Determinants of recall and recall bias in studying drug and chemical exposure in pregnancy

Abstract

Case-control studies on effects of drugs in pregnancy rely heavily on maternal recall. At the Motherisk Program in Toronto we counsel women during early pregnancy on the risk of drug and chemical exposure; subsequently, we follow up the outcome of pregnancy after birth. This cohort has given us an opportunity to assess the magnitude of recall of early pregnancy exposure and determinants likely to affect it in 145 consecutive cases. The mean recall of exposure identity was 62%, while accurate recall of timing of exposure was 37% and of dosage 24%. Exposures that prompted the clinic visit, chronic therapeutic exposures, environmental agents, and known teratogens were recalled significantly better than were other exposures. Accurate report of smoking was significantly higher than of alcohol use (79.4% vs. 59%, respectively, P = .0002). The number of agents consumed by the pregnant woman negatively correlated with her recall; mean recall of 1 agent was 85% vs. only 40% recall of 4 agents. Women greater than or equal to 30 years of age recalled significantly worse (mean +/- SEM, 52 +/- 4%) than women younger than 30 (70 +/- 4%), P = .002) despite a similar mean number of exposures. No difference in mean recall was found between women having normal (n = 112) or adverse pregnancy outcome (n = 33). There was a recall bias in reporting alcohol consumption; postnatally, women with adverse outcome tended to report significantly less than the amount initially reported by them.