Due to the seriousness of adverse drug reactions, life sciences organizations perform something called “signal detection,” which is the process of evaluating potential safety signals. This entails looking at previously known associations between a drug and an adverse event, and looking at previously known drug-related issues that may be either improving or worsening. Not only do companies routinely do this type of analysis, they are obligated to do so under current regulatory guidelines.
This guide analyzes how artificial intelligence – including machine learning – can be used by pharmaceutical and medical device companies to improve the clinical data review and cleansing process.
Signal detection leverages a variety of methodologies, some of which are more comprehensive than others. One evaluation method, proposed by a number of signal detection specialists, is called the SNIP methodology. With SNIP, researchers consider the signal’s Strength, whether the signal is New, the Importance of the signal, and whether the signal can be prevented.
- Strength: This refers to the mathematical or scientific strength. Normally, the statistic used in this market analysis is an offshoot of an odds ratio of Proportional Reporting Ratio (PRR) or Bayesian statistic.
- New: How new is the event-drug relationship? Is it part of a well-established risk-benefit for the drug of interest or is it brand new?
- Importance: Is the ailment causing severe issues within the patient population or is this something that is a little more manageable, from a clinical perspective?
- Prevention: How easy is it to prevent this problem, from a public health perspective?
Various types of evaluation tools, like SNIP, are used to identify signals, regardless of whether they come from spontaneous reports, clinical studies, or scientific literature.
Check back soon for the next post in this series on drug safety: pharmacovigilance. In the meantime, check out this guide on drug safety and register for our upcoming webinar titled “The 5 Most Significant Changes in Argus Safety 8.1.”