Association of Criminal Statutes for Opioid Use Disorder With Prevalence and Treatment Among Pregnant Women With Commercial Insurance in the United States

IMPORTANCE Inadequatetreatmentofopioidusedisorder(OUD)inpregnantwomenincreasesthe risk of life-threatening consequences on maternal and fetal outcomes. Untreated OUD during pregnancy is associated with higher rates of adverse outcomes among newborns. OBJECTIVE To examine the variation in the prevalence of OUD and the use of medication-assisted treatment among commercially insured pregnant women according to region and state legislature. DESIGN, SETTING, AND PARTICIPANTS Cohort study in which the patient cohort used was derived from a 10% random sample of enrollees within the IQVIA PharMetrics Plus adjudicated claims and enrollment database from 2007 to 2015. The database consists of a 10% random sample of private health insurance recipients in the United States and contains claims and enrollment data that are representative of the commercially insured US population. The cohort comprised women (n = 110285) between 18 and 45 years of age with a code indicating a delivery and continuous insurance enrollment 9 months before and 12 months after delivery. Data analysis was performed from December 2017 to May 2018. CONCLUSIONS AND RELEVANCE These results appeared to show significant variations in the patterns of OUD diagnosis and receipt of medication-assisted treatment among pregnant women, suggesting the need to further explore the source of these variations. receipt of treatment did not. The proportion of women with OUD and receiving treatment within this insured population wasrelativelylowcomparedwithwomeninotherpopulations.Thesefindingsindicateadiscrepancy in the diagnosis and treatment of OUD both on a regional and a legal basis. This study demonstrates the need for further studies into policies that reduce stigma and discrimination and that encourage the proper identification of OUD during pregnancy as part of a national quest to address the opioid epidemic among society’s most vulnerable women.


Introduction
In the United States, opioid misuse is a growing and prevalent concern to the care and well-being of pregnant women as it can lead to adverse maternal and neonatal outcomes. Opioid use among women of childbearing age (15-44 years) has reached epidemic proportions, and a substantial increase in opioid use among pregnant women has been reported. 1,2 Opioid use among pregnant women was approximately 5.6 per 1000 live births in 2012. 3,4 In a previous study of the Optum commercially insured population, an opioid was dispensed to 14.4% (1 in every 7) of pregnant women and 2.2% were given an opioid 3 or more times during pregnancy, with some variation by region and state between 2005 and 2011. 5 With the rising opioid use in this population comes a concurrent increase in opioid use disorder (OUD). Opioid use disorder is defined in the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) as a problematic pattern of opioid use resulting in symptoms such as tolerance, withdrawal, craving, or an inability to cut down or control opioid use. 6,7 Among all women hospitalized for delivery, opioid dependence increased from 0.17% in 1998 to 0.39% in 2011, representing a 127% increase. 8 In pregnancy, opioid withdrawal is associated with decreased neonatal birth weight; illicit drug use; relapse; and resumption of high-risk behaviors such as intravenous drug use, prostitution, and criminal activity. 6,9 Medication-assisted treatment, specifically methadone hydrochloride or buprenorphine hydrochloride, in combination with behavioral and psychosocial counseling is the standard practice for treating pregnant and postpartum women with OUD. Medication-assisted treatment minimizes opioid withdrawal, reduces risk-taking behavior, and decreases the risk of acquisition and transmission of infectious diseases. [10][11][12] Various barriers exist for pregnant women to receive these treatments, such as accessibility, insurance coverage, region, and state laws that permit charges against pregnant women with OUDs. Such barriers are associated with a substantially decreased use of medication-assisted treatment.
Inadequate treatment of OUD in pregnant women is associated with increases in the risk of lifethreatening consequences on maternal and fetal outcomes. [13][14][15] Untreated OUD during pregnancy is associated with higher rates of low birth weight in the newborn, fetal growth restriction, and continued use of opioids by pregnant women. 16 Exposure to opioids in utero places infants at risk for neonatal abstinence syndrome, a drug withdrawal syndrome that causes severe withdrawal symptoms, such as tremors, irritability, and respiratory distress, almost immediately after birth. 17 The incidence of neonatal abstinence syndrome in newborns increased nationally from 3.4 to 5.8 per 1000 hospital births from 2009 to 2012. 3 This spike represents an increased cost of $43 900 per hospital birth. 4 Pregnant women with OUDs face not only medical consequences, such as an increased risk of obstetric morbidity and mortality, but also a predisposition to the potential loss of child custody and even criminalization in some states. 8,18 A growing number of states over the past several years have debated and passed legislation criminalizing women with substance use disorders. Overall, 18 states since 2012 have required health professionals to report substance use disorder in pregnant women and have established civil or criminal laws that consider substance use to be child abuse. 19 Escalating trends in OUD during pregnancy and the associated adverse maternal and neonatal outcomes have captured the attention of policymakers. The most recent initiatives are directed toward improving the accessibility and affordability of substance use treatment services, such as medication-assisted treatment for pregnant women. 20 Most studies evaluating the prevalence of OUD and the use of medication-assisted treatment among pregnant women have either involved solely the Medicaid population or have shown that most women are covered by Medicaid. 4,8,19 No studies thus far have shown the variations in prevalence of OUD and the use of medication-assisted treatment among pregnant women within a commercially insured population.

JAMA Network Open | Substance Use and Addiction
The objective of this study was therefore to examine the variation in the prevalence of OUD and the receipt of medication-assisted treatment among commercially insured pregnant women according to region and state legislature. We hypothesized that the diagnosis of OUD and receipt of treatment would vary by region and state legislature. Evaluating the prevalence of OUD and the receipt of medication-assisted treatment is important to define the current practices among commercially insured individuals and to inform future research within this population in the context of policy on civil or criminal laws on substance use.

Study Design and Setting
This retrospective observational cohort study included pregnant women with commercial insurance across the United States. Pregnant women were assessed for the outcomes 9 months before and 12 months after their recorded delivery date. The data obtained were deidentified, and the study was

Study Data
The patient cohort used in this study was derived from a 10% random sample of enrollees within the

Identification of Sample
The study period spanned from June 30, 2007, to June 30, 2015. Female patients younger than 18 years and women older than 45 years were excluded from the cohort. To identify pregnancy, we collected 2995 ICD-9-CM codes and 88 CPT-4 codes related to pregnancy. Codes were identified using published literature [22][23][24][25] and the IQVIA PharMetrics Plus data dictionary. Based on the descriptions of the collected ICD-9-CM and CPT-4 codes, the codes were classified into 1 of 8 to be continuously enrolled 9 months before the index date and 12 months after the index date to identify the outcomes.

JAMA Network Open | Substance Use and Addiction
The primary outcome was a diagnosis of OUD during the 9 months before the index date. 26 From the published literature, we used 20 ICD-9-CM codes to identify an OUD. 27 Receipt of medication-assisted treatment was observed 9 months before the index date to reflect the mean duration of pregnancy and 12 months after the index date. The postdelivery time frame of 12 months was chosen to allow ample time after delivery to capture receipt of treatment.
This time frame was primarily a concern for women who lived in states with statutes criminalizing  19 The second population resided in states without these statutes.

Statistical Analysis
Data analysis was performed from December 2017 to May 2018. The prevalence of OUD diagnosis in the commercially insured US population was calculated using the unique number of pregnant women between 18 and 45 years of age with an OUD diagnosis during the 8-year study period as the numerator and the total number of pregnant women within the age restriction as the denominator.
The prevalence of treatment was calculated in a similar manner. Prevalence calculations were stratified by region and by the presence of legislation that criminally or civilly prosecutes pregnant women with an OUD diagnosis. All states categorized as having such legislation had passed laws before 2012 and had cases convicting women before 2012. 19 Cochran-Mantel-Haenszel (χ 2 ) statistics were calculated to determine whether a substantial difference exists between the prevalence of OUD diagnoses and receipt of treatment by region and by presence of civil or criminal statutes. Statistical significance was determined at 2-sided P = .05 using a 2-sided Pearson χ 2 test. All statistical analyses were performed in SAS, version 9.4 (SAS Institute Inc).

Results
Of a total 12 416 600 individuals in the commercially insured population, 2 683 387 (21.6%) were women between 18 and 45 years of age, of whom 295 837 (11.0%) had a code for a delivery date. The final cohort consisted of 110 285 pregnant women between 18 and 45 years of age who were continuously enrolled 9 months before the delivery date and 12 months after the delivery date ( Figure 1) Of the 44 683 women living in states in which women with OUD diagnoses were civilly or criminally prosecuted, 81 (0.07%) had a diagnosis code for OUD and 129 (0.12%) received treatment.

Discussion
These results differ from findings in a similar study by Angelotta et al, 19 who found that criminal or civil statutes did affect whether a pregnant woman would receive treatment. Note that the population in the Angelotta et al 19 study included a Medicaid population, in which OUD was more prevalent and treatment more likely to be covered than in commercially insured populations, such as this study population. Medicaid patients are twice as likely as patients with no insurance or private insurance to receive OUD treatment. 29 Bateman et al 5 explored the Optum data and found results similar to those in the present study, showing a low prevalence of methadone and buprenorphine dispensing for commercially insured pregnant women. Bateman et al 5 found that buprenorphine was dispensed to 0.03% of women during pregnancy and methadone was dispensed to 0.02% of women in pregnancy. Among the women exposed to opioids during pregnancy, 0.5% had a diagnosis of opioid dependence and 0.02% had an OUD diagnosis. 5 A notable finding in the present study pertains to the analysis by region and by civil and criminal statutes. Regional variations may account for the limited availability of treatment centers or practitioners specializing in medication-assisted treatment within a region. This lack would indicate that even if practitioners who treat the commercially insured have good practices and want to provide treatment, they may not be able to send women to receive care. We found a lower prevalence of recorded OUD and lower prevalence of medication-assisted treatment in states with statutes civilly or criminally prosecuting women with an OUD diagnosis.
These results show a statistically significant association between the presence of these statutes and a lower prevalence of recorded OUD. The low prevalence of OUD in the states with statutes may be the result of fewer women seeking care for OUD, perhaps out of the fear of being criminally charged and losing custody of their newborn. In addition, the same fear compounded by stigma may lead some women to avoid receiving care from medical practitioners, even when they have insurance coverage. One study that interviewed pregnant women who were substance users reported that 73% feared being identified as substance users, and many often used tactics, such as skipping appointments or forgoing prenatal care altogether, to avoid any detection of their drug use. 15

Limitations and Strengths
Even as we note the importance of these findings, we also advise that they should be carefully considered because of the inherent limitations of claims data. First, practitioners may be hesitant to code for an OUD diagnosis for a number of reasons, including the concern that the patient may face consequences and may no longer seek prenatal care. Second, information in commercial claims databases is limited. The database does not contain information on the gestational age at birth. Given that women with OUD are more likely to give birth preterm, the look-back period of 9 months from delivery may yield a conservative estimate and may be longer than some pregnancies. 8  However, despite the challenges of identifying pregnancy and treatment frequency in claims databases, several strengths of this study should be noted. This study is the first, to our knowledge, to attempt to identify the OUD prevalence and treatment during pregnancy within the IQVIA PharMetrics Plus database commercially insured population. We used a novel method of identifying pregnancy within claims and found a possible underdiagnosis and treatment of OUD among pregnant women.
The most striking finding and contribution from this study to an ongoing effort to address the opioid epidemic is that the criminality statute is the basis of geographic variations. The process leading to the unintended consequences of such statutes is very complex and should be examined very carefully lest it produces the opposite of its intended purpose. The geographic variations indicate that differential factors within regions may play a role in the diagnosis of OUD among pregnant women.

Conclusions
The prevalence of OUD in commercially insured pregnant women and receipt of treatment within the United States appeared to vary by region. The diagnosis of OUD also significantly varied by the presence of criminal or civil statutes within the state of residence, whereas the receipt of treatment did not. The proportion of women with OUD and receiving treatment within this insured population was relatively low compared with women in other populations. These findings indicate a discrepancy in the diagnosis and treatment of OUD both on a regional and a legal basis. This study demonstrates the need for further studies into policies that reduce stigma and discrimination and that encourage the proper identification of OUD during pregnancy as part of a national quest to address the opioid epidemic among society's most vulnerable women.