Neonatal Abstinence Syndrome: Leveraging Health Information Technology to Develop a Data-Driven National Policy Approach

Neonatal abstinence syndrome, also referred to as neonatal opioid withdrawal syndrome (NAS/NOWS), is a consequence of the opioid epidemic that affects many US communities. NAS/NOWS is a form of drug withdrawal that can occur in infants after in utero exposure to opioids and other substances. ¹ From 2004 to 2014, the rate of NAS/NOWS increased more than 5-fold, from 2.8 to 14.4 per 1000 Medicaid-covered NAS/NOWS-related births. ² Ongoing efforts to address the acute effects of the opioid epidemic are foremost, but they are only part of the solution for the mother–infant dyad. Although notable gaps in knowledge exist, no fundamental, scalable mechanisms are in place to understand the treatment and recovery needs of mothers, the potential developmental needs of infants, and how the complexity of interactions among biological, social, and environmental factors may influence outcomes.

Strategies aimed at integrally supporting opioid-exposed mothers and infants are vital for long-term clinical care, including an emphasis on prepregnancy, prenatal, and postnatal pain management; opioid use disorder treatment and recovery services; and neurocognitive, behavioral, and academic development through childhood and adolescence. Taking into consideration the potential intersections and interactions between the needs of opioid-exposed mothers and their infants may inform longitudinal best practices for both. Above all, a data-driven approach that focuses on the dyad, rather than separately on the mother and infant, is essential. The purpose of this Executive Perspective is to highlight 4 key findings of a collaborative effort that the US Department of Health and Human Services (HHS) is leading with national experts, propose ideas for further applications of this initiative, and describe implications for national policy for NAS/NOWS.

The primary scope of this HHS initiative on NAS/NOWS is to develop and implement strategies aimed at improving long-term follow-up care of mothers and infants exposed to opioids and other substances by taking a health information technology (IT)–centered approach. This effort has entailed a series of listening sessions with state, federal, and private-sector partners specializing in maternal and infant care who shared their insights on gaps in care. At the community and state levels, experts have expressed concern about variability in approaches to inpatient and outpatient assessment and management of opioid exposure in pregnant mothers and their infants. This variability makes it unclear what the longitudinal outcomes of these approaches may be. At the national level, much work remains to be done to understand the long-term needs—both clinical and social—of vulnerable mothers and infants.

Because of the need for an approach that encompasses longitudinal continuums of care, key participants in this discussion have included experts in public health, obstetrics/gynecology, primary care and developmental pediatrics, neonatology, family medicine, early intervention, psychiatry and mental/behavioral health, internal medicine, social work, bioinformatics, and bioethics. To advance discussions, in fall 2018, HHS held a 1-day national convening on NAS/NOWS in Philadelphia with an overarching focus on health IT–centered strategies while engaging broader perspectives on mother–infant dyadic care. The convening had a 2-pronged policy–academic focus. First Lady of the United States Melania Trump, HHS Secretary Alex M. Azar II, Assistant Secretary for Health and Senior Adviser for Opioid Policy ADM Brett P. Giroir, and Chief Medical Officer for the Office of the National Coordinator (ONC) for Health IT Thomas Mason led the policy charge. A scholarly symposium of national experts led the academic charge, with a focus on illuminating current gaps in long-term care and bridging gaps with health IT. This convening was the first meeting about NAS/NOWS to uniquely integrate clinical and IT specialties; it marked an important advancement in raising, on a national level, (1) visibility about the effect of the opioid epidemic on vulnerable mother–infant patient populations, (2) dialogue on the dyadic challenge of caring for substance-exposed mothers and infants, and (3) discussion on destigmatized, improved access to care.

The first key message emerging from these discussions was the importance of establishing a data element set, or standardized health care data vocabulary, for the longitudinal monitoring of clinical and supportive needs of mothers and children, as the primary strategy toward building frameworks for data standardization. Development of basic clinical tools using a common data element set for communicating longitudinal clinical care for mothers and children could begin at the individual electronic health record (EHR) level. This underlying premise leans on advancing health IT in support of clinicians’ diagnostic, therapeutic, and long-term care management in the EHR without adding to clinician burden. Starting with a standard mother–child data element set, the concept would be to align electronic clinical workflow to match clinician workflow for routine outpatient visits, such as maternal–infant continuity follow-up. Potential tool development in EHRs may include longitudinal clinical flowsheets, decision support, and population-based data analytics. Such innovations could unlock opportunities for quality improvement at the patient level, organizational level, and multicenter level.

Second, these discussions would not be complete without consideration of the role of interoperability in leveraging health IT to bridge data gaps in long-term dyadic care. Focusing on interoperable pediatric health records, potential data-sharing challenges may include the following clinical intersections: (1) routine prenatal care data with the inpatient newborn record, (2) inpatient newborn record with the outpatient pediatric record, and (3) outpatient pediatric records across EHRs. Some of these data-sharing challenges may be related to technology, and others may be related to policy. However, as one federal resource and viable roadmap, the ONC for Health IT’s 2015 Edition Health IT Certification Criteria and the Interoperability Standards Advisory focus on how EHRs collect, use, and exchange health data. ³ Accordingly, ONC’s proposed US Core Data for Interoperability includes standardized demographic and medical data (ie, diagnoses, medications, allergies, diagnostics/laboratory work, immunizations) that may serve as a basis for core data exchange among EHR systems.

Third, before data and information systems are engaged, at least 2 important items must be considered: (1) given broad diagnostic variability, establishing a standard clinical definition for NAS/NOWS and (2) defining purposes for privacy-protected data sharing within the health system, whether for the improvement of patient care, population-level health research, or both. Data sharing raises the issue of privacy protection of mothers and children and ensuring that both the Health Insurance Portability and Accountability Act of 1996 (HIPAA) Privacy Rule standards and the Confidentiality of Substance Use Disorder Patient Records, Title 42 Code of Federal Regulations (CFR) Part 2, requirements are met. The HIPAA Privacy Rule standards address the use and disclosure of individual health information and individuals’ privacy rights to understand and control how their health information is used. ⁴ The 42 CFR Part 2 regulations serve to protect patient records created by federally funded programs for the treatment of substance use disorder. ⁵ Important to these privacy discussions must be nonpunitive uses of any health-related data for the purpose of improving health care for mother, child, and any siblings. Of equal importance is an examination of bioethical considerations pertinent to preventing potential unintended consequences of public health policies and further stigmatization in an already vulnerable mother–child dyad population.

The fourth and perhaps most valuable result from these discussions is the development of a mother–child data element set based on experts’ recommendations of markers that are key to long-term needs assessment. HHS invited experts to suggest data elements that can support existing or emerging clinical priorities and the development of interoperable health IT tools for care delivery. Starting with a framework for longitudinal data categories (Figure), a modular set of mother–child data elements was created focusing primarily on clinical and supportive needs markers, with the ability for inclusion of research indicators and content updates. The set partly consists of data elements routinely encountered in pediatric primary care follow-up, starting with growth and development at the first newborn visit and continuing with comprehensive milestone assessment spanning the preschool, school-aged, and adolescent years. As such, it follows the existing format and flow of pediatric health maintenance visits. The maternal categories include (1) clinical data (eg, infectious disease, medication-assisted treatment, other prescribed medications, obstetric history, medical and psychiatric diagnoses) and (2) supportive care data (eg, demographic characteristics, social history, determinants of health, and administrative data). The pediatric categories include (1) clinical data (eg, birth history, positive laboratory tests, prescribed medications, nonpharmacologic care, developmental milestones, medical and psychiatric diagnoses) and (2) supportive care data (eg, social determinants of health, administrative data, referral to or use of services). Subsequent cross-mapping (not shown) of the full maternal–child data element set to the ONC 2015 Edition Health IT Certification Criteria for EHRs revealed that 67% of the proposed data elements is supported by current IT certification standards and represented in structured (defined) code sets in certified EHR technology (Figure). That is, most of the maternal–child data elements that experts recommended as important for long-term dyadic needs assessment already exist as structured data available in most EHRs across the country (nationwide, 96% of hospitals and 80% of office-based practices have certified EHR adoption). ^6,7 Given these findings, one important next step may be an examination of how electronic clinical workflow based on these supported data elements could be improved; that is, improving the flow of health care data already collected in the EHR as part of routine mother–child care. Most importantly, these findings represent untapped opportunities for the development of clinical tools that could leverage existing EHR architecture to improve long-term needs and outcomes assessment, for opioid/substance exposure and for a host of chronic conditions, other risk exposures, and social determinants of health.

Figure.

A framework for standardization of maternal–child data elements to support longitudinal needs assessment. (A) Priority data categories were proposed for consideration in long-term clinical and supportive needs assessment for the mother–child dyad. Within maternal and child categories, individual data elements (not shown here) were developed for each subcategory under Clinical and Supportive headings (total maternal data elements = 70, total child data elements = 110). (B) Proposed data elements were then individually cross-mapped (not shown) to national health information technology (IT) certification criteria for the electronic health record (EHR). ³ Shown here are the total percentages of proposed maternal and child data elements: supported by IT certification criteria and represented in structured (defined) code sets available in certified EHRs nationwide, 67% (maternal = 79%, child = 60%); not supported (as structured data) in certified EHR technology, 32% (maternal = 20%, child = 39%); and requiring special (maternal) consent (1%).

In considering long-term outcomes, it is essential that we seek effective approaches to integrate how social determinants, for both mother and infant, may be influential. Growing evidence indicates a disproportionate increase in the incidence rate of NAS/NOWS and opioid use disorder in rural vs urban communities in the United States; data from 2009-2015 of 580 counties in 8 US states indicate that higher long-term unemployment levels, particularly in remote rural counties, were associated with higher NAS/NOWS rates. ^8,9 A critical understanding of the principal social determinants of health underlying such disparities could shed light on better ways to address NAS/NOWS and opioid use disorder in communities that are highly affected by the opioid epidemic. Unique advantages of EHR-based data analytic methodologies may include opportunities for (1) focused, community-centric assessment of distinct public health gaps and social-supportive care needs for mothers and children and (2) real-time and long-term assessments of clinical management approaches to NAS/NOWS and opioid/substance use disorder.

Collectively, key components of long-term care include standardized clinical data elements and a dyadic approach centering on maternal support and childhood development with (1) education to decrease stigma about opioid use disorder; (2) improved access to opioid use disorder medication-assisted treatment, recovery, and family-focused supportive social services; and (3) enhanced pain management and opioid use disorder prevention services. ^10,11 With the urgency of issues affecting maternal and infant populations, this national opioid crisis illustrates the importance of understanding and improving long-term dyadic care. Beginning with opioid exposure as one important risk marker, universal and unbiased mechanisms for longitudinal needs assessment for the dyad are vital. Interoperable approaches toward developing EHR clinical tools may offer one viable pathway. Notably, before considering the creation of any IT tool, standard clinical terminology is needed so that health care providers are speaking the same language around opioid exposure and other biological, social, and environmental risk factors for the dyad. Through the dedicated efforts of experts across the country, the data element set developed as part of this HHS initiative on NAS/NOWS is an important step in this direction and offers a primer for clinical tool development.