Standardization of State Definitions for Neonatal Abstinence Syndrome Surveillance and the Opioid Crisis

Abstract

Rates of neonatal abstinence syndrome (NAS) have increased fivefold in the past decade. To address this expanding and complex issue, state public health agencies have addressed the opioid crisis affecting newborns in diverse ways, leading to a variety of methods to quantify the burden of NAS.

In an effort to understand this variability, we summarized clinical case and surveillance definitions used across jurisdictions in the United States. We confirmed that the rapid progression of the nation’s opioid crisis resulted in heterogeneous processes for identifying NAS. Current clinical case definitions use different combinations of clinician-observed signs of withdrawal and evidence of perinatal substance exposure. Similarly, there is discordance in diagnosis codes used in surveillance definitions. This variability makes it difficult to produce comparable estimates across jurisdictions, which are needed to effectively guide public health strategies and interventions.

Although standardization is complicated, consistent NAS definitions would increase comparability of NAS estimates across the nation and would better guide prevention and treatment efforts for women and their infants.

Neonatal abstinence syndrome (NAS) can occur following in utero exposure to several substances, including barbiturates, benzodiazepines, and opioids, and is characterized by signs of withdrawal, such as irritability, tremors, and poor feeding.¹ In the past 10 years, rates of newborns diagnosed with NAS have risen fivefold in the United States, from 1.5 per 1000 hospital births in 2004 to 8.0 in 2014.² Increased incidence is attributed to rising maternal opioid use, including use of prescription opioids as prescribed, misuse of prescription opioids, use of illicit opioids (e.g., heroin), and use of medication-assisted treatment of opioid use disorder.¹ Among pregnant women, opioid use disorder has increased from 1.1 per 1000 delivery hospitalizations in 2000 to 6.5 in 2014.³ NAS-related hospital costs totaled $563 million in 2014 and $3.1 billion over 10 years (2004–2014).² Data on health outcomes attributable to perinatal opioid exposure and NAS are limited, and further research on the long-term developmental outcomes of newborns diagnosed with NAS is needed.^1,4

As the opioid crisis continues, NAS remains a public health concern, and state agencies have begun to quantify their NAS burdens using different methods. Some states, such as Tennessee and Arizona, have made NAS a reportable condition to facilitate timely data collection, whereas others, such as Alaska and New York, are using existing hospital discharge data for passive surveillance. Despite these efforts, several challenges remain in accurately measuring NAS and the burden of NAS related to opioid exposure in pregnancy, which are needed to effectively guide public health strategies and interventions. The primary challenge, as we noted, is that withdrawal signs in neonates can be attributed to many different substances, including opioids.¹ Because polysubstance use is common, the identification of withdrawal signs exclusively related to opioid use is complicated.⁵ Additionally, because NAS is a syndrome, there is a wide range of signs that a neonate could display, and clinical assessment of these signs is not standardized across practices.⁵ Furthermore, not all substance-exposed newborns experience withdrawal, necessitating differentiation between cases of perinatal opioid exposure and newborns with NAS.¹

Taken together, these obstacles have led to different clinical definitions of NAS, as well as variation in both the assignment of International Classification of Diseases, Clinical Modification (ICD-CM)^6,7 diagnosis codes for NAS and the application of these diagnosis codes to conduct NAS surveillance. Although these data are commonly used for surveillance purposes, ICD-CM diagnosis codes related to neonatal withdrawal did not specify the source of substance exposure before 2018. Thus, without clear definitions to adopt and little guidance to follow, state public health agencies’ efforts to quantify the burden of NAS and sources of exposure have varied.

To further understand the variability in what is considered an NAS case, we summarized clinical case and surveillance definitions currently being used across jurisdictions in the United States, regardless of whether NAS is a state reportable condition. This information may be useful to states as they consider initiating NAS surveillance or adding NAS to their lists of reportable conditions.

METHODS

Three independent reviewers conducted Web-based reviews of publicly available information on current and past state-level NAS surveillance activities, clinical case definitions, and surveillance definitions for each state in July and December 2018. We collected details on current state surveillance activities by reviewing each state’s public health agency Web site with specific searches for the following: NAS surveillance reports, reportable condition lists and reporting instructions, opioid data dashboards, and opioid-related press releases. We did not include facility-level or hospital system reports concerning NAS unless they were associated with a state’s public health agency.

We used results of the initial search efforts to further expand the search strategy by pairing the search terms “neonatal abstinence syndrome,” “neonatal withdrawal,” “perinatal substance exposure,” “779.5,” and “P96.1” with each state in a Google search. Additionally, to capture all definitions states use, we reviewed state perinatal quality collaborative Web sites for NAS-related initiatives related to the development of clinical case definitions in collaboration with state public health agencies. As this review was focused on state-level activities, we did not include regional perinatal quality collaboratives in the search. Each reviewer independently coded the collected information, and we resolved discrepancies in data collection and coding by consensus.

We defined clinical case definitions as criteria provided by state public health agencies to help health care providers, across their respective states, identify NAS in hospitalized neonates, and they could include maternal or neonatal elements. For instance, if a state agency noted that maternal history of prescription or nonprescription opioid use was required to diagnose NAS then we considered it a maternal element of the clinical case definition. Similarly, if a state agency noted that clinical signs of withdrawal were needed to diagnose NAS, we listed it as a neonatal element of the clinical case definition.

We defined surveillance definitions as criteria on the basis of ICD-CM diagnosis codes used to identify NAS cases from registry, administrative, or surveillance system data. If a state agency noted the use of any diagnosis codes as criteria for NAS case identification, we listed those respective codes as its surveillance definition (all abstracted ICD-CM diagnosis codes we used for surveillance are detailed in Table 1). When a state differentiated between NAS and perinatal substance exposure surveillance definitions, we also noted this. If we identified multiple definitions for a state because of changing surveillance methodologies, we included the definition related to the most recent activity. In states where surveillance activities employed hospital discharge data but we could not identify a surveillance definition, we considered the state’s surveillance definition unspecified.

TABLE 1—

International Classification of Diseases, Clinical Modification (ICD-CM) Diagnosis Codes Used by US States for Neonatal Abstinence Syndrome or Perinatal Substance Exposure Surveillance Activities: July and December 2018

Diagnosis Code	Definition
ICD-9-CM
Maternal
292.0	Drug withdrawal
304.0	Opioid type dependence
304.1	Sedative, hypnotic, or anxiolytic type dependence
304.7	Combinations of opioid type drug with any other drug dependence
304.8	Combinations of drug dependence excluding opioid type drug
304.9	Unspecified drug dependence
305	Nondependent abuse of drugs
305.5	Nondependent opioid abuse
305.6	Nondependent cocaine use
V3×	Liveborn infants according to type of birth
V58.69	Long-term use of other drugs
630–679	Complications of pregnancy, childbirth, and the puerperium
Neonatal
760.70	Unspecified noxious substance affecting fetus or newborn via placenta or breast milk
760.71	Alcohol affecting fetus or newborn via placenta or breast milk
760.72	Narcotics affecting fetus or newborn via placenta or breast milk
760.73	Hallucinogenic agents affecting fetus or newborn via placenta or breast milk
760.75	Cocaine affecting fetus or newborn via placenta or breast milk
779.5	Drug withdrawal syndrome in newborn
Exclusion
765.01–765.05	Extreme immaturity, 0–1499 grams
770.7	Chronic respiratory disease arising in the perinatal period
772.1×	Intraventricular hemorrhage of fetus or newborn
777.5×	Necrotizing enterocolitis in newborn
777.6	Perinatal intestinal perforation
779.7	Periventricular leukomalacia
ICD-10-CM
Maternal
F11	Opioid-related disorders
F11.20	Opioid dependence, uncomplicated
F12	Cannabis-related disorders
F13	Sedative-, hypnotic-, or anxiolytic-related disorders
F13.20	Sedative, hypnotic, or anxiolytic dependence, uncomplicated
F14	Cocaine-related disorders
F15	Other stimulant-related disorders
F16	Hallucinogen-related disorders
F17	Nicotine dependence
F18	Inhalant-related disorders
F19	Other psychoactive substance–related disorders
Z38	Liveborn infants according to place of birth and type of delivery
Neonatal
P04.4	Newborn affected by maternal use of drugs of addiction
P04.49	Newborn affected by maternal use of other drugs of addiction
P96.1	Neonatal withdrawal symptoms from maternal use of drugs of addiction
P96.2	Withdrawal symptoms from therapeutic use of drugs in newborn
Exclusion
P27×	Chronic respiratory disease originating in the perinatal period
P52×	Intracranial nontraumatic hemorrhage of newborn
P77×	Necrotizing enterocolitis of newborn
P78.0×	Perinatal intestinal perforation
P91.2×	Neonatal cerebral leukomalacia

Source. National Center for Health Statistics.^6,7

We used descriptive statistics to analyze the abstracted information. We report counts and percentages of states with identified clinical case or surveillance definitions and describe variations in these definitions.

RESULTS

We identified 44 states (88%) with NAS or perinatal substance exposure–related surveillance and data activities through our Web-based search. Of these, 7 (16%) had specified clinical case definitions only, 22 (50%) had specified surveillance definitions only, and 3 (7%) had both specified clinical case and surveillance definitions. In two states with surveillance activities (Illinois and Utah), a clinical case definition was specified, but we did not identify the surveillance definition in our search. Additionally, we found surveillance activities without corresponding specified surveillance definitions for 10 states (23%). Nine states (20%) had mandated NAS or perinatal substance exposure as a reportable condition for public health surveillance at the time of our search (Table 2).

TABLE 2—

US State-Level Neonatal Abstinence Syndrome and Perinatal Substance Exposure Surveillance Activities by Reporting Requirements and Use of Clinical Case or Surveillance Definitions: July and December 2018

Activity or Type of Definitions	No. of States	States
Any NAS or perinatal substance exposure–related surveillance and data activities	44	AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NH, NJ, NM, NY, NC, ND, OH, OK, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI
NAS or perinatal substance exposure reportable condition	9	AZ, FL, GA, KY, MA, OH, PA, TN, VA
Clinical case definition only	7	AZ, GA, IN, KY, PA, TN, WV
Surveillance definition only	22	AK, CA, CO, CT, FL, IA, KS, LA, ME, MA, MI, MN, MS, MO, MT, NM, NY, OK, RI, SD, VT, WI
Clinical case definition and surveillance definition	3	DE, OH, VA
Surveillance activities with unspecified surveillance definition	10	AL, AR, MD, NH, NJ, NC, ND, SC, TX, WA
Clinical case definition and unspecified surveillance definition	2	IL, UT
No identified activities	6	HI, ID, NE, NV, OR, WY

Note. NAS = neonatal abstinence syndrome. Results are from information gathered from publicly available, Web-based sources. Researchers did not directly contact states to confirm or obtain additional information regarding activities, reporting requirements, or definitions. Additionally, categories are not mutually exclusive. States can be included in more than 1 category.

Clinical Case Definitions

Details for each state’s clinical case definition can be found in Table A (available as a supplement to the online version of this article at http://www.ajph.org). Of the 12 states with identified clinical case definitions, 11 require diagnosed neonates to display behavioral signs consistent with substance withdrawal (e.g., irritability, tremors, and poor feeding).¹ Of the 12 states, only Georgia allows either clinical signs or evidence of perinatal substance exposure from a positive newborn drug screen to meet required diagnosis elements. Six states (Delaware, Illinois, Indiana, Ohio, Tennessee, and Utah) instruct health care providers to use a standardized scoring instrument, such as the Finnegan Scoring Tool, to help diagnose and quantify withdrawal sign severity.

Two states (Arizona and Indiana) require evidence of perinatal substance exposure in addition to neonatal withdrawal signs. Perinatal substance exposure evidence thresholds vary widely and are not identical across any two states (Table A). For example, a maternal history of perinatal opioid use meets reporting criteria in Arizona, but a positive newborn substance screening test is necessary for evidence of perinatal substance exposure in Georgia. Indiana allows either positive maternal verbal history, maternal toxicology tests, or neonatal toxicology tests as evidence. Clinical case definitions in Pennsylvania and West Virginia also necessitate perinatal substance exposure, but neither state requires reporters to document evidence of exposure or specifies whether exposure should be supported by maternal or neonatal elements. Kentucky requires either a history or suspicion of perinatal substance exposure. In Tennessee, only clinical withdrawal signs are necessary to establish a clinical case of NAS; neonatal screening or confirmatory substance testing and documented maternal opioid use are noted as supportive elements of a NAS diagnosis.

Additionally, states differ in the type of perinatal substance exposure they indicate as supporting a NAS diagnosis. For example, in Arizona and Pennsylvania, diagnosis of NAS is specific to perinatal exposure to opiates, whereas in West Virginia, NAS can result from any neuroactive substance. NAS clinical case definitions in Georgia and Tennessee refer to general substance use, and neither Indiana nor Kentucky specify type of perinatal substance use.

Surveillance Definitions

Definitions for maternal and neonatal ICD-CM diagnosis codes used for NAS surveillance by states are detailed in Table 1.^6,7 Each of the 25 state surveillance definitions include the newborn substance withdrawal codes ICD-9-CM 779.5 or ICD-10-CM P96.1. Surveillance definitions for 18 (72%) of the 25 states (Alaska, Colorado, Connecticut, Delaware, Florida, Kansas, Louisiana, Maine, Michigan, Minnesota, Montana, New Mexico, Ohio, Oklahoma, Rhode Island, Vermont, Virginia, and Wisconsin) include newborn substance withdrawal codes alone, whereas definitions for 7 (28%) of the remaining states (California, Iowa, Massachusetts, Mississippi, Missouri, New York, and South Dakota) include additional diagnosis codes associated with perinatal substance exposure (Table A).

NAS surveillance definitions that include perinatal substance exposure diagnosis codes can be further broken down into two groups: states where NAS is related specifically to maternal opioid use and states where NAS is also related to general perinatal substance use in addition to maternal opioid use. In California, Mississippi, New York, and South Dakota, neonates with opioid-related diagnosis codes ICD-9-CM 760.72 and ICD-10-CM P04.49 are classified as NAS cases. By contrast, in Iowa and Missouri, diagnosis codes related to generalized maternal substance use or other types of perinatal substance exposure are also included in surveillance definitions (Table A).

Massachusetts has created separate surveillance definitions for perinatal substance exposure, differentiating between neonatal and maternal diagnosis codes. Cases of neonatal withdrawal are identified using ICD-10-CM P96.1 or P04.49, whereas neonates delivered to mothers with drug dependence are identified using maternal codes related to opioid, benzodiazepine, and opioid combination dependence (Table A). South Dakota similarly differentiates between NAS, infant drug exposure, and maternal opioid use by assigning distinct surveillance definitions. Cases of infant drug exposure include the NAS-related codes ICD-9-CM 779.5 and 760.72 in addition to diagnosis codes related to perinatal hallucinogenic and cocaine exposure. Maternal opioid use is identified from maternal substance dependence and abuse diagnosis codes (Table A).

Nuances in a few surveillance definitions should be noted. Missouri is the only state to include ICD-10-CM diagnosis code P96.2 in its NAS surveillance definition. Furthermore, three states (Kansas, Louisiana, and Vermont) exclude diagnosis codes associated with iatrogenic withdrawal signs from NAS surveillance definitions (Table A). Iatrogenic withdrawal signs are typically seen in neonates treated with opioids while in intensive care treatment.⁸ Finally, three states use alternative terminology to refer to cases of neonatal withdrawal. Iowa specifically refers to “newborns with in utero exposure to opioids and other substances” rather than NAS. In Maine, the phrasing used is “drug-affected newborns,” whereas in Vermont they are surveilled as “opioid-exposed infants.”

DISCUSSION

We have presented a comprehensive summary of NAS definitions examining clinical case and surveillance definitions used across jurisdictions. We confirmed that the rapid progression of the nation’s opioid crisis has resulted in heterogeneous processes for identifying NAS-affected neonates.⁹ Our search found that clinical case definition criteria vary across states because of different combinations of withdrawal signs and evidence of perinatal substance exposure specified for health care providers in identifying NAS cases. Definition variation persists among states where NAS is a reportable condition. Similarly, definitions vary for the 25 states using ICD-CM diagnosis codes to identify NAS for surveillance.

Discordance in NAS clinical case and surveillance definitions may result in erroneous comparisons and conclusions that could negatively affect public health responses. For instance, a state that identifies either newborns with withdrawal signs or newborns with perinatal substance exposure as NAS cases will estimate a different NAS burden than will a state that identifies NAS cases as only newborns with withdrawal signs or a state that requires cases to have both withdrawal signs and evidence of exposure. Flawed comparisons could then result in investments of limited resources in jurisdictions with the greatest estimated NAS burdens when, in fact, demand is higher elsewhere.

Our review suggests that standard NAS clinical case and surveillance definitions should be considered, as the current myriad of definitions makes it difficult to produce comparable estimates across jurisdictions. Variation in definitions likely reflects differences in data availability and jurisdiction-specific interests regarding burden. For instance, some jurisdictions may wish to capture all substance-exposed infants, whereas others may want to capture only infants exposed to opioids with signs of NAS, resulting in different populations of newborns being surveilled. A consistent standardized definition could be supplemented with additional definitions to meet the needs of local jurisdictions.

By harmonizing definitions and increasing comparability of NAS estimates across the country, public health agencies can better guide prevention and treatment of women and their infants. During the time of this research, the Council for State and Territorial Epidemiologists convened a workgroup to develop a standard case definition for NAS. Council members recently voted to adopt this standardized definition during the council’s annual conference in June 2019. Similar standardization initiatives of the council have been helpful in addressing other public health concerns, such as elevated blood lead levels and hepatitis C.¹⁰

Furthermore, guidance for the appropriate coding of NAS diagnoses could also be considered in subsequent revisions to the ICD-10-CM official guidelines for coding and reporting.¹¹ For instance, in addition to establishing a NAS clinical case definition, West Virginia provided instructions to health care providers and medical coders on how to assign perinatal substance exposure and NAS diagnosis codes.¹² Similar guidance could be useful in states where NAS is not a reportable condition, and surveillance depends on registry, administrative, and surveillance system data.

This review was limited, as researchers used publicly available information and did not directly contact states to confirm or obtain additional information regarding NAS surveillance activities and definitions. However, the Council for State and Territorial Epidemiologists is conducting a rigorous environmental scan, which will include activities and definitions directly obtained from all states. We excluded unpublished definitions, and we did not include changes occurring in state policies following the data collection phase in December 2018. Further, in some states, county or regional public health departments may have different definitions or reporting criteria, creating even greater variation in NAS clinical and surveillance definitions. The researchers interpreted publicly available information; it is possible that we misclassified some definitions in the review process.

As the opioid crisis continues to demand attention, there is a pressing need for timely and accurate data to help public health practitioners characterize the full scope of the NAS burden and to provide comparable estimates for developing effective interventions and determining where and for whom prevention and treatment efforts are best focused. This summary provides an understanding of the different clinical case and surveillance definitions used across the United States, a first step toward convening jurisdictions to discuss standardizing NAS definitions.