Acute adverse event signalling scheme using the Saskatchewan Administrative health care utilization datafiles: results for two benzodiazepines.

Linked administrative health care utilization databases offer potential benefits for postmarketing surveillance. The value of the Saskatchewan datafiles in an acute adverse event signalling scheme has been evaluated using two benzodiazepines. The first 20,000 patients dispensed lorazepam and the first 8525 patients dispensed alprazolam were followed through the datafiles over the year after their initial prescription of the relevant drug, and all medical services occurring during treatment were recorded. The most frequent adverse drug reactions to benzodiazepines are drowsiness, depression, impaired intellectual function and memory, lethargy, impaired coordination, dizziness, nausea and/or vomiting, skin rash, and respiratory disturbance. Data from our study showed that sleep disorders, depressive disorders, dizziness and/or vertigo, respiratory symptoms, esophagus and stomach disorders, and inflammatory skin conditions occurred significantly more often in the first 30 days after the initial prescription than in the succeeding six months in both drug groups, indicating that they are important adverse events. There are several limitations to the methodology; however, the results of the analysis indicate that the use of administrative health care utilization datafiles in a systematic assessment to signal potential acute adverse drug reactions is a feasible proposition, but further studies are required to assess whether events are real adverse reactions.