One-third of pregnant cannabis users don't disclose, especially minorities

This study examines a critical gap in cannabis research: the reliability of self-reported cannabis use during pregnancy. Using data from a Michigan population-based pregnancy cohort of 606 participants, researchers compared what women reported about their cannabis use to actual urine toxicology test results. The findings reveal a significant discrepancy—self-reports captured only 14.2% prevalence while urine tests detected 19.1%, indicating that nearly one-third of pregnant cannabis users were not disclosing their use. The accuracy of self-reports varied substantially, with sensitivity (ability to correctly identify users) at 57.8% and notably poorer performance among certain populations.

The most concerning finding is that Non-Hispanic Black participants had sensitivity of only 45.5% compared to 73.0% for Non-Hispanic White participants, suggesting systematic differences in disclosure patterns across racial groups. This disparity likely reflects legitimate concerns about legal and social consequences, as pregnant women may fear child protective services involvement or legal action. The study found that participants were highly accurate when they did disclose (77.9% positive predictive value), but many chose not to disclose at all. No other significant demographic differences were found between those who disclosed and those who didn't.

This research has important implications for cannabis science and public health: self-reported data systematically underestimates prenatal cannabis exposure, particularly in disadvantaged and minority populations, which can skew research findings and misrepresent actual usage patterns. The researchers conclude that biospecimen testing (urine toxicology) should become standard in cannabis research to reduce misclassification errors and improve the validity of findings, especially when studying vulnerable populations where social stigma and legal concerns create barriers to honest reporting.