Validity of perinatal pharmacoepidemiologic studies using data from the RAMQ administrative database.

Background: The RAMQ prescription claims database (RAMQ-Rx) is increasingly being used in perinatal pharmacoepidemiologic studies; but, there are reasons to believe that results generated with the RAMQ-Rx might not be generalizable to all patient populations.

Objectives: Compare characteristics between pregnant women insured by the RAMQ-Rx and those insured by a private drug insurance plan.

Methods: A prospective study performed within the population of pregnant women receiving prenatal care at different obstetrics and gynecology clinics affiliated to the University of Montreal, Canada, was conducted from October 2004 to March 2006. Women were eligible if they were > or =18 years of age, < or =16 weeks of gestation at the time of their first prenatal visit, and able to read and understand French or English. Eligible women were asked to fill out a self-administered questionnaire.

Results: Three hundred and sixty-three women met inclusion criteria, of which 99 (27%) had RAMQ-Rx coverage, and 264 (73%) had a private drug insurance coverage. Compared to those who were covered by private drug insurance plans, those insured by the RAMQ-Rx were younger (30.7yrs vs. 32.1yrs; P=0.03), more likely to be immigrant (60% vs. 24%; P<0.01), and have a household income below poverty level (39% vs. 2%; P<0.01). They were also less likely to be Caucasian (69% vs. 86%; P<0.01), employed (51% vs. 87%; P<.01), and have a post-secondary education (76% vs. 95%; P<0.01). No differences were observed on smoking status and alcohol use during pregnancy.

Conclusions: There are substantial differences between pregnant women insured by the RAMQ-Rx and those insured by private drug insurance plans. However, these differences will most likely limit generalizability, but not internal validity, of studies using data from the RAMQ-Rx database.