Module 5
Lesson 1
Welcome to this introduction to statistical reasoning and algorithms in pharmacovigilance!
As you may remember from our module on signal detection and causality assessment, expert review
is the single most important step in pharmacovigilance analysis and statistical reasoning can play an
important role here. Especially for those signals that are driven not by the individual strength of one
or a few reports but by their sheer number.
Can we reasonably expect this many reports on this drug and this adverse event by coincidence? Or
if we suspect another cause, such as a co-reported drug, does data lend support to that alternative
explanation?
So, statistical reasoning brings value in challenging and enhancing our thinking ideally in interplay
with the clinical and pharmacological reasoning.
If we turn our focus to statistical algorithms, they bring value in much the same way by raising
awareness and pointing us to patterns worthy of further consideration and review. We can use them
to look for reports that are suspiciously similar and may actually refer to the same case, so-called
duplicates. We can use them to look in longitudinal medical records for patterns of time to onset
that suggests that a certain drug increases the risk of an adverse event or decreases that risk.
We can hopefully use them to identify sensitive information in free text, so that we can anonymize
the valuable case narratives and make sure that we can share them for the purpose of
pharmacovigilance without compromising patient confidentiality.
Where algorithms have had the greatest impact in pharmacovigilance though, is as an early step in
signal detection to focus our expert review when there are more reports than we can look at every
incoming report, or if there are more reports than any one individual can keep in mind at the same
time.
A specific class of algorithms used for this purpose that is now used by many organizations across
the world is disproportionality analysis and this is what you'll hear more about next.
Enjoy!