Implementation of universal screening for substance use in pregnancy in a public healthcare system

Abstract

Objective

Screening questionnaires are one option for identification of at-risk substance use and substance use disorder (SUD) during pregnancy. We report the experience of a single institution following universal implementation of a brief screening tool for self-reported substance use at the first prenatal encounter.

Study Design

This is a prospective implementation study evaluating screening for substance use in pregnancy in a large safety net healthcare system. Universal screening with the National Institute of Drug Abuse (NIDA) Quick Screen V1.0 was integrated into the electronic medical record (EMR) and administered at the first point of contact with the healthcare system. SUD was identified initially with diagnosis within the EMR by a healthcare provider and was confirmed with toxicology (maternal or neonatal) results corroborating a pattern of substance use and maternal and neonatal ICD-10 codes for SUD. Patients identified with SUD were then classified as moderate or severe SUD based on criteria established by the Diagnostic and Statistical Manual of Mental Disorders, 5th edition. We measured rates of NIDA implementation across different healthcare settings, evaluated NIDA concordance with ascertainment of SUD, and compared adverse pregnancy outcomes associated with moderate and severe SUD.

Results

From July 28, 2021, through June 25, 2022, 14,634 unique pregnant individuals accessed care at ambulatory and acute care sites. Universal implementation of the NIDA Quick Screen identified at-risk substance use in 2146 (14.7%) of those who accessed our system, or 17.1% of 12,550 screened across the system, with greater screen completion in ambulatory over acute care settings. SUD was identified in 256 (1.7%) of 14,634 individuals and moderate or severe SUD was identified in 184 (1.3%). Among those with moderate or severe SUD, 90 (48.9%) were NIDA positive, 22 (12.0%) NIDA negative, and 72 (39.1%) unscreened. Of 94 individuals with NIDA discordance or who were unscreened 76 (81%) accessed initial care through an acute care setting. Of 96 individuals with opioid use disorder, 68 (70.8%) were treated with medication-assisted therapy, and 56 (58.3%) were screened with the NIDA Quick Screen. Among delivered individuals with available outcomes, those with moderate or severe SUD were less likely to seek prenatal care (71 (76%) vs 9852 (98%), <0.001)) and more likely to deliver before 37 weeks, (18 (20%) vs 909 (9%), RR (95% CI) 2.13 (1.40, 3.24)) compared to individuals without SUD. Neonates exposed to moderate or severe SUD were more likely to have birth weight <10th centile for gestational age (20 (22%) vs 1147 (12%), RR (95% CI) 1.92 (1.29, 2.85)) and require admission to the neonatal intensive care unit (NICU) (19 (21%) vs 964 (10%), RR (95%) 1.95 (1.30, 2.93)).

Conclusion

Universal screening was implemented across a large public healthcare system at a high rate, with higher rates of implementation in ambulatory settings. NIDA successfully identified at-risk substance use in 17% of the SUD cohort but failed to identify more than 50% of patients with moderate or severe SUD. Patients with moderate and severe SUD accessed care primarily through the emergency department and experienced higher rates of adverse obstetric and neonatal outcomes. Future efforts to identify, engage, and retain this highest-risk group are needed.