Self-report and urine tests moderately agreed on cannabis use during pregnancy | Research

The takeaway

Self-report and urine drug screening for cannabis during pregnancy showed only moderate agreement (kappa=0.49), with discordant results explained by low-level use undetected by urine tests and pre-awareness-of-pregnancy use detected by urine but not reported.

Prenatal cannabis researchers, obstetricians, epidemiologists studying exposure measurement

Kappa=0.49 (moderate agreement)

What the researchers found

Concordance between self-report and urine drug screen was moderate (k=0.49). SR+/UDS+ (n=107), SR-/UDS- (n=142), SR+/UDS- (n=44, less frequent use), SR-/UDS+ (n=40, often used before knowing they were pregnant). Over 50% of discordant cases became concordant by trimester 2. Using SR+ and/or UDS+ as exposed changed classification minimally.

Why it matters

Accurately measuring prenatal cannabis exposure is essential for research on fetal outcomes. This study shows neither self-report nor urine screening alone captures the full picture, with important implications for how prenatal cannabis studies classify exposure.

The numbers in context

333 participants (88.6% Black, 45.4% below poverty). Kappa=0.49. SR+/UDS+: 107. SR-/UDS-: 142. SR+/UDS-: 44. SR-/UDS+: 40. 35% of SR-/UDS+ reported secondhand exposure or blunt use. Over 50% of discordant cases resolved by trimester 2.

How the study worked

CUDDEL Study analyzing 333 pregnant individuals (88.6% Black, mean age 26.6) with both self-report and first-trimester urine drug screens. Concordance assessed with kappa statistic and groups characterized by demographics, use patterns, and exposure history.

What this study cannot tell us

Predominantly Black, urban, lower-income sample may not generalize. CUDDEL Study recruited women with lifetime cannabis use history. UDS has limited detection window. Secondhand exposure may cause positive UDS without active use.

How to read the evidence

Well-designed concordance study with both measurement methods, but limited to one site with a specific demographic profile.

When this study was published

2024 analysis from the CUDDEL Study of pregnant individuals

The bigger picture

The finding that relying on self-report provides "adequate approximation" of cannabis use is reassuring for large epidemiological studies that cannot collect biological specimens. However, the 12% SR-/UDS+ rate means significant underestimation occurs in some populations.

Questions still open

Should prenatal cannabis research use both self-report and biological testing? How does secondhand cannabis exposure affect fetal outcomes?

Common questions

Can urine tests always detect cannabis use during pregnancy?

No. 44 women who reported using cannabis had negative urine tests, likely because they used infrequently enough that levels were below detection thresholds.

Why did some women test positive but deny use?

40 women denied use but tested positive. Many had used before knowing they were pregnant (62.5%), and 35% reported secondhand exposure or blunt use that may have contributed to positive tests.

Read the original research

Characteristics of women concordant and discordant for urine drug screens for cannabis exposure and self-reported cannabis use during pregnancy.

Neurotoxicology and teratology, 103, 107351

View on PubMed ↗Publisher’s paper ↗

Citation

Bogdan, Ryan; Leverett, Shelby D; Constantino-Petit, Anna M; Lashley-Simms, Nicole; Liss, David B; Johnson, Emma C; Lenze, Shannon N; Lean, Rachel E; Smyser, Tara A; Carter, Ebony B; Smyser, Christopher D; Rogers, Cynthia E; Agrawal, Arpana. (2024). Characteristics of women concordant and discordant for urine drug screens for cannabis exposure and self-reported cannabis use during pregnancy.. Neurotoxicology and teratology, 103, 107351. https://doi.org/10.1016/j.ntt.2024.107351