Signal generation in the New Zealand Intensive Medicines Monitoring Programme: A combined clinical and statistical approach

• Published in: Drug safety 2002, Vol.25 (6), p.433-439
• Main Author: COULTER, David M
• Format: Article
• Language: English
• Subjects: Adverse Drug Reaction Reporting Systems - statistics & numerical data; Analysis. Health state; Automatic Data Processing - methods; Biological and medical sciences; Clinical trial. Drug monitoring; Cohort Studies; Data Interpretation, Statistical; Drug Interactions; Drug Monitoring - methods; Drug Prescriptions - statistics & numerical data; Epidemiology; General aspects; General pharmacology; Humans; Medical sciences; New Zealand; Pharmacology. Drug treatments; Product Surveillance, Postmarketing - statistics & numerical data; Public health. Hygiene; Public health. Hygiene-occupational medicine; Risk Assessment
• ISSN: 0114-5916
• DOI: 10.2165/00002018-200225060-00007

The New Zealand Intensive Medicines Monitoring Programme (IMMP) undertakes prospective observational cohort studies on selected new drugs in the early postmarketing period using prescription-event monitoring (PEM) methodology with the purpose of identifying signals of previously unrecognised ADRs and establishing risk profiles for each drug. Events are reviewed by a physician and a relationship is established between each event and the drug. The events are then sorted into reactions and incidents. The latter are used to assist signal detection and control for bias. Rates for reports, reactions and incidents are used to assess the adequacy of reporting, signal detection and identification of confounders. Most signals are identified by clinical evaluation of the reports at a stage when statistical analyses are unlikely to have the power to detect them with confidence. The incident group is used for signal detection and controlling for bias. A low reporting rate indicates that certain types of event are unlikely to be reported. A systematic review of the original case reports at the site of collection provides the best opportunity for early signal detection. More resources need to be invested in the training and support of clinical evaluators. Categorising events into reactions and incidents gives added value to the data. Rates of reporting should be quoted with the results of cohort studies to facilitate assessment of their power to detect new signals.