Premium
TESTING FOR ADVERSE REACTIONS USING PRESCRIPTION EVENT MONITORING
Author(s) -
ANDREW J. E.,
PRESCOTT P.,
SMITH T. M. F.,
INMAN W. H. W.,
KUBOTA K.
Publication year - 1996
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/(sici)1097-0258(19960530)15:10<987::aid-sim210>3.0.co;2-g
Subject(s) - estimator , statistics , event (particle physics) , rule of thumb , econometrics , medical prescription , confidence interval , drug reaction , computer science , medicine , mathematics , drug , algorithm , physics , quantum mechanics , psychiatry , pharmacology
The Drug Safety Research Unit's current methods of investigating adverse drug reactions using prescription event monitoring are discussed. The statistical properties of estimators of rates of occurrence of events in post‐marketing surveillance using prescription event monitoring are considered, and a simple model is proposed based on an exponential distribution of time to first occurrence of the event. It is shown that current methodology closely relates to the use of maximum likelihood estimation under this assumption and the distributions of the estimators are shown to be approximately normal, which allows simple confidence intervals and tests to be developed. Two recent applications are considered and corresponding simulations are presented to verify the approximate properties of the test statistics, based on ratios of rates over time and between drugs. Sources of bias in the rates and rate ratios are considered, including under‐reporting in later months. A rule‐of‐thumb, developed from many years experience, is shown to be generally conservative, except when these under‐reporting biases are large.