Among 5540 total pregnancies in the MAT-LINK OUD cohort, there were 3911 pregnancies (70.6%) with any documented MOUD. Among pregnancies with any MOUD, the majority of people were White (n = 3351, 85.7%), not Hispanic or Latino (n = 2722, 69.6%), receiving public health insurance (ie, Medicaid) (n = 3285, 84.0%), or living in urban core zip codes (n = 3206, 82.0%) (Table 1). Almost half (n = 1780, 45.5%) of pregnancies were among people who had 1 to 2 prior pregnancies. Considering that multiple types of MOUD could have been documented during each pregnancy, among the 3911 pregnancies included in this analysis, there were 1642 (42.0%) with methadone, 1662 (42.5%) with BUP, 1101 (28.2%) with BUP-NAL, and 22 (0.6%) with naltrexone documented at least once in the EHR. When pregnancies were assessed by MOUD type in nonmutually exclusive categories, patterns in age at pregnancy outcome and race were similar across each MOUD type; however, patterns in ethnicity differed by MOUD type. Among pregnancies with methadone and BUP, 34.6% (n = 568) and 35.6% (n = 592), respectively, were among Hispanic or Latino people, whereas among pregnancies with BUP-NAL, only 6.6% (n = 73) were among Hispanic or Latino people. Among pregnancies with methadone, 90.5% (n = 1486) were among people covered by public insurance, which was greater than the corresponding percentages among pregnancies with BUP (n = 1290, 77.6%), BUP-NAL (n = 899, 81.7%), or naltrexone (n = 15, 68.2%). Compared with other MOUD types, pregnancies with BUP-NAL had the highest percentage of rural zip code residence (n = 128, 11.6%), followed by BUP (n = 149, 9.0%) and methadone (n = 83, 5.1%). Pregnancies with naltrexone had the highest percentage of nulliparous pregnancies (n = 8, 36.4%), whereas BUP-NAL had the lowest percentage of nulliparous pregnancies (n = 204, 18.5%). From January 2014 to August 2021, the proportion of pregnancies with documented BUP-NAL and naltrexone significantly increased over time, whereas the proportion with methadone significantly decreased over time (Figure, Supplemental Digital Content 1, http://links.lww.com/JAM/A580). Long-acting BUP and naltrexone products were exclusively seen in pregnancies with an outcome occurring during 2017 and beyond.

Table 2 shows that among most pregnancies where trimester information was available for both first MOUD and prenatal care initiation, prenatal care was initiated during the same trimester as the first documented MOUD during pregnancy. However, more than 30% (n = 450/1454) of pregnancies with prenatal care information documented in the first trimester did not have MOUD documented until the second or third trimester. Additionally, among the pregnancies with MOUD in the first trimester, 12.3% (n = 141/1145) had no prenatal care until the second or third trimester. The median gestational age at first MOUD during pregnancy was 20 weeks, whereas the median gestational age at prenatal care initiation was 15 weeks.

In later trimesters, the percentage of pregnancies with documented MOUD notably increased, as shown in Table 3; 37.0% (n = 1447/3911) of pregnancies had MOUD documented in the EHR during the first trimester, 61.3% (n = 2324/3791 continuing pregnancies) during the second trimester, and 90.8% (n = 3347/3688 continuing pregnancies) during the third trimester. Within 2 weeks prior to the pregnancy outcome, which could have occurred during any trimester, 80.1% (n = 3134/3911) of pregnancies had any MOUD. Across all time frames of pregnancy, BUP was the most commonly documented MOUD type. Overall, BUP-NAL was also commonly documented in more than a quarter of pregnancies.

Among pregnancies with any MOUD, most had a single MOUD type documented (n = 3412, 87.2%), whereas 12.3% (n = 482) pregnancies had 2 and 0.4% (n = 17) pregnancies had 3 MOUD types documented, respectively (Fig. 1). Among the pregnancies with 2 or 3 MOUD types, the most common transitions were from BUP-NAL to BUP and from BUP to methadone. As further shown in Figure 2, whereas the first documented MOUD type per trimester stayed consistent for most pregnancies, there were also pregnancies either with multiple types of MOUD or with discontinuity in MOUD across trimesters (represented as a transition from any MOUD type to the “No MOUD” category).

Among 5540 pregnancies with an OUD diagnosis and a pregnancy outcome between January 1, 2014, and August 31, 2021, 7 in 10 had any MOUD in the EHR, and the majority had only one single MOUD type documented during pregnancy. However, this report highlights opportunities to enhance timely access to MOUD during pregnancy. Approximately 30% of pregnancies impacted by OUD did not have any documented MOUD. Among those with any MOUD documented during pregnancy, demographic differences were observed by MOUD type, such as health insurance status. Notably, 12.3% (n = 141/1145) of pregnancies with MOUD documented in the first trimester did not have prenatal care until the second or third trimester, and 30.9% (n = 450/1454) of pregnancies with prenatal care in the first trimester did not have MOUD documented until the second or third trimester. Furthermore, less MOUD was documented during the first trimester compared with the second and third trimesters, and discontinuity in reported MOUD across trimesters was observed.

In this analysis of MAT-LINK’s OUD cohort, a higher proportion of pregnancies with any MOUD was observed compared with existing estimates from 1996 to 2017 (50%–60%) based on the Medicaid administrative data or the Treatment Episode Data Set.^9,21–23 This difference might be attributed to the origin of the MAT-LINK data, which are collected from 7 clinical sites renowned for their expertise in managing OUD during pregnancy.¹⁷ These clinics provide care using a multidisciplinary, comprehensive, and nonstigmatizing approach, likely leading to higher linkage to and continuation of MOUD during pregnancy. In this report, methadone and BUP were documented most frequently overall, and transitioning between MOUD types most often involved shifts from BUP-NAL to BUP. However, BUP-NAL was still commonly observed in more than a quarter of pregnancies and showed an increasing trend in recent years. These patterns were expected because methadone and BUP are the MOUD types currently recommended in clinical guidelines for OUD during pregnancy.² However, following emerging evidence supporting BUP-NAL for OUD during pregnancy,^5,24 BUP-NAL was also included in the updated 2020 ASAM national practice guideline as a safe and effective MOUD option during pregnancy. Results from this report demonstrate an increase in use of BUP-NAL over time, correlating with the release of clinical guidance.

Naltrexone was uncommon but increasingly documented in more recent years. In addition to not being endorsed in clinical guidance for MOUD during pregnancy, its limited use in this cohort might also be explained by scarcity of sites nationwide that routinely provide naltrexone as a treatment option for pregnant people.²⁵ Emerging evidence on the use of naltrexone for OUD during pregnancy has shown comparable outcomes to buprenorphine and methadone for gestational age at delivery, mode of delivery, and preterm birth.²⁶ Barriers to offering naltrexone during pregnancy might include concerns around requiring a 7-to-10-day opioid-free period before naltrexone initiation, which might present as a period of vulnerability for people to return to nonprescribed substance use.²⁷ Due to this required period of opioid withdrawal, naltrexone has been proposed as a potential option for some pregnant populations, such as those who were already receiving and successfully maintaining treatment before pregnancy.^6,26

Certain demographic characteristics appeared to differ by MOUD type documented during pregnancy, including pregnant person ethnicity, health insurance coverage, urbanicity, parity, and year of pregnancy outcome. A previous report of MAT-LINK data comparing any MOUD versus no MOUD during pregnancy observed a significant difference according to pregnant person race.¹⁷ However, in the current report, no differences in the distribution of race were observed by MOUD type. These findings contradict previously published literature using other data sources, which have shown racial disparities affecting MOUD access during pregnancy, specifically timely receipt of MOUD,^14,23 receipt of BUP/BUP-NAL versus methadone,^14,15 and consistent use of MOUD during pregnancy and postpartum periods.^14,15 These examples demonstrate the persistence of systemic racism and implicit bias toward pregnant populations from racial minority groups,^14,15 and further analyses are needed to understand disparities affecting the MAT-LINK OUD cohort. There is an opportunity and need at the federal, state, and local levels to address racial and other disparities in treatment by identifying treatment gaps by patient characteristics, such as race and residence, and implementing targeted interventions where services are most needed.²⁸ This report did, however, find different patterns in ethnicity by MOUD type: 7% of pregnancies with BUP-NAL were among Hispanic or Latino people, whereas more than 30% of pregnancies with BUP or methadone were among Hispanic or Latino people. This is one of the few analyses to observe differences in buprenorphine by ethnicity across multiple clinical sites.^14,15 Prior reports, such as that by Xu et al, have suggested that analyses on MOUD in Hispanic or Latino pregnant people have lacked sufficient statistical power.¹³ Further exploration is needed to understand the differences in ethnicity by MOUD type, especially because these differences may reflect variations in MOUD practices due to implicit biases or structural barriers to treatment, such as the accessibility of clinical services by Hispanic or Latino pregnant people.

Different patterns in health insurance were also observed by MOUD type. More than 90% of pregnancies with methadone were to publicly insured pregnant people, surpassing the percentage for any other MOUD type examined. Health insurance status in this report might correlate with an individual’s financial resources or ability to afford out-of-pocket costs for MOUD, particularly considering that buprenorphine typically may cost more than methadone.^29,30 “Secret shopper” studies, where trained callers pose as people with assigned demographic profiles needing services, and surveys of MOUD prescribers revealed that many clinicians did not accept private or public insurance and instead only accepted cash payments from both pregnant and nonpregnant women seeking treatment.^29–32 Researchers posing over phone calls as a pregnant person with Medicaid insurance were less likely to be offered an appointment with a buprenorphine prescriber than at an OTP and encountered more barriers compared with nonpregnant or privately insured women.³² Prioritizing the needs of pregnant people without socioeconomic biases when making decisions about MOUD use might improve affordability and accessibility of MOUD prior to and during pregnancy.²

The percentage of pregnancies with any MOUD at these clinical sites more than doubled over the course of pregnancy, from the first to the third trimester. During most pregnancies, prenatal care was initiated during the same trimester as their first documented MOUD during pregnancy, suggesting that prenatal care and initiation of MOUD may be facilitators for one another. However, 3 of 10 pregnancies with first-trimester prenatal care information had no MOUD documented until the second or third trimester, and 12% of pregnancies with MOUD in the first trimester had no prenatal care documented until the second or third trimester. This potentially indicates missed opportunities for early linkage to prenatal care and MOUD treatment³³ and prompts the need to support and address barriers to linking people with OUD to prenatal care and MOUD during early pregnancy. Stigma against substance use and MOUD during pregnancy, along with punitive laws and policies, persist as barriers to care in the United States.^34–36 Prior researchers have shown that punitive state-wide prenatal substance use policies were associated with reduced admission of women of reproductive age to substance use disorder (SUD) pharmacological and psychosocial treatment, whereas policies funding SUD treatment programs for pregnant people were associated with reduced opioid overdoses and increased access to MOUD.^34,35 As of 2023, almost all MAT-LINK clinical sites were located in states with supportive policies funding programs for pregnant people with SUD, but approximately half of these states simultaneously had punitive or mandatory reporting policies.^37,38 A nonstigmatizing approach to the management of pregnant people with OUD might facilitate MOUD access during pregnancy and potentially have downstream benefits such as reducing the rate of entry into foster care.³⁹

This analysis is subject to some limitations. First, these results are not generalizable to all US individuals who access OUD treatment during pregnancy. Second, because multiple MOUD types were documented during some pregnancies, MOUD type categories were nonmutually exclusive, and χ² tests could not be performed to determine statistically significant differences in demographic characteristics or MOUD timing by MOUD type. However, because MAT-LINK captured data on a census of pregnancies with OUD in these 7 clinical sites, sampling error is not expected among this analytic sample, so statistical testing is not required to interpret descriptive comparisons. Third, observed differences in ethnicity could have been influenced by clustering of people documented as Hispanic or Latino at certain clinical sites with specific MOUD treatment patterns. Fourth, EHR-documented MOUD includes inpatient and outpatient prescriptions but may not always reflect dispensation or consumption. Fifth, certain variables are prone to misclassification in the EHR or were not included in the analysis because they are not collected systematically within or across EHR systems at clinical sites. Information on MOUD timing was subject to some missingness and misclassification because some clinical sites reported difficulties obtaining MOUD data from external sources, especially OTPs. Consequently, methadone data might be less accurate than other data. Lastly, data were not available to differentiate whether the MOUD documented in the first trimester was initiated prior to versus after conception.

The MAT-LINK surveillance system offers ample opportunity for further analyses on MOUD during pregnancy as a unique collection of extracted and abstracted EHR data spanning nearly 8 years of pregnancy outcomes and involving 7 US clinical sites.^16,17 Future analyses on MOUD during pregnancy might consider dosing and frequency, continuity during the postpartum period, and predictors and effects of transitioning between MOUD types. MAT-LINK will continue collecting MOUD data through December 31, 2024, with 3 additional clinical sites to monitor evolving clinical guidance and legislative environments related to the opioid crisis.

In the MAT-LINK OUD cohort, the overall frequency of MOUD appeared to improve over the course of pregnancy, with rates increasing from the first to the second and third trimester. From 2014 to 2021, BUP-NAL and naltrexone were increasingly documented MOUD types during pregnancy, indicating that clinical practice appears to be evolving over time in response to growing evidence and updated clinical guidance. Contextual factors, such as insurance status and year of pregnancy outcome, might influence MOUD types documented during pregnancy. Although the findings of this report suggest that prenatal care and MOUD might be facilitators for one another, there are still opportunities to improve early linkage and continuous access to both prenatal care and MOUD during pregnancy. Continuity of care for pregnant people with OUD might improve when decisions on OUD and prenatal care are prioritized based on each person’s unique needs.