Hospital pharmacy dispensing records for pharmacoepidemiology research into late gestation exposure to antidepressants

Aim: To determine the validity of electronic hospital pharmacy dispensing records as an indicator of late gestation exposure to antidepressants using medical records as the gold standard. Method: Medical records (gold standard) were reviewed for consecutive sample of 400 women who delivered at a large maternity hospital during 2005. Data on antidepressants used in pregnancy from the medical records were compared with the electronic pharmacy dispensing records to ascertain sensitivity and specificity. Results: According to the medical records 12 (3%) women were identified as taking antidepressants during late gestation. The electronic pharmacy dispensing records correctly identified 9 women as exposed to antidepressants (sensitivity score 75%; 95%CI 43-95). No women were classified as exposed according to pharmacy dispensing records and classified as not exposed according to the medical records (specificity score 100%; 95%CI 99-100). 7 (1.8%) women were identified in the medical records as taking antidepressants but had either stopped prior to pregnancy or on finding out that they were pregnant. These patients were not identified in the pharmacy dispensing records. Conclusion: Electronic hospital pharmacy dispensing records may provide an efficient alternative to paper records for identifying late gestation exposure to antidepressants but will underestimate the exposure by 25%.