Maternal Life History- Versus Gestation-Focused Assessment of Prenatal Exposure to Substances of Abuse

Abstract

Purpose: Substance use by pregnant women is socially stigmatized and may be legally punishable. This societal condemnation raises concerns about underascertainment of prenatal substance exposure of offspring if mothers are asked specifically about their behavior during gestation, versus their life histories without reference to gestational dates. This study assessed agreement between life history-focused and pregnancy-focused assessments of prenatal exposure, and percentages of offspring classified as exposed to a range of substances by each measure, in a sample of school-aged children of methadone-maintained, opioid-dependent parents.

Methods: Prenatal exposure was assessed in 172 offspring of 109 mothers by: (a) questionnaires administered to mothers about substance use during pregnancy; and (b) best-estimate (BE) diagnoses of substance use disorders in mothers overlapping with pregnancy dates. BE diagnoses were based on interviews with the Schedule for Affective Disorders and Schizophrenia-Lifetime Version, conducted by trained mental health professionals with mothers about their life histories of psychiatric and substance use disorders, as well as mothers' medical records. Chance-corrected agreement between the measures was examined using κ statistics. Percentages of offspring classified as exposed by each method were compared using McNemar χ² tests.

Results: Except for cigarettes, agreement between the measures was poor. Except for alcohol, diagnosed episodes of substance use disorders in mothers with dates overlapping pregnancy classified more offspring as exposed than mothers' responses to the questionnaire focusing on behavior while pregnant, though the differences in proportions identified as exposed were not always large or statistically significant.

Implications: When retrospective ascertainment of prenatal exposure is necessary, asking mothers for their own life histories, without reference to pregnancy dates, may be the preferred approach.