change in restrictiveness.

Together, the results of these studies suggest that new BBWs and incremental updates to existing BBWs are frequently added to drug labels after regulatory approval, and that these warnings commonly contain vital clinical information about which clinicians should be aware. Further, Medicare formularies did not restrict coverage to drugs receiving BBWs, even when the warning was for a serious risk such as death or in a commonly used chronic medication. Formulary restrictions therefore present an underutilized mechanism to reduce use of potentially unsafe medications.


This thesis is restricted to Yale network users only. This thesis is permanently embargoed from public release.