Validity of Self-Reported Drug Use Information Among Pregnant Women

Mahek Garg, Laura Garrison, Lawrence Leeman, Ajna Hamidovic, Matthew Borrego, William F. Rayburn, Ludmila Bakhireva

This study assesses validity of self-report for the use of major classes of illicit drugs and opioid-maintenance therapy among pregnant women at a substance abuse treatment program. Analyses used data collected from 83 pregnant women in a prospective cohort study at the University of New Mexico. Study participants with a history of substance abuse were screened and, if eligible, enrolled during an early prenatal care visit. A follow-up interview was conducted shortly after delivery. Self-reported information about drug use later in pregnancy was compared with urine drug screen (UDS) results collected during the third trimester. Simple kappa (k) and prevalence-and-bias-adjusted kappa (PABAK) coefficients were calculated as the measures of agreement. Sensitivity and specificity of self-report for each drug class were estimated using UDS as the 'gold standard'. The sample included a large proportion of ethnic minority (80 % Hispanic/Latina and 7 % American Indian) and socially disadvantaged (50 % less than high school education and 94 % Medicaid-insured) pregnant women. On average, patients had 4.8 ± 3.0 urine drug screens during the third trimester. Sensitivity of self-report was low (<60 %) for all classes of illicit drugs; however, marijuana and opioids demonstrated slightly higher sensitivity (57.9 and 58.3 %, respectively) than other classes (<47 %). This study found substantial underreporting for all classes of illicit drugs among pregnant women in a substance abuse treatment program. Rates of underreporting are expected to be higher among the general population of pregnant women.