Abstract
Adverse perinatal outcomes are a significant contributor to neonatal and infant deaths. Mental illness, substance use disorders, and interpersonal trauma are often prevalent within obstetrical populations. Previous literature has documented the individual associations between these psychosocial factors and adverse perinatal outcomes. The co-occurrence of these three psychosocial factors might represent a syndemic among pregnant women, although they have not been described as such in the literature. Analysis of the interrelatedness and aggregate effect of these factors may allow for a more effective screening process that may reduce adverse perinatal outcomes. The objective of this article is to examine whether psychosocial factors (mental illness, substance use disorders, and interpersonal trauma) were independently and synergistically associated with adverse perinatal outcomes. This is a retrospective cohort study of 1,656 pregnant women at a single institution. Perinatal outcome and psychosocial data were abstracted from each participant’s electronic medical record. Univariate and bivariate analyses, and multiple logistic regression were performed. Mean age was 27.5 (SD = 6.2) years. The majority was Black (60.6%) and single (58%). Psychosocial factors were reported in 35% of women. The incidence of adverse perinatal outcomes increased with greater number of psychosocial factors: 21.2% if no psychosocial factor, 27.0% if one psychosocial factor, 27.4% if two, and 35.3% if all three (for trend, p = .01). Women who reported all three psychosocial factors had twice the odds of adverse perinatal outcomes (adjusted odds ratio = 2.04, 95% confidence interval = [1.09, 3.81], p = .03) compared with those who reported none. Our data suggest there is a synergistic relationship between the psychosocial factors that is associated with increased adverse perinatal outcomes. A validated screening tool is needed to stratify patient’s risk of adverse perinatal outcomes based on psychosocial factors. Such screening could lead to tailored interventions that could decrease adverse perinatal outcomes.
Introduction
Adverse perinatal outcomes, which include preterm births and low birth weight infants, are a significant contributor to neonatal and infant deaths (Mathews, MacDorman, & Thoma, 2015). Decreases in gestational age at delivery and birth weight are associated with poorer survival (Mathews et al., 2015). Preterm births account for 70% of neonatal deaths, 36% of infant deaths, and 25% to 50% of cases of long-term neurologic impairment in children (Volpe, 1997). The causes of adverse perinatal outcomes are complex and poorly understood. Centers for Disease Control and Prevention (CDC) recommends identifying risks associated with poor maternal, infant, and child health outcomes, and implementing interventions that can improve these outcomes and have the potential to reduce societal burden (U.S. Department of Health and Human Services, 2014). Substantial literature exists linking physical maternal factors to adverse perinatal outcomes (Lawn, Cousens, Zupan, & Lancet Neonatal Survival Steering Team, 2005; Lopez, Mathers, Ezzati, Jamison, & Murray, 2006; Ouyang et al., 2013); however, the role of psychosocial factors is only beginning to develop.
Individually, psychosocial factors such as mental illness, substance use disorders, and interpersonal trauma (i.e., experiences of physical and/or sexual abuse as a child or an adult) have been associated with poor maternal and infant health outcomes (Connelly, Hazen, Baker-Ericzén, Landsverk, & Horwitz, 2013; Holden, McKenzie, Pruitt, Aaron, & Hall, 2012; Mason & O’Rinn, 2014). Mental illness has been shown to occur in 10% to 13% of peripartum women (World Health Organization, 2016). Depression has been shown to be an independent risk factor for adverse perinatal outcomes such as stillbirth and neonatal deaths (Hughes, Turton, & Evans, 1999; King-Hele et al., 2009; Thornton, Guendelman, & Hosang, 2010). Based on a study by the Substance Abuse and Mental Health Services Administration (2014), 5.4% of pregnant women aged 15 to 44 years use illicit substances, which increases the risk of adverse perinatal outcomes (Creanga et al., 2012; David et al., 2014; Hayatbakhsh et al., 2012; Maeda, Bateman, Clancy, Creanga, & Leffert, 2014). In addition, a systematic review from Taillieu and Brownridge (2010) reported the prevalence of intimate partner violence during pregnancy ranged from 0.9% to 30% for physical abuse, 1% to 3.9% for sexual abuse, and 1.5% to 36% for emotional abuse. Pregnant women who report abuse (physical, sexual, or emotional) have higher odds of delivering a low birth weight baby (odds ratio [OR] = 1.4, 95% confidence interval [CI] = [1.1, 1.8]; Murphy, Schei, Myhr, & Du Mont, 2001). Although these psychosocial factors individually affect perinatal outcomes, data are lacking that comprehensively account for contexts when all three factors are present.
Experience of these psychosocial factors early in life has been associated with negative health outcomes later in life. For example, childhood trauma is associated with unhealthy risky behaviors and lifestyle choices in adulthood (i.e., excessive alcohol use, smoking, suicidal attempt), diseases (i.e., obesity, sexually transmitted infection), or sexual revictimization as an adult (Brenner & Ben-Amitay, 2015; Felitti et al., 1998; Norman et al., 2012;Whitfield, Anda, Dube, & Felitti, 2003). Furthermore, interpersonal trauma has been associated with increased risk of mental illness and substance use disorders (Cocozza et al., 2005; Coker et al., 2002). In one study, among the 51% of the women who reported a lifetime history of intimate partner violence, 39% also reported mental illness and 31% reported illicit and licit substance use disorders (Hink, Toschlog, Waibel, & Bard, 2015). One study found that pregnant women who are assaulted are more likely to report a history of substance abuse and mental illness (e.g., anxiety and depression; Campbell, Poland, Waller, & Ager, 1992).
The Perspectives of Psychiatry framework provides a multifaceted approach to the individual diagnosis, treatment, and research of these factors (McHugh & Slavney, 2011), which comprises the biological, cognitive, and behavioral traits that contribute to a person’s mental experience. This framework would suggest that the lifetime occurrence of mental illness, substance use disorders, and interpersonal trauma could lead to a negative effect on the person’s physiology (e.g., physical trauma, hormonal shifts) and social setting (e.g., family, friends, social support) that can be cyclical in nature, perhaps resulting in adverse perinatal outcomes. The co-occurrence of these psychosocial factors therefore might represent a syndemic among pregnant women. A syndemic refers to two or more health problems that coexist or cluster in certain populations with an interaction that may worsen or lead to negative health consequences (Singer, 2009). A syndemic approach offers a tangible framework for analysis and prevention of co-occurring disorders. The value of finding a synergistic connection between certain health conditions or diseases can offer direction for effective management (Singer & Clair, 2003). Although mental illness, substance use disorders, and interpersonal trauma are often reported during prenatal care risk screens, and each are well-recognized as risk factors for adverse perinatal outcomes individually (Gold & Marcus, 2008; McFarlane, Campbell, Sharps, & Watson, 2002), there are no known data on whether there is a synergistic association between these psychosocial factors that might be associated with increased adverse perinatal outcomes. Hence, this study sought to (a) determine the prevalence of psychosocial factors (i.e., mental illness, substance use disorders, and interpersonal trauma), (b) assess whether these psychosocial factors are interrelated and represented a syndemic, and (c) determine if there was a synergistic association between these psychosocial factors and adverse perinatal outcomes in a cohort of urban pregnant women.
Materials and Methods
Study Design and Participants Characteristics
A retrospective cohort study was conducted among pregnant women in Baltimore, Maryland, who delivered at a large tertiary care center in 2012. Prenatal care services were provided by resident physicians, advanced practice nurses, and attending physicians. Women with elective termination of pregnancy, transfer of prenatal care, and multiple gestations were excluded.
Demographic and clinical data were collected from the electronic medical record by trained medical and nursing research personnel which included age, race, mental health history, substance use disorder history, interpersonal trauma history, education attainment, type of insurance (private or public), marital status, parity, and gestational age at first prenatal visit. Household income was not collected; therefore, type of insurance served as a proxy for socioeconomic status. A standardized abstraction form was used to guide data collection. The Johns Hopkins Medicine Institutional Review Board approved this study.
Psychosocial Factors Measurement Tools
Lifetime history of mental illness (depression, anxiety, schizophrenia, bipolar disorder), lifetime illicit or licit substance use disorders (marijuana, crack cocaine/stimulants, opioids, and tobacco), and interpersonal trauma (childhood or adulthood physical, sexual, or emotional abuse) were selected as the psychosocial factors and were self-reported. All patients were interviewed by a nurse at the initial prenatal care visit using internally developed, standardized risk screening tools for mental illness, substance use disorders, and interpersonal trauma. Mental illness was assessed by asking the following: “Were you ever depressed enough to hurt yourself?” “are you depressed enough to hurt yourself now?” and “were you ever diagnosed with a mood disorder like depression, schizophrenia, bipolar disease, or anxiety?” Illicit or licit substance use disorders was assessed by asking the following: “Do you smoke?” “have you ever used street/recreational drugs?” “do you inject drugs?” and “are you currently using street/recreational drugs?” If a response was positive, then the name of the drug, amount, and frequency of use was obtained. Active substance use disorders were also verified by urinary toxicology screen, which was performed at the first prenatal care visit and repeated on the labor floor. A CAGE questionnaire was used to screen for alcohol use disorder (Ewing, 1984). A CAGE test score ≥2, which has a sensitivity of 93% and specificity of 76% for alcohol abuse, was termed “excessive alcohol use.” Interpersonal trauma was assessed by asking the following: “Were you ever physically, verbally, or emotionally abused by a partner or anyone else?” “were you forced to have sex with anyone?” and “are you afraid of your partner or anyone else?” If a woman answered yes to abuse, then she was asked, “did you receive medical treatment for this abuse?”
Adverse Perinatal Outcomes
The primary outcome variable was adverse perinatal outcome, which was a dichotomous composite variable of the presence of at least one of the following variables: preterm delivery (<37 gestational weeks), low birth weight (<2,500 g), neonatal intensive care unit (NICU) admission, or stillbirth. A composite variable was created because of the low frequency of each individual variable.
Statistical Analysis
Each psychosocial factor was dichotomized based on the presence or absence of the risk factor. A syndemic variable was created as a count score based on the number of psychosocial factors present. The score was calculated as a sum of the three psychosocial factors ranging from 0 to 3. Normally distributed variables were summarized by mean (SD) and compared with the student’s t test. Skewed variables were summarized by median (interquartile range [IQR]) and compared using the Wilcoxon rank sum test or Kruskal–Wallis test. Binary and categorical variables were compared using chi-square or Fisher’s exact test. Missing data were assumed to be missing at random and a listwise deletion approach was used in the analyses. Education and insurance variables were the only two variables with missing data and were missing in less than 2% of charts.
We followed published literature on syndemics for the analysis (Chakrapani, Newman, Shunmugam, Logie, & Samuel, 2015; Milstein, 2002; Singer, 2009). First, the prevalence of the psychosocial factors—mental illness, substance use disorders, and trauma—and adverse perinatal outcome were estimated. Second, bivariate OR for each pair of psychosocial factors and their association with adverse perinatal outcome were calculated. Third, to test for synergistic relationships, we calculated the prevalence of adverse perinatal outcome for each possible count score of psychosocial factors (from 0 to 3) and used the chi-square test for linear trend and Goodman and Kruskal’s gamma to measure an association. Last, multiple logistic regression (MLR) was performed to determine the association between the syndemic count variable and adverse perinatal outcome. MLR models were built by entering variables significantly associated (p < .05) with the primary outcome in univariable analyses, and results were reported as unadjusted and adjusted odds ratios (aOR) and 95% CI. A listwise deletion approach was used to handle missing data. For some analyses, data were stratified by race, that is, Black versus other races. Statistical significance was defined by a two-sided p value of less than .05. All the analyses were conducted in Stata 13.1 (StataCorp LP, College Station, Texas).
Results
Demographic Characteristics
There were 1,988 women who delivered during the study period. After excluding 14 women with elective termination of pregnancy, 274 women who transferred care, and 44 women with multiple gestations, 1,656 women remained. Mean (SD) age was 27.5 (6.2) years. The majority of participants were Black (60.6%; n = 1,003), were single (58%; n = 961), completed high school (79.9%; n = 1,307), and had public health insurance (54.7%; n = 906; Table 1). The median (IQR) gestational week at entry into prenatal care was 10 (8-17), and 39.6% (n = 655) of women attended at least 10 prenatal visits.
Univariable Analysis by Psychosocial Factors.
Note. Statistically significant findings are bolded. IQR = interquartile range.
Prevalence of Psychosocial Factors and Adverse Perinatal Outcomes
Among all participants, 16.3% (n = 270) reported lifetime mental illness, 19.5% (n = 323) reported lifetime illicit or licit substance use disorders, and 11.8% (n = 196) reported lifetime interpersonal trauma. There were 35% (n = 579) of women who reported any psychosocial factor. Mental illness diagnoses among the cohort included depression (8.3%; n = 137), anxiety (1.9%; n = 31), bipolar disorder (1.2%; n = 26), schizophrenia (0.06%; n = 1), developmental delay (0.2%; n = 3), and multiple diagnoses (4.3%; n = 72). Type of substance used during pregnancy among the cohort included tobacco smoking (4.1%; n = 68), excessive alcohol use (0.1%; n = 1), and illicit drugs (15.2%; n = 252). Urinary toxicology screen confirmed ongoing illicit drug use in 51.6% (n = 130) of women who reported illicit drug use. The number of participants who self-reported ongoing illicit or licit substance use disorders was greater than those who tested positive by urinary toxicology (252 vs. 214, p < .001); therefore, we selected the self-reported variable for all analyses, which is an approach previously validated in different populations (Beaton, Hogg-Johnson, & Bombardier, 1997; Darke, 1998). Types of interpersonal trauma reported among the cohort included physical abuse (4.3%; n = 71), sexual abuse (3.8%; n = 63), verbal or emotional abuse (0.9%; n = 15), and any combination of abuse (2.7%; n = 44).
Among all participants, there were 9.5% (n = 157) preterm deliveries, 9.7% (n = 161) low birth weight infants, 16% (n = 256) NICU admissions, and 0.6% (n = 10) stillbirths. As a composite, 23.5% (n = 389) of women had an adverse perinatal outcome. Demographic variables positively associated with adverse perinatal outcomes included Black race (p = .04), single marital status (p < .001), lack of a high school diploma (p = .002), and public insurance (p = .02). Other variables are included in Table 2. Individually, mental illness (p = .006) and interpersonal trauma (p = .02) were associated with adverse perinatal outcomes, while substance use disorders was not (p = .1).
Univariable Analysis by Adverse Perinatal Outcomes.
Note. Statistically significant findings are bolded. IQR = interquartile range; STI = sexually transmitted infection.
Bivariate Association of Psychosocial Factors and Adverse Perinatal Outcomes
Among women who reported lifetime mental illness, the odds of lifetime interpersonal trauma were 5 times as large (aOR = 5.12, 95% CI = [3.65, 7.19]) and the odds of lifetime substance use disorders were 3 times as large (aOR = 3.27, 95% CI = [2.39, 4.48]) as the odds for women who did not report mental illness after adjusting for parity, race, marital status, education attainment, gestational age at initial prenatal care visit, and insurance status (Table 3). Among women who reported lifetime substance use disorders, the odds of lifetime interpersonal trauma were 3 times as large (aOR = 3.15, 95% CI = [2.26, 4.41]) as the odds for women who did not report substance use disorders. The magnitude of these associations was greater for women of other racial groups than Black. Two of the psychosocial factors were significantly associated with larger odds of an adverse perinatal outcome, mental illness (aOR = 1.53, 95% CI = [1.13, 2.08]) and interpersonal trauma (aOR = 1.48, 95% CI = [1.06, 2.06]), whereas substance use disorders was not (aOR = 1.10, 95% CI = [0.81, 1.49]).
Psychosocial Risk Factor Bivariate Analysis and Multiple Logistic Regression.
Note. OR = odds ratio; CI = confidence interval.
Syndemic and Adverse Perinatal Outcome
We assessed the additive effect of the three psychosocial factors on the odds of having an adverse perinatal outcome. Based on the syndemic count variable, there were 65% (n = 1,077) of women who did not report any psychosocial factors, 24.4% (n = 404) who reported one psychosocial factor, 7.5% (n = 124) who reported two psychosocial factors, and 3.1% (n = 51) who reported all three psychosocial factors. Among women who did not report any psychosocial factors, 21.2% (n = 228) had an adverse perinatal outcome, and 27.0% (n = 109) with one factor, 27.4% (n = 34) with two factors, and 35.3% (n = 18) with three factors had adverse perinatal outcomes. Having a greater number of psychosocial factors was significantly and positively associated with adverse perinatal outcomes (for trend, p = .01, Gamma = .16).
After adjusting for race, marital status, education attainment, insurance status, gestational age at initial prenatal care visit, Table 4 shows that among women who reported any psychosocial factor, the odds of an adverse perinatal outcome were larger than those who did not report any psychosocial factors (aOR = 1.34, 95% CI = [1.04, 1.73], p = .03), and among women who reported all three psychosocial factors, the odds of an adverse perinatal outcome were twice as large (aOR = 2.04, 95% CI = [1.09, 3.81], p = .03) as those who did not report any psychosocial factors. There was a trend toward larger OR for an associated adverse perinatal outcome as the number of psychosocial factors increased, although such an increase was not statistically significant in MLR. When stratified by race, among Black women who reported all three psychosocial factors, the odds of an adverse perinatal outcome were over 2.5 times as large (aOR = 2.64, 95% CI = [1.22, 5.69], p = .01) as the odds for other racial groups.
Syndemic Association of Psychosocial Factors With Adverse Perinatal Outcomes.
Note. OR = odds ratio; CI = confidence interval.
Discussion
We present the first evidence that suggests there is a synergistic relationship between the lifetime occurrence of three psychosocial factors—mental illness, substance use disorders, interpersonal trauma—and adverse perinatal outcomes. Over a third of our cohort of diverse pregnant women self-reported psychosocial factors, and we identified a linear dose-response trend between the identified psychosocial factors and adverse perinatal outcomes.
Similar to our findings, Connelly and colleagues (2013) studied the co-occurrence of depression, intimate partner violence, and substance use and reported that over one third of their obstetrical cohort had one or more psychosocial factor. This and our study also found a small percentage of pregnant women who reported all three factors. Although this and other studies have shown the prevalence of these psychosocial factors in pregnancy and their association with birth outcomes (Connelly et al., 2013; Holden et al., 2012), our study expanded this field of investigation to include a synergistic association between these psychosocial factors and adverse perinatal outcomes.
Each of the three psychosocial factors in our study was strongly associated with each other, even after adjusting for confounders. Among pregnant women who reported a mental illness or trauma, the odds of a substance use disorder were 3 to 5 times as large as those without a mental illness or trauma. This interrelatedness between psychosocial factors is similar to a study by Holden and associates (2012), which showed psychosocial relationships between depression and intimate partner violence, depression and substance use disorders, and substance use disorders and intimate partner violence. In another study, 50% of interpersonal trauma victims described symptoms of mental illness and had a higher rate of binge drinking than those not exposed to trauma (Zahnd, Aydin, Grant, & Holtby, 2011). Thirty-three percent of these traumatized adults requested assistance with mental illness and substance use disorders. The strong interrelatedness of these psychosocial factors among pregnant women warrants attention by prenatal care providers.
Central to the study aim, the presence of mental illness, substance use disorders, or interpersonal trauma, both individually and collectively, was associated with adverse perinatal outcomes, even after adjusting for demographic variables and gestational age at the initial prenatal care visit. Our findings suggest that women who experience these psychosocial factors have larger odds of having an adverse perinatal outcome. Furthermore, when all three psychosocial factors are reported, the odds of an adverse perinatal outcome may be twice as large as the odds for women who do not have any psychosocial factors, which is cause for concern. This association was even larger for Black women.
Limitations to our study included the lack of data on medical comorbidities that might have been associated with adverse perinatal outcomes. Next, our sample was predominantly Black women, who tend to report increased abuse than other races, and limits the generalizability of our findings (McFarlane et al., 2002). Given the known variations in the regularity of self-reported traumatic episodes (Krinsley, Gallagher, Weathers, Kutter, & Kaloupek, 2003), we may have underestimated the true proportion of women who experience trauma in their lifetime (misclassification bias). Furthermore, we reported on lifetime experiences of the psychosocial factors and we do not know the association between a resolved or treated psychosocial factor and adverse perinatal outcomes. However, some psychosocial factors like past or childhood trauma may have lasting effects that may negatively affect the other psychosocial factors directly or indirectly (Kendler et al., 2000). Next, this study was underpowered to formally test statistical interactions among the psychosocial factors, although there was a trend of increasing magnitude of association as the number of psychosocial factors increased. Last, our definition of a syndemic and its methodology is under debate. Some argue for sophisticated modeling to address interaction effects versus an additive effect (Tsai & Burns, 2015); whereas, others agree with methodology similar to ours (Stall, Coulter, Freidman, & Plankey, 2015).
Despite the limitations, the statistical significance of the presented data suggests that mental illness, substance use disorders, and interpersonal trauma are interrelated and are likely to have a synergistic association with adverse perinatal outcomes. Psychosocial factors are complex, and each one provides its own challenges. Professional training and education to identify at-risk pregnant women and the opportunity to stratify appropriate and timely interventions are required. It would be essential to consult psychiatric and substance abuse practitioners to initiate innovative strategies to address the psychosocial risk factors (Danhausen et al., 2015). Mental illness and substance use disorder severity could be assessed using Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013) criteria. In addition, encouraging trauma-informed prenatal care, where the clinical environment is safe, supportive, and welcoming, in select at-risk women, can potentially have a major impact on mothers, newborns, families, and society as a whole (Machtinger, Cuca, Khanna, Rose, & Kimberg, 2015).
Implications to Public Health
Federally sponsored programs such as Women, Co-Occurring Disorders, and Violence Study and “trauma-informed skills” training program could educate prenatal care providers to help them understand how psychosocial factors may impact their patients’ lives, focus on interventions at a macrolevel, and implement care with effective communication (McHugo et al., 2005; Substance Abuse and Mental Health Services Administration, 2014; Torchalla, Linden, Strehlau, Neilson, & Krausz, 2015). Providing comprehensive and structured counseling services to address psychosocial factors, even in minute amounts, can improve overall well-being, and appropriate trauma-related literature can reinforce counseling efforts (Machtinger et al., 2015; Torchalla et al., 2015).
The co-occurrence of the identified psychosocial risk factors and their association with adverse perinatal outcomes can lead to an economic burden. The estimated cost of preterm birth in the United States is US$26 billion (Behrman & Butler, 2007) and intimate partner violence exceeds US$5.8 billion with considerable expenditures to address medical and mental health issues, along with lost wages (CDC, 2003). This economic burden is worsened with accompanying risk factors such as mental illness and substance use disorders (CDC, 2003). Thus, it is worth any effort toward preventive intervention to avoid the psychological and economic burden placed on individuals and society due to adverse perinatal outcomes (van Baaren et al., 2015).
Overall, this study confirms the high prevalence of mental illness, substance abuse, and interpersonal trauma within an urban obstetrical population. Together, these psychosocial factors may contribute to a substantial proportion of adverse perinatal outcomes, which could be partly preventable. Future research should first seek to identify the most appropriate validated assessment tools for psychosocial factors reported in this study, which can be used to confirm our findings. Such research could explore the validation of a risk stratification tool for adverse perinatal outcomes based on psychosocial factors, which could then prompt an appropriate referral for proper treatment (e.g., psychiatric consultation). Risk stratification tools have been utilized for other medical morbidities (e.g., stroke, myocardial infarction) to prevent poor health outcomes, and such a tool would be invaluable to prevent adverse perinatal outcomes. Given the prevalence of psychosocial factors within the presented population, research could also aim to identify patients’ protective factors and perceived barriers to obtaining treatment. This information might identify needed changes on a systems level. Such systems-level research could explore potential interventions such as substance use disorder and psychiatric consultation services within OB/GYN clinics. These embedded services might increase the likelihood of effective engagement of patients in longitudinal treatment, which might conversely lead to a demonstrable decrease in adverse perinatal outcomes. Greater attention to the combination of psychosocial factors described in this study is warranted before further gains can be made in reducing adverse perinatal outcomes, especially among urban populations.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.