If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence to: Katy B. Kozhimannil, PhD, MPA, Division of Health Policy and Management, University of Minnesota School of Public Health, 420 Delaware Street SE, MMC 729, Minneapolis, MN 55455. Phone: 612-626-3812; fax: 612-624-2196.
Department of Pediatrics, Vanderbilt University School of Medicine, Nashville, TennesseeDepartment of Health Policy, Vanderbilt Center for Health Services Research, Vanderbilt University School of Medicine, Nashville, Tennessee
Nonmedical use of opioids during pregnancy is associated with adverse outcomes for women and infants, making it a prominent target for prevention and identification. Using a nationally representative sample, we determined characteristics of U.S. pregnant women who reported prescription opioid misuse in the past year or during the past month.
Methods
We used data from the National Survey on Drug Use and Health (2005–2014) in a retrospective analysis. The sample included 8,721 (weighted n = 23,855,041) noninstitutionalized women, ages 12 to 44, who reported being pregnant when surveyed. Outcomes were nonmedical use of prescription opioid medications during the past 12 months and during the past 30 days. Multivariable logistic regression models were created to determine correlates of nonmedical opioid use after accounting for potential confounding variables.
Results
Among pregnant women in the United States, 5.1% reported nonmedical opioid use in the past year. In adjusted models, depression or anxiety in the past year was strongly associated with past year nonmedical use (adjusted odd ratio [AOR], 2.15; 95% CI, 1.52–3.04), as were past year use of alcohol (AOR, 1.56; 95% CI, 1.11–2.17), tobacco (AOR, 1.72; 95% CI, 1.17–2.53), and marijuana (AOR, 3.44; 95% CI, 2.47–4.81). Additionally, 0.9% of U.S. pregnant women reported nonmedical opioid use in the past month. Past year depression or anxiety and past month use of alcohol, tobacco, and marijuana each independently predicted past month nonmedical use.
Conclusions
Characteristics associated with nonmedical opioid use by pregnant women reveal populations with mental illness and co-occurring substance use. Policy and prevention efforts to improve screening and treatment could focus on the at-risk populations identified in this study.
Over the last decade, prescription opioid use for nonmedical purposes during pregnancy nearly doubled (
Prevalence of hospitalized live births affected by alcohol and drugs and parturient women diagnosed with substance abuse at liveborn delivery: United States, 1999-2008.
Maternal and Child Health Journal.2013; 17: 667-676
). More than 20% of Medicaid beneficiaries filled an opioid prescription during pregnancy, with substantial and growing implications for the health of women and infants (
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
). In a recent national study, approximately 22% of women who reported past year nonmedical opioid use also met diagnostic criteria for opioid use disorder (
), which carries heightened negative consequences generally and during pregnancy. Opioid dependency may place a pregnant woman at greater risk of experiencing violence, illness (such as hepatitis C, HIV, or other sexually transmitted infections), as well as other high risk behaviors (
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
). Chronic nonmedical opioid use in pregnancy is associated with increased risk of newborn withdrawal, known as neonatal abstinence syndrome (NAS), and preterm birth, the largest contributor to infant mortality (
). As the opioid epidemic has impacted more pregnant women, there has been a concomitant increase in the number of infants diagnosed with the syndrome. From 2000 to 2012, the national NAS rate increased from 1.2 to 5.8 per 1,000 live hospital births, resulting an estimated $1.5 billion in hospital charges (
There is a need to better understand the antecedent factors that lead to nonmedical opioid use by pregnant women to inform prevention, screening, and treatment efforts (
Substance Abuse and Mental Health Services Administration (SAMHSA) Results from the 2013 National Survey on Drug Use and Health: Summary of National Findings. NSDUH Series H-48, HHS Publication No. (SMA) 14-4863.
), these reports describe broad trends. They do not identify groups at higher risk for prenatal substance use, nor do they specify individual characteristics that may predict nonmedical use of prescription opioids. Studies of hospital discharge data report trends in and correlates of maternal diagnosis of substance abuse (including opioid dependency) complicating childbirth (
), but these studies only capture cases that were diagnosed and documented at the time of childbirth. They leave open the question about the scope of the problem of nonmedical opioid use before and during pregnancy. Women are highly motivated to care for their health during pregnancy, and they typically interact with clinicians on a regular basis, making this time period a unique window of opportunity for screening, diagnosis, and treatment.
The effects of nonmedical opioid use on women and infants have a high potential for prevention and appropriate management(
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
). New prescribing guidelines raise awareness about the risks associated with prescription opioid use during pregnancy, but offer few specific recommendations (
). For pregnant women with opioid dependency, comprehensive treatment, including medication-assisted therapy combined with adequate prenatal care, reduces the risk of obstetric complications and of NAS (
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
). To develop strategies to prevent and appropriately treat nonmedical use of prescription opioids during pregnancy, there is a need for greater precision in knowledge about the correlates of nonmedical opioid use before and during pregnancy. There is limited national data on pregnant women who report nonmedical use of prescription opioids, and this knowledge gap may constrain current policy efforts to address NAS and other effects of the opioid epidemic. The goal of this study was to use national survey data to describe characteristics of pregnant women who reported nonmedical opioid use in the prior year or over the past month.
Materials and Methods
Data and Study Population
We analyzed data from the National Survey of Drug Use and Health (NSDUH). The NSDUH provides population estimates of substance use and health-related behaviors in U.S. adolescents and adults using multistage area probability sampling. We pooled 10 years of data (2005–2014) to create a sample of female respondents who reported being currently pregnant at the time of the survey. The sample included 8,721 (weighted n = 23,855,041) noninstitutionalized pregnant women ages 12 to 44. Descriptive characteristics and model estimates were weighted to be nationally representative, accounting for the NSDUH sample design.
Variable Measurement
The dependent variables were based on a pregnant respondent's having reported using prescription opioids nonmedically or “for the feeling it caused.” Past year nonmedical opioid use occurred either “within the past 30 days” or “more than 30 days ago but within the past 12 months.” Past month use was coded among those with nonmedical opioid use only “within the past 30 days.” The NSDUH survey does not explicitly ask about the frequency or duration of substance use, nor does the survey ask if use was during pregnancy. We expect that most reports of past month nonmedical opioid use reflected use during pregnancy for many of those in the study population (who reported being currently pregnant at the time of the survey), because most pregnancies cannot be detected or confirmed before day 20 of the pregnancy.
An indicator for anxiety or depression was included if the respondent reported having either condition diagnosed in the past 12 months. We measured substance use, including alcohol, tobacco, and marijuana, in the past month and in the past year. Other relevant covariates included age (12–25, ≥26), year (2005–2014), trimester of pregnancy, race/ethnicity (non-Hispanic Caucasian, African American, other, and Hispanic), health insurance (private, public [including Medicaid/CHIP/CHAMPUS], and none), marital status (married and unmarried), self-reported health status (excellent, very good, good, and fair or poor), and education (age 12–17 or less than high school, high school graduate, and some postsecondary or more).
Analysis
We estimated the proportion of pregnant women reporting nonmedical opioid use during the past year or during the past month. We also described weighted demographic and clinical characteristics of the women in these groups and compared characteristics between those with nonmedical opioid use and those who do not report use, using χ2 tests.
We constructed survey-weighted multivariable logistic regression models to examine the associations between patient characteristics and a) past year nonmedical opioid use and b) past month nonmedical opioid use. We modeled past year nonmedical use including the following independent variables: year (entered continuously as a restricted cubic spline with three knots), trimester, age, race, insurance, an indicator for anxiety/depression in the past year, separate indicators for past year use of alcohol, tobacco, and marijuana, marital status, self-reported health status, and education. We then fit a model of past month nonmedical opioid use. Owing to the limited frequency of events in the past month model (unweighted n = 122; weighted n = 217,106), we limited the number of covariates in multivariable models to avoid overfitting; covariates included year (using a restricted cubic spline with three knots), trimester, age, race, insurance, an indicator for anxiety/depression in the past year, and indicators for past 30 day use of alcohol, tobacco, and marijuana.
Predicted marginal probabilities of selected covariates based on model estimation are displayed in Figure 1, Figure 2. The bars in the figures indicate the probability of nonmedical opioid use for each level of, for example, past year alcohol use (yes and no) and setting all other covariate values to each individual's own values. We then calculated the absolute difference in predicted marginal probabilities of nonmedical opioid use between alcohol users and nonusers and tested the statistical significance of the difference. We also calculated predicted marginal probabilities for women with multiple risk factors, focusing on those who reported use of alcohol, tobacco, and marijuana. This method puts into clinical perspective the absolute risk differences, based on key correlates of nonmedical opioid use in the past year or in the past month by pregnant women.
Figure 1Predicted marginal probabilities of past year nonmedical opioid use by pregnant women across multiple characteristics–anxiety or depression, alcohol use, tobacco use, marijuana use, and multiple substance use (alcohol, tobacco, and marijuana) in the past year. ***p < .001, **p < .01.
Figure 2Predicted marginal probabilities of past month nonmedical opioid use by pregnant women across multiple characteristics–anxiety or depression, alcohol use, tobacco use, marijuana use, and multiple substance use (alcohol, tobacco, and marijuana) in the past month. ***p < .001, **p < .01.
Analyses were conducted using R version 3.1.3, and two-sided p values of less than .05 were considered significant. Missing values accounted for less than 2% of our sample across all variables. This study was designated exempt from review by the University of Minnesota Institutional Review Board.
Results
More than 5% of pregnant women (5.1%; 95% CI, 4.6–6.0) reported nonmedical use of prescription opioids over the past year; nearly 1% (0.9%; 95% CI, 0.7–1.0) reported nonmedical opioid use during the past month (Table 1).
Table 1Descriptive Characteristics of Pregnant Women by Nonmedical Opioid Use in the Past Year and Past Month, 2005–2014
Nonmedical Opioid Use, Past Year (%) (n = 1,222,060)
No Nonmedical Opioid Use, Past Year (%) (n = 22,632,981)
p-value
Nonmedical Opioid Use, Past Month (%) (n = 217,106)
No Nonmedical Opioid Use, Past Month (%) (n = 23,637,934)
p Value
5.1
94.9
0.9
99.1
Age (y)
<.001
<.001
≤25
63.3
38.8
67.8
39.8
≥26
36.7
61.2
32.2
60.2
Race/ethnicity
.007
.149
Non-Hispanic Caucasian
66.8
57.3
51.9
57.8
Non-Hispanic African American
14.0
14.5
24.6
14.4
Non-Hispanic other
4.2
8.3
5.5
8.1
Hispanic
15.0
19.9
18.0
19.7
Marital status
<.001
<.001
Married
29.6
61.4
23.3
60.1
Unmarried
70.4
38.3
76.7
39.9
Household income ($)
<.001
.004
<20,000
35.6
23.2
44.3
23.7
20,000-49,999
33.5
31.6
32.2
31.7
50,000-74,999
13.8
17.2
10.1
17.1
≥75,000
17.1
28.0
13.4
27.6
Education
<.001
<.001
Age 12–17 or less than high school
30.4
18.7
37.9
19.1
High school graduate
32.7
25.6
27.9
25.9
Some post-secondary
36.9
55.8
34.2
55.0
Insurance
<.001
<.001
Private
31.8
54.2
25.2
53.3
Medicaid/CHIP/CHAMPUS
48.5
32.9
46.9
33.5
None
19.7
12.9
28.0
13.2
Trimester
<.001
<.001
First
45.2
30.7
49.1
31.3
Second
29.3
35.9
31.7
35.6
Third
25.5
33.4
19.2
33.1
Self-reported health status
<.001
<.001
Excellent
19.5
35.3
15.1
34.7
Very good
36.5
37.7
19.2
37.8
Good
30.4
22.0
43.5
22.2
Fair or poor
13.6
5.0
22.2
5.3
Anxiety or depression
Past year
25.9
8.1
<.001
29.8
8.8
<.001
Alcohol use
Ever used in past month
23.9
8.1
<.001
49.2
8.6
<.001
Ever used in past year
80.1
59.2
<.001
79.9
60.1
.002
Tobacco use
Ever used tobacco in past month
43.5
14.5
<.001
59.3
15.6
<.001
Ever used tobacco in past year
65.2
27.1
<.001
65.8
28.7
<.001
Marijuana use
Past month
22.9
2.6
<.001
41.6
3.3
<.001
Past year
46.3
9.8
<.001
61.8
11.2
<.001
Note: All descriptive characteristics incorporate NSDUH survey sampling design and weights, and Pearson χ2p values are reported.
Compared with pregnant women who did not report past year nonmedical opioid use, those who did have opioid nonmedical use in the past year had a higher prevalence of anxiety or depression diagnoses (25.9% vs. 8.1%), and more frequently reported alcohol, tobacco, and marijuana use in the past year (80.1% vs. 59.2%, 65.2% vs. 27.1%, and 46.3% vs. 9.8%, respectively). Women reporting nonmedical opioid use over the past year also tended to be younger (63.3% age 12–25 vs. 38.8%), unmarried (70.4% vs. 38.3%), lower income (35.6% <$20,000 annual income vs. 23.2%), not completed high school (30.4% vs. 18.7%), have fair or poor health (13.6% vs. 5.0%), and to have government or no health insurance (68.2% vs. 45.8%), compared with nonusers (Table 1; p < .05 for all comparisons shown). Characteristics of pregnant women reporting past month nonmedical use of prescription opioids revealed similar patterns, compared with nonusers. Pregnant women with nonmedical opioid use in the past 30 days reported comparatively higher rates of anxiety or depression and past year as well as higher rates of substance use in the past month, relative to those who did not report nonmedical opioid use in the past 30 days (Table 1).
In adjusted models, the following characteristics were associated with pregnant women's reports of nonmedical opioid use within previous year: depression or anxiety diagnosis within the previous year (AOR, 2.15; 95% CI, 1.52–3.04), and previous year use of alcohol (AOR, 1.56; 95% CI, 1.11–2.17), tobacco (AOR, 1.72; 95% CI, 1.17–2.53), and marijuana (AOR, 3.44; 95% CI, 2.47–4.81). Past year opioid nonmedical use was associated with younger age (12–25 vs. >35: AOR, 3.37; 95% CI, 1.20–9.46), being unmarried versus married (AOR, 1.38; 95% CI, 1.00–1.90), and having Medicaid (AOR, 1.40; 95% CI, 1.02–1.94) or no health insurance (AOR, 1.57; 95% CI, 1.08–2.29), compared with private health insurance. Pregnant Caucasian women had higher odds of reporting past year nonmedical opioid use compared with African American women (AOR, 1.61; 95% CI, 1.09–2.36), and women reporting fair or poor health had more than twice the odds of past year nonmedical opioid use relative to those reporting excellent health (AOR, 2.40; 95% CI, 1.20–4.76; Table 2).
Table 2Characteristics Associated With Past Year Nonmedical Opioid Use Among U.S. Pregnant Women, 2005–2014
Characteristic
AOR (95% CI)
p-value
Age (y)
12-25
1.38 (1.00–1.90)
.05
≥26
Reference
Trimester
First
1.34 (0.93–1.92)
.12
Second
0.92 (0.69–1.24)
.60
Third
Reference
Anxiety or depression in past year
Yes
2.15 (1.52–3.04)
<.001
No
Reference
Alcohol in past year
Yes
1.56 (1.11–2.17)
.01
No
Reference
Tobacco in past year
Yes
1.72 (1.17–2.53)
.01
No
Reference
Marijuana in past year
Yes
3.44 (2.47–4.81)
<.001
No
Reference
Race/ethnicity
Non-Hispanic Caucasian
1.61 (1.09–2.36)
.02
Non-Hispanic other
0.96 (0.48–1.90)
.90
Hispanic
1.16 (0.72–1.87)
.56
Non-Hispanic African American
Reference
Insurance
Medicaid/CHIP/CHAMPUS
1.40 (1.02–1.94)
.04
None
1.57 (1.08–2.29)
.02
Private
Reference
Marital status
Unmarried
1.60 (1.10–2.32)
.02
Married
Reference
Self-reported health status
Very good
1.28 (0.85–1.92)
.25
Good
1.38 (0.89–2.14)
.16
Fair or poor
2.40 (1.20–4.76)
.02
Excellent
Reference
Education
12–17 or some high school
1.18 (0.85–1.65)
.32
High school graduate
1.01 (0.69–1.48)
.98
Some postsecondary or more
Reference
Note: All model estimates incorporate National Survey of Drug Use and Health survey sampling design and weights and include a linear spline with three knots to control for year.
Figure 1 displays predicted marginal probabilities of past year nonmedical opioid use among pregnant women by anxiety or depression, alcohol use, tobacco use, and marijuana use in the past year. Differences in predicted marginal probabilities were statistically significant for all characteristics evaluated. For a pregnant woman with a past-year diagnosis of anxiety or depression, the predicted probability of past year nonmedical opioid use was 8.6%, compared with 4.5% for a similar woman without anxiety or depression. Pregnant women who used alcohol, tobacco, or marijuana (vs. nonusers) also had higher predicted marginal probability of nonmedical opioid use in the past year (5.6% vs. 3.8%, 6.3% vs. 3.9%, and 10.6% vs. 3.6%, respectively). The predicted probability of past year nonmedical opioid use was markedly higher for a pregnant woman who reported using three substances (alcohol, tobacco, and marijuana) in the past year (14.9%) compared with a predicted probability of 4.3% for a similar woman who was a nonuser.
Shifting focus from the past year to the past month, Table 3 presents AORs of characteristics associated with pregnant women's nonmedical opioid use within the previous 30 days, adjusted for covariates. Pregnant women with anxiety or depression in the past year had higher odds of past month nonmedical opioid use compared with women without these conditions (AOR, 1.78; 95% CI, 1.02–3.09). Those who reported using alcohol, tobacco, or marijuana in the past month had two to five times the odds of nonmedical opioid use during that same time period, compared with nonusers (AOR, 5.00 [95% CI, 2.66–9.40]; AOR, 2.86 [95% CI, 1.49–5.51]; AOR, 4.71 [95% CI, 2.42–9.16], respectively).
Table 3Characteristics Associated With Past Month Nonmedical Opioid Use Among U.S. Pregnant Women, 2005–2014
Note: All model estimates incorporate National Survey of Drug Use and Health survey sampling design and weights and include a linear spline with three knots to control for year.
Figure 2 displays predicted marginal probabilities of past month nonmedical opioid use among pregnant women by the same characteristics of interest, namely, anxiety or depression in the past year, and alcohol, tobacco, and marijuana use in the past month, as well as a measure for women who reported past-month use of multiple substances (alcohol, tobacco, and marijuana). Across all categories, the difference in predicted probability of past month nonmedical use of prescription opioids was significantly higher for a woman with a depression/anxiety diagnosis or who reported past month substance use. The most striking differences were predicted for a pregnant woman who reported past month use of alcohol, tobacco, and marijuana, for whom the predicted probability of nonmedical opioid use was 20.2%, compared with 1.4% for a similar woman who was a nonuser.
Discussion
More than 6 million U.S. women become pregnant annually (
), and this analysis shows that approximately 300,000 pregnant women (5.1%) reported nonmedical use of prescription opioids during the prior year. More than 44,000 pregnant women (0.9%) reported nonmedical opioid use in the past month, where—for many—their use may coincide with and affect the pregnancy. Notably, pregnant women reporting use of multiple substances (alcohol, tobacco, and marijuana) have a substantially higher predicted probability of nonmedical opioid use.
Physiologic dependence on opioids at birth poses considerable health risks to infants, including the development of NAS (
Prevalence of hospitalized live births affected by alcohol and drugs and parturient women diagnosed with substance abuse at liveborn delivery: United States, 1999-2008.
Maternal and Child Health Journal.2013; 17: 667-676
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
), but this requires detection and access to appropriate services. In this study, the characteristics associated with nonmedical opioid use by pregnant women reveal populations in contact with the medical system, including women with depression/anxiety diagnoses and concurrent substance use. These populations may also have broader social needs related to their mental illness and substance use. Emerging clinical and policy efforts currently underway to address the opioid epidemic nationally may offer opportunities for targeted efforts to reach pregnant women at risk of nonmedical opioid use.
Implications for Practice and/or Policy
Clinical Implications for the Detection and Treatment of Maternal Opioid Misuse
Screening for substance use is an integral part of preconception and maternity care (
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
American College of Obstetricians and Gynecologists Committee on Health Care for Underserved Women AGOG Committee Opinion No. 473: Substance abuse reporting and pregnancy: The role of the obstetrician-gynecologist.
). Despite clear recommendations for routine screening for nonmedical opioid use (including opioid dependence), screening patterns are variable, and—where screening is reported—validated instruments are not consistently used, owing in part to limited options for referral and treatment access (
). For women, early recognition of nonmedical opioid use (and detection and treatment of opioid dependence) may allow for tapering before planned pregnancies, and—when detected during pregnancy—appropriate treatment for opioid dependence can effectively reduce its pregnancy-related health risks (
American College of Obstetricians and Gynecologists Committee on Health Care for Underserved Women AGOG Committee Opinion No. 473: Substance abuse reporting and pregnancy: The role of the obstetrician-gynecologist.
Substance Abuse and Mental Health Services Administration (SAMHSA) Substance abuse treatment: Addressing the specific needs of women. Treatment Improvement Protocol (TIP) Series 51.
). Additionally, although prescribing guidelines for opioid pain medications have received attention and updates in recent years, little specific attention is paid to reproductive-age women generally and to pregnant women specifically (
). Clinicians who care for women before and during pregnancy need clear, actionable guidance regarding appropriate pain management strategies that minimize the potential risks of opioid dependency.
Appropriate treatment for opioid dependence, a clinically diagnosed condition that may be related to nonmedical opioid use, when detected during pregnancy, requires carefully managed ongoing use of opioids through medication-assisted treatment (
). Patient needs are substantial, as is the level of effort and commitment required on the part of both the patient and the health care delivery system to effectively meet these needs. In reality, the ideal laid out in guidelines is rarely achieved owing to myriad barriers and challenges such as stigma, lack of resources, limited clinician capacity, and absence of qualified personnel (or great distance from qualified personnel;
Substance Abuse and Mental Health Services Administration (SAMHSA) Substance abuse treatment: Addressing the specific needs of women. Treatment Improvement Protocol (TIP) Series 51.
). Our analysis revealed that some of the nation's most vulnerable pregnant women—those with mental illness and/or alcohol, tobacco, or marijuana use—are also most likely to report nonmedical opioid use, highlighting the potential maternal and neonatal clinical effects of underdetection and undertreatment. Some of the women with nonmedical opioid use that co-occurs with substance use or mental illness may also seek and receive treatment for their co-occurring conditions. For example, among the adults in our study sample, 25% of pregnant women with past month nonmedical opioid use and 38% of those with past year use also reported receiving mental health treatment in the past year, implying an important potential role for mental health care providers in screening for nonmedical opioid use.
Policy Implications
In 2013, the Department of Health and Human Services estimated that the Affordable Care Act combined with the Mental Health Parity and Addiction Equity Act would result in an expansion of mental health and substance use disorder benefits and parity protections for 62 million Americans (
). However, it is unclear whether these predictions have yet come to fruition, especially for pregnant women whose access to opioid-related and other substance use treatment may be hampered by costs, clinician supply shortages, and the choice of their state not to expand Medicaid (
Many low-income pregnant women are eligible for Medicaid coverage during pregnancy and for 60 days after childbirth. Nationally, Medicaid finances 48% of births (
), and many lose eligibility for Medicaid coverage after 60 days, because income eligibility thresholds plummet from 250% to 300% of the Federal Poverty Level to 133% in most states (
). Even though the Affordable Care Act provides for greater access to private health insurance coverage, the variability in mental health coverage in private markets mirrors those in Medicaid programs (
Also, workforce constraints severely limit the availability of clinicians who are qualified to treat nonmedical opioid use in pregnant women. Professional guidelines recommend co-management of opioid-dependent pregnant women by an obstetrician-gynecologist and an addiction medicine specialist (
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
). State and federal policy can influence clinician supply, as indicated by differential growth in physicians credentialed to provide opioid treatment based on state-level Affordable Care Act implementation (
Recently, the Government Accountability Office released a report highlighting the federal government's approach to nonmedical opioid use in pregnancy and infants with NAS. The report discussed fragmentation in federal programs and concluded that federal efforts need more coordination and planning (
). Since that time, two notable pieces of legislation have been signed into law, aiming to improve access to treatment and outcomes for pregnant women with opioid dependence and infants with NAS—The Protecting Our Infants Act and the Comprehensive Addiction and Recovery Act. Taken together, these pieces of legislation aim to improve coordination federal efforts, improve access to treatment for pregnant and postpartum women, and improve implementation of safe care plans for infants discharge home after substance exposure. Importantly, however, as of this writing, the future of these pieces of legislation and the broader health reforms on which they are premised is uncertain.
Limitations
There are several limitations of this analysis. First, repeat cross-sections of NSDUH data from 2005 to 2014 were pooled to increase the analytic sample size, but a longitudinal assessment was not possible. Also, geographic location variables were not added to NSDUH until 2007 and therefore were not included. Although self-report is considered a reliable measure for pregnancy status (
), it is possible that respondents may have misreported or been unaware of their pregnancy status. Substance use may be under-reported among pregnant individuals, especially owing to social desirability bias (
). Also, although we examined a number of important covariates, the NSDUH does not include important biological (e.g., gestational age) and contextual (e.g., prenatal care) factors that may inform prevention efforts, nor does it include information on duration or repetition of nonmedical opioid use over time. Finally, the reports by currently pregnant survey respondents of past year and past month measures of nonmedical opioid use do not provide a precise categorization of use before and during pregnancy.
Conclusions
This analysis of a national sample of pregnant women revealed that depression or anxiety, alcohol, tobacco, and marijuana use each independently predicted nonmedical opioid use in this population. Policy and prevention efforts to improve both screening and treatment could focus on the at-risk populations identified in this study.
References
ACOG Committee on Health Care for Underserved Women; American Society of Addiction Medicine
ACOG Committee Opinion No. 524: Opioid abuse, dependence, and addiction in pregnancy.
Prevalence of hospitalized live births affected by alcohol and drugs and parturient women diagnosed with substance abuse at liveborn delivery: United States, 1999-2008.
Maternal and Child Health Journal.2013; 17: 667-676
Katy B. Kozhimannil, PhD, MPA, is Associate Professor, Division of Health Policy and Management, University of Minnesota. She conducts research to inform the development, implementation, and evaluation of policy impacting health care delivery, quality, and outcomes during critical times in the lifecourse.
Amy J. Graves, SM, MPH, is a Statistician in the Division of Health Policy and Management, University of Minnesota. She conducts data analyses to inform policy making and clinical care for women, children, and beneficiaries enrolled in Medicare and Medicaid programs.
Robert Levy, MD, MS, is Assistant Professor in the Department of Family Medicine and Community Health at the University of Minnesota Medical School. He is a family physician and addiction medicine specialist who treats pregnant women for substance use disorders.
Stephen W. Patrick, MD, MPH, is Assistant Professor of Pediatrics and Health Policy at Vanderbilt University School of Medicine. He is a practicing neonatologist, health services researcher, and national expert on neonatal abstinence syndrome and the opioid epidemic.
Article info
Publication history
Published online: April 10, 2017
Accepted:
March 1,
2017
Received in revised form:
February 23,
2017
Received:
November 16,
2016
Footnotes
Funding statement: Research reported in this publication was supported by a Faculty Seed Grant from the University of Minnesota Academic Health Center (K. Kozhimannil, A. Graves, R. Levy) and the National Institute on Drug Abuse of the National Institutes of Health under award number K23DA038720 (S. Patrick). The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.
30329-2&id=gr1.jpg){kind=link}
30329-2&id=gr2.jpg){kind=link}