FDA researchers find promise in custom medical queries to identify safety signals
Current and former US Food and Drug Administration (FDA) researchers say an initial analysis to evaluate potential drug and biologic adverse events (AE) using a new terminology system has shown promising results. While further research is needed, they said the system could be used to detect safety signals for product labeling.The article, written by several current and former FDA officials from the Center for Drug Evaluation and Research (CDER) and the Center for Biologics Evaluation and Research (CBER), and published in Drug Safety this week, examines a new set of medical queries to evaluate AEs.
For years, drug and biologics sponsors have used standardized Medical Dictionary for Regulatory Activities (MedDRA) terms to list Preferred Terms (PT) to help researchers identify potential AEs in clinical trials. However, as the study's authors point out, subjective use of the PTs may make it harder to detect safety signals.
“In some cases, MedDRA adds specificity to PTs that may be counterproductive with respect to AE analyses,” the article said.
As an example, the authors noted that there are several PTs for insomnia, and researchers may use different terms for the same issue, which may lead to an underestimation of insomnia AEs. To address this concern, FDA’s Office of New Drugs (OND) developed 104 custom groupings of AE terms known as OND Custom Medical Queries (OCMQs) that are intended to identify commonly labeled adverse reactions (AR).
The authors conducted a validation study of 459 clinical trials using OCMQ groupings to identify AEs. They found that the initial analysis successfully identified AEs that can be used to inform the product labeling.
"This validation demonstrated that OCMQs are broadly successful at detecting signals in clinical trial data for drugs that are labeled for ARs represented by the OCMQs," said the authors. "The strength of this validation was the large database of clinical trial safety data used for the validation work.
"An important limitation was that the OCMQs were not tested in the specific manner for which they are intended to be used," they added. "Specifically, when a difference is detected in an OCMQ grouping between control and treatment arms, reviewers are expected to assess the individual AE terms within the OCMQ to determine whether there are a limited number of terms driving the disparity, a more subtle imbalance across many terms, or a combination of both."
While the validation was successful, the authors noted that clinical reviewers recommended minor revisions to the OCMQs to ensure more consistency in the terminology. They also added that they will consider additional validation studies to better understand how OCMQs can be used, including assessing differences in signal detection when using only individual AEs compared to OCMQ groupings, and the use of OCMQs by clinical review teams.
Drug Safety article
