Relating Subjective and Objective Pharmacovigilance Association Measures - Clustering Challenges in Biological Network - page 279

Chapter 15
Relating Subjective and Objective Pharmacovigilance Association
Measures
Ronald K. Pearson
ProSanos Corporation
225 Market St., Suite 502, Harrisburg, PA, USA, 17101
[e-mail]
The field of pharmacovigilance is concerned with the detection and interpretation
of associations between drugs and adverse medical events that may be related to
the use of those drugs. These assocations can be measured in various ways, and
this paper considers five: two are aggregate statistical measures derived from an
entire adverse event database, two are case-specific objective measures, and one
is a subjective measure related to the way adverse events are reported. Exami-
nation of the available data suggests that these measures are all interrelated, but
in a complicated manner. This finding motivates the use of cluster analysis to
explore these relationships, with the ultimate objective of constructing an index
of blame that quantifies the tendency for some drugs to be subjectively blamed
for adverse events even in the absence of objective evidence for an association
with those events.
15.1. Introduction
Pharmacovigilance analysis attempts to understand and quantify the relationship
between drugs and adverse events that they may have caused or exacerbated. In
the U.S., the primary data source for pharmacovigilance analysis is the FDA's
Adverse Event Reporting System (AERS) database [19], which is organized by
Individual Safety Reports (ISR's). Each ISR describes a suspected adverse drug
reaction experienced by a single patient, lising the drugs that patient was taking,
the adverse reactions they reported, and a limited amount of additional data (e.g.,
patient age, gender, reporting source information, etc.). Each ISR typically lists
several drugs and several adverse events, a fact that plays an important role in the
results presented here.
Since the AERS database is built from spontaneous reports, there is an element
of subjectivity inherent in the decision that a particular combination of drugs and
281