Validity of retrospectively reported behaviors during the periconception window

Area of Science:

• Reproductive Health
• Perinatal Epidemiology
• Behavioral Science

Background:

• Accurate reporting of maternal behaviors during the periconception period is crucial for understanding pregnancy outcomes.
• Retrospective recall of these behaviors may be subject to inaccuracies, potentially biasing research findings.

Purpose of the Study:

• To evaluate the validity of retrospectively reported maternal behaviors (smoking, alcohol, vitamin, caffeine use, and sport fish consumption) during pregnancy attempts.

Main Methods:

• A prospective pregnancy cohort study compared daily diary entries (gold standard) with self-administered questionnaire data collected a decade later.
• Agreement was assessed using percent agreement and Kappa coefficients for five key behaviors.

Main Results:

• Validity for the five behaviors ranged from poor to moderate (54-74% agreement).
• Smoking (Kappa=0.43) and fish consumption (Kappa=0.32) showed higher validity than caffeine (Kappa=0.21) and alcohol (Kappa=0.20).
• Misclassification of behaviors significantly impacted associations with pregnancy outcomes.

Conclusions:

• Retrospective reporting of periconception behaviors demonstrates poor to moderate validity.
• More regular behaviors like smoking and caffeine consumption were reported with higher accuracy.
• Prospective data capture methods are essential to mitigate misreporting and ensure reliable research inferences.