Identifying parents in pharmacy data: a tool for the continuous monitoring of drug exposure to unborn children

Objective: This study aimed to develop and validate a method for retrospectively identifying parents in pharmacy data during pregnancy. Study Design and Setting: The principle of the method was to select all children 0-2 years in pharmacy records, and to consider men/women 15-50 years older with the same address as fathers/mothers. Results: Applying this method to the records of all 4 pharmacies in 1 town (33,000 inhabitants) resulted in identification of 807 fathers and 765 mothers, corresponding with 68.5% of all fathers, and 64.9% of all mothers from the town. Additionally, the method was applied to one selected pharmacy, resulting in 151 fathers and 170 mothers. Validation criterions, evaluated by pharmacy employees and GPs, disproved one of these fathers (0.7% of all identified fathers) and one mother (0.6%). Conclusion: We conclude that automatic retrospective identification of parents in pharmacy data is feasible in a valid way. The main limitation is that not all parents were found, possibly resulting in selection bias. (C) 2004 Elsevier Inc. All rights reserved.