Abstract
This cross-sectional study examines data across 17 birthing hospitals before and after a policy change at Boston Medical Center in how reporting decisions are made in cases of prenatal substance exposure.
Clinicians are mandated reporters of suspected child abuse and neglect. In at least 26 US states, this includes cases of prenatal substance exposure (PSE) without exceptions for prescribed opioids or medications for opioid use disorder (MOUD).1 Yet mandatory reporting for PSE may stigmatize evidence-based care while contributing to unnecessary family surveillance and inefficient use of Child Protective Services (CPS) resources.1,2,3 In Massachusetts, only half of PSE reports filed in fiscal year 2021 were found to be consistent with abuse or neglect.4
In May 2021, Boston Medical Center (BMC) implemented a clinical practice guideline to inform reporting decisions for PSE, focusing on trauma-informed care and substance use treatment engagement with reporting to CPS only if protective concerns are identified. BMC’s guidance explicitly noted that CPS reports should not automatically be filed for prescribed opioid or MOUD use, positive toxicology tests, or positive substance use screens in the absence of protective concerns. We examined short-term impacts of this policy change.
Methods
The Boston University Medical Campus institutional review board approved this study with a waiver of informed consent. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
To evaluate BMC’s guideline (eMethods 1 in Supplement 1), we conducted a difference-in-differences analysis of opioid-exposed birthing parent–infant dyads across 17 birthing hospitals reporting quarterly to the Massachusetts Perinatal-Neonatal Quality Improvement Network (eMethods 2 in Supplement 1) from January 1, 2017, to May 31, 2023. We examined 3 newborn outcomes: proportion with CPS reporting prior to discharge, proportion discharged in a biological parent’s custody, and length of stay (LOS). Analyses used linear regression models with an implementation washout period from April to June 2021, an interaction term between binary indicators for “treatment” group (BMC vs non-BMC) and time period (before or after implementation), and adjustment for hospital and year fixed effects, prenatal exposure to nonprescribed opioids, and neonatal intensive care unit admission. Regression specifications, confirmation of difference-in-differences assumptions, and justification for our analytic approach are provided in eMethods 3 and 4 in Supplement 1. Statistical analyses were performed using R version 4.3.1 (R Foundation), with a significance threshold of P < .05.
Results
We identified 3658 opioid-exposed parent-infant dyads; 440 (12.0%) had birth hospitalizations at BMC (Table). The Figure shows outcome trends by quarter-year. Adjusted linear regressions confirmed preintervention trends without significant differences. In adjusted difference-in-differences analyses, guideline implementation was associated with a 44.9 percentage-point decrease (95% CI, −52.6 to −37.2) in the proportion of newborns reported to CPS without significant change in the proportion discharged with a biological parent (4.5 percentage points; 95% CI, −1.0 to 9.9), or LOS (0.0 day; 95% CI, −1.9 to 1.9).
Table. Characteristics of Birthing Parent–Infant Dyads at BMC vs All Other Sites, 2017-2023.
| Characteristic | No. (%) | ||
|---|---|---|---|
| BMC (n = 440) | All other sites (n = 3218) | Overall (N = 3658) | |
| Birthing parent racea | |||
| Another race | 17 (3.9) | 129 (4.0) | 146 (4.0) |
| Black or African American | 49 (11.1) | 116 (3.6) | 165 (4.5) |
| White | 327 (74.3) | 2839 (88.2) | 3166 (86.6) |
| Missing | 47 (10.7) | 134 (4.2) | 181 (4.9) |
| Birthing parent ethnicitya | |||
| Hispanic | 30 (6.8) | 244 (7.6) | 274 (7.5) |
| Not Hispanic | 384 (87.3) | 2794 (86.8) | 3178 (86.9) |
| Missing | 26 (5.9) | 180 (5.6) | 206 (5.6) |
| Infant sex assigned at birth | |||
| Female | 210 (47.7) | 1538 (47.8) | 1748 (47.8) |
| Male | 226 (51.4) | 1672 (52.0) | 1898 (51.9) |
| Missing | 4 (0.9) | 8 (0.2) | 12 (0.3) |
| Gestational age, wk | |||
| Mean (SD) | 38.0 (2.22) | 38.4 (2.06) | 38.4 (2.09) |
| Missing | 6 (1.4) | 33 (1.0) | 39 (1.1) |
| Receipt of any NICU-level care | |||
| Yes | 281 (63.9) | 2667 (82.9) | 2948 (80.6) |
| No | 159 (36.1) | 551 (17.1) | 710 (19.4) |
| Type of in utero opioid exposure | |||
| Only prescribed opioids, including MOUD | 249 (56.6) | 2102 (65.3) | 2351 (64.3) |
| Only nonprescribed opioids or unknown sources | 47 (10.7) | 450 (14.0) | 497 (13.6) |
| Both prescribed and nonprescribed opioids | 136 (30.9) | 607 (18.9) | 743 (20.3) |
| Unknown exposure | 8 (1.8) | 59 (1.8) | 67 (1.8) |
| Other in utero substance exposuresb | |||
| Alcohol | 23 (5.2) | 100 (3.1) | 123 (3.4) |
| Cannabis | 44 (10.0) | 948 (29.5) | 992 (27.1) |
| Cocaine | 129 (29.3) | 594 (18.5) | 723 (19.8) |
| Nicotine/tobacco | 244 (55.5) | 1563 (48.6) | 1807 (49.4) |
| Amphetamine/methamphetamine | 38 (8.6) | 321 (10.0) | 359 (9.8) |
| Benzodiazepine | 89 (20.2) | 480 (14.9) | 569 (15.6) |
| Gabapentin | 36 (8.2) | 256 (8.0) | 292 (8.0) |
| Other | 41 (9.3) | 349 (10.8) | 390 (10.7) |
| Newborn receipt of NOWS pharmacotherapy | |||
| No | 244 (55.5) | 1982 (61.7) | 2226 (60.9) |
| Yes | 196 (44.5) | 1236 (38.4) | 1432 (39.1) |
Abbreviations: BMC, Boston Medical Center; MOUD, medication for opioid use disorder; NICU, neonatal intensive care unit; NOWS, neonatal opioid withdrawal syndrome.
Race and ethnicity are reported as documented in the electronic health record, with options as designated by each study site. The category Another Race includes American Indian and Alaska Native, Asian, Native Hawaiian and Pacific Islander, and other races not otherwise listed.
Totals may sum to >100% because multiple exposures may be present.
Figure. Changes in Newborn Hospitalization Outcomes at Boston Medical Center (BMC) vs All Other Sites.
Each panel depicts observed rates of outcomes aggregated for BMC and all other sites by quarter. The vertical dotted line indicates implementation of the BMC guideline change. The shaded area represents the washout period during which observed outcomes were excluded from analyses to allow time for guideline implementation. CPS indicates Child Protective Services; Q2, second quarter.
Discussion
Adoption of a trauma-informed clinical practice guideline requiring identification of specific protective concerns in the setting of PSE led to a large reduction in CPS reports. There were no significant changes in custody at discharge or LOS, suggesting the guideline change did not alter short-term disposition or discharge readiness.
Birthing parents prescribed MOUD perceive mandatory reporting to be unjust, stigmatizing, and harmful to family health.1 National quantitative studies corroborate these narratives: punitive reporting policies decrease access to substance use treatment and fail to improve neonatal outcomes.2,3 Our early findings should encourage health systems and policymakers to consider a possible counterfactual: if CPS reporting is redefined based on expert-identified protective concerns rather than categorical mandates, the benefits could include conservation of limited state resources, reduction in family surveillance, and mitigation of stigma.5
Limitations include generalizability, as this guideline change was implemented at an institution with dyad-focused prenatal and postpartum addiction care, including multidisciplinary case review and care coordination.6 Strengths include use of rigorous quasi-experimental methods and statewide comparative data to strengthen causal inferences.
Our study highlights that policies can shift from the harmful practice of categorically reporting all cases of PSE toward a trauma-informed process without short-term unintended consequences. These findings should motivate longitudinal analysis of dyad outcomes associated with judicious, guideline-directed filing practices for PSE.
eMethods 1. BMC Clinical Practice Guideline: “When to Report Protective Concerns in the Setting of Prenatal Substance Use”
eMethods 2. Details Regarding Data Source and Reporting Hospitals
eMethods 3. Main Difference-in-Differences Regression Specification
eMethods 4. Pre-Policy Trend Comparisons between BMC and Non-BMC Dyads (Parallel Trends Assumption)
Data sharing statement
References
- 1.Work EC, Muftu S, MacMillan KDL, et al. Prescribed and penalized: the detrimental impact of mandated reporting for prenatal utilization of medication for opioid use disorder. Matern Child Health J. 2023;27(suppl 1):104-112. doi: 10.1007/s10995-023-03672-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Austin AE, Naumann RB, Simmons E. Association of state child abuse policies and mandated reporting policies with prenatal and postpartum care among women who engaged in substance use during pregnancy. JAMA Pediatr. 2022;176(11):1123-1130. doi: 10.1001/jamapediatrics.2022.3396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Atkins DN, Durrance CP. State policies that treat prenatal substance use as child abuse or neglect fail to achieve their intended goals. Health Aff (Millwood). 2020;39(5):756-763. doi: 10.1377/hlthaff.2019.00785 [DOI] [PubMed] [Google Scholar]
- 4.Massachusetts Department of Children and Families . Annual report FY2021. Accessed September 24, 2023. https://www.mass.gov/doc/dcf-annual-reportfy2021/download
- 5.Raz M, Gupta-Kagan J, Asnes AG. Using child abuse specialists to reduce unnecessary Child Protective Services reports and investigations. JAMA Pediatr. 2023;177(12):1249-1250. doi: 10.1001/jamapediatrics.2023.3676 [DOI] [PubMed] [Google Scholar]
- 6.Boston Medical Center . Project RESPECT. Accessed September 29, 2023. https://www.bmc.org/obstetrics/pregnancy/addiction
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods 1. BMC Clinical Practice Guideline: “When to Report Protective Concerns in the Setting of Prenatal Substance Use”
eMethods 2. Details Regarding Data Source and Reporting Hospitals
eMethods 3. Main Difference-in-Differences Regression Specification
eMethods 4. Pre-Policy Trend Comparisons between BMC and Non-BMC Dyads (Parallel Trends Assumption)
Data sharing statement
