Mental health, demographic, and risk behavior profiles of pregnant survivors of childhood and adult abuse

Abstract

Our objective was to address the gap in knowledge about the extent to which perinatal mental health and risk behaviors are associated with childhood and adult experiences of abuse that arises because of barriers to screening and disclosure about past and current abuse. Survey data from an ongoing study of the effects of posttraumatic stress on childbearing were used to describe four groups of nulliparous women: those with no abuse history, adult abuse only, childhood abuse only, and abuse that occurred during both periods. The rates of abuse history disclosure were higher in the research context than in the clinical settings. Mental health morbidity and risk behaviors occurred in a dose-response pattern with cumulative abuse exposure. Rates of current posttraumatic stress disorder ranged from 4.1% among those never abused to 11.4% (adult only), 16.0% (childhood only), and 39.2% (both periods). Women abused during both periods also were more likely to be using tobacco (21.5%) and drugs (16.5%) during pregnancy. We conclude that mental health and behavioral risk sequelae affect a significant portion of both childhood and adult abuse survivors in prenatal care. The integration into the maternity setting of existing evidence-based interventions for the mental health and behavioral sequelae of abuse is needed.

INTRODUCTION

Screening for domestic or intimate partner violence and abuse has been a standard of care in maternity health care settings for the last ten years.^1–6 Additional detailed screening related to childhood sexual abuse and adult sexual assault as elements of social history has also been encouraged because of the long-term negative health consequences of childhood maltreatment and sexual trauma.^7–10 Specific sequelae of having experienced childhood or adult abuse include posttraumatic stress disorder (PTSD), ^11,12 depression,¹³ and substance abuse,¹⁴ along with physical comorbidities including both chronic pain conditions and diseases. ¹⁵

There are barriers to health care providers screening for childhood maltreatment and adult abuse.^16–27 Additionally, women who have had these types of abuse in their lives also experience barriers to disclosure when asked in health care settings.^28–34 Rates of disclosure by clients are far below population estimates based on epidemiologic survey research.^{11–12, 35–36} Thus, the actual proportion of women who have experienced childhood and adult abuse among maternity care clients, their posttraumatic stress-related mental health profiles and the extent of their trauma-related health risk(s) remain unknown to maternity care providers.

The primary aim of this article is to describe the mental health status, demographic and behavioral risk profiles of 1,259 nulliparous women survivors of childhood and adult abuse who participated in a research interview prior to 28 weeks of gestation. A second aim is to explore responses of a postpartum subset of women about their experiences of being screened for their abuse history and disclosing it in prenatal clinic and labor and delivery settings. Their maternity care and labor medical records were reviewed to confirm if they contained any notation of past or current abuse. We will situate their qualitative statements in the context of selected literature on screening for abuse. Data for these analyses are from an on-going psychobiological study of the effects of posttraumatic stress disorder on perinatal outcomes; the results reported here do not contain pregnancy outcomes as that portion of the study is not yet complete.

BACKGROUND

Reviews of the prevalence and adverse effects of gender-based violence on women’s health and pregnancy outcomes found in the nursing, obstetric, midwifery, and forensic literature provide compelling reasons to include screening assessments for violence against women in health care settings.^{1–6, 37} The long-term nature of the maternity care relationship and the potential to positively affect the well-being of both a woman and her child make prenatal care a unique and valuable opportunity to address both on-going abuse and long-term adverse consequences of past abuse.

Current research shows that there are still many barriers to screening for past and current physical and sexual abuse and to disclosure about abuse. Recent studies indicate that the reasons health care providers give for not screening is changing from lack of education and confidence in their ability to screen to concerns that there is lack of evidence that screening is effective.^{18, 38}. This concern is complicated by a fear that once screening has identified someone with a positive history there is a paucity of interventions available to aid survivors.¹⁸ Despite these fears, newer studies continue to find that patients consider screening to be valuable if done sensitively^28–29 and there is also recent literature that identifies best practices for screening and interventions (See Appendix A for a self-assessment of knowledge and best practices.)

The overall value of screening for health conditions or risk factors related to a specific health condition is traditionally measured against the prevalence and severity of the condition in the population that is to be screened. As noted above, rates of history of physical or sexual abuse found in maternity care are believed to be well below that reported in epidemiological survey research due to barriers to both screening and disclosure. Therefore the significance and value of screening and intervening may be underestimated because it is based on perceptions of prevalence that is erroneously low. Yet if barriers to screening and disclosure were reduced, the importance of having these dialogues during maternity care could increase. Components of prenatal care might shift to address more psychological and health behavior issues with resources and evidence-based interventions.³⁹ Integrating such interventions into maternity settings has the potential to positively affect health care outcomes for childbearing women who have a history of childhood or adult abuse.

The three landmark studies that measured posttraumatic stress disorder (PTSD) in large samples, including PTSD from childhood and adult abuse, provide compelling evidence of the importance of addressing abuse as a means to improve the nation’s health. The National Women’s Study¹¹, conducted with a nationally representative sample of 4,008 U.S. women, found that 12.3% of women met diagnostic criteria for PTSD in their lifetime and 4.6% currently have the disorder. The original National Comorbidity Survey (NCS) ¹² conducted with a nationally representative sample of 5,877 people, advanced knowledge by showing that 80% of persons with PTSD have comorbid disorders, including depression or substance dependence and that PTSD becomes chronic in 40% of cases. In the NCS the lifetime rate of PTSD among women was 10.4%. In the more recent NCS Replication (NCS-R) ⁴⁰ 5,692 people were assessed for PTSD. The lifetime PTSD rate for women was 9.7% and the current rate was 5.2%. Data from the Detroit Area Study, ^35–36 which included 1,007young adult HMO members, has been used to refine this knowledge, showing that 16–20 years of age is the peak incidence of trauma exposure for women, that African American urban populations have twice the rate of PTSD, and that depression and substance use are most often secondary disorders with onset if PTSD becomes chronic. ¹⁴ These findings have informed mental health practice and research over the past decade related to PTSD and women’s health. But their impact on delivery of health care to women seems to have been less pervasive. The gap between the proportion affected by abuse trauma and the proportion disclosing their experience may render the importance of past and current as a cause of morbidity across the woman’s lifespan and during childbearing less visible.

METHODS

This descriptive analysis was conducted using survey data from an on-going study of the effects of posttraumatic stress disorder on childbearing outcomes (NIH R01 NR008767, Psychobiology of PTSD & Adverse Outcomes of Childbearing; common name STACY Project, an acronym for Stress, Trauma, Anxiety, and the Childbearing Year). This study is a prospective, longitudinal cohort study that combines multiple sources of data. Data in this analysis are from early pregnancy and postpartum telephone interviews and analysis of one element of the maternal medical record. The early prenatal interview is conducted shortly after women initiate prenatal care; it collects data about trauma history including abuse, psychiatric diagnoses, substance use, demographics, and mental health treatment history. The postpartum interview is conducted at six weeks postpartum; this interview includes items about screening and disclosure. Medical records are reviewed after delivery; this review includes noting if past or current abuse has been recorded by a clinician or disclosed on a form.

Women were recruited from prenatal clinics at three medical centers in the region, one in a university town and two in a major urban center. All three institutions provided Institutional Review Board approval for the study, and data collection was conducted in accordance with recommendations for safety in studies of violence against women.^41–42 Prenatal clients were eligible to participate if they were 18 or older, able to speak English without an interpreter, expecting their first child, and initiating prenatal care at less than 28 weeks of gestation. The nurses who conduct the obstetric patient history identified those who were eligible and invited them to take part in “a telephone survey about stressful events that happen to women, emotions, and pregnancy.” If they agreed to be contacted, a telephone interview was later conducted at the time and location chosen by the woman. During the phone interview women were also informed about ongoing ways to continue participation in the same study.

In the first telephone survey measures pertinent to our first aim were completed. These included widely used instruments for history of trauma and psychiatric diagnoses. A history of trauma history was ascertained using the Life Stressor Checklist.⁴³ This instrument assesses 29 potential trauma exposures specific to women’s experiences, including six items using behaviorally-specific wording to inquire about childhood and adult physical abuse and sexual abuse involving contact only versus penetration. Of the five instruments most frequently used in research to measure trauma exposures, this is the one with highest sensitivity to trauma among women.⁴⁴ PTSD diagnosis was determined using the National Women’s Study PTSD module. This instrument was used in the largest epidemiological study of PTSD specific to women that was conducted via the National Crime Victim Center.¹¹ It is designed as a structured telephone diagnostic interview to be administered by lay interviewers. It was validated in a clinical sample of 528 women during the DSM-IV PTSD Field Trial.⁴⁵ The NWS-PTSD module attained a sensitivity of .99 and specificity of 0.79 compared with the Structured Clinical Interview for DSM-III-R.⁴⁶

The diagnosis of comorbidities including major depressive and generalized anxiety disorder was determined using the Composite International Diagnostic Interview (CIDI) modules.⁴⁷ The CIDI is a structured psychiatric diagnostic interview designed to be used by trained interviewers who are not clinicians. Information about current substance use and demographics was obtained using items from the Centers for Disease Control and Prevention’s (CDC) Perinatal Risk Assessment Monitoring System (PRAMS).^48–49 This epidemiologic surveillance instrument was created by the CDC to collect perinatal data routinely across the United States.

Measures pertinent to our second aim were investigator-generated items in the postpartum survey asking about the participants’ experiences with screening and disclosure in prenatal clinic and labor and delivery settings. This included an open-ended question asking their reasons for disclosing or choosing not to. Their responses were typed verbatim by the interviewers. We noted any recorded history of past or current abuse within the perinatal medical record. This information was the only medical record item included in the analysis.

The analyses compare the demographic, psychiatric, and risk behavior profiles of abused women, and describes the extent to which they report being screened for abuse, and the extent to which they disclose abuse. Statistical analysis was conducted using SPSS v. 14.⁵⁰ We present rates for four mutually-exclusive groups: (1) those who were never physically or sexually abused (though they may have other traumatic events in their history), (2) those abused only when they were age 16 or older, (3) those abused only prior to the age of 16, and (4) those who had both childhood and adult abuse representing re-victimization. Content analysis was completed of the verbatim responses to open-ended survey items asking about disclosure. Two analysts reached consensus on the categories that summarized below.^51–52

RESULTS

Approximately 10% of prenatal patients in all three settings met the eligibility criteria because they were adults, able to speak English, less than 28 weeks gestation, and expecting their first infant. As of September 2007, a total of 2,140 had agreed to be contacted about the study. Of these, 1,259 were contacted, confirmed to be eligible, and completed the initial interview. It is their data which are used to describe demographic, abuse history, psychiatric, risk behavior, and mental health treatment profiles. Of these 1,259 women, 377 (30) reported a history of physical or sexual abuse. Of these, 202 (16.0%) experienced abuse as adults, and 254 (20.2%) experienced abuse in childhood. Seventy-nine women (6.3%) reported abuse occurring in both periods of their life. A total of 123 (9.8%) reported abuse occurring at age 16 or older only. Another 175 (13.9%) reported abuse occurring in childhood only, at age 15 or younger. Odds of being abused again as an adult if the woman was abused in childhood were three times greater than for women who did not experience childhood abuse: 79 (31.1%) versus 123 (12.2%); OR = 3.2, 95% CI [2.3, 4.5]. Rates of being abused again as an adult were higher among women living in the inner city: 51 (8.7%) versus 28 (4.2%); OR = 2.2, 95% CI [1.4, 3.5].

Demographic characteristics considered in this analysis (Table 1) were those associated with increased risk for adverse perinatal or psychiatric outcomes: being adolescent (age 18–20 in this study), being African American, having household income less than $15,000, having high school education or less, and living in the inner city. All five of those risk characteristics were present in a dose-response pattern, with lowest rates among non-abused women. Higher rates of all five risk factors occurred in those reporting abuse during one period, with very similar rates in those reporting only adult or only childhood abuse. The highest rates of these risk factors were reported among women who were abused in both periods.

Table 1.

Rates of Specific Risk Factors and Mean Number of Risk Factors for Each Group of Women

Risk factors	No abuse (n=882) n(%)	Adult only (n =123) n (%)	Child only (n =175) n(%)	Both periods (n =79) n(%)	Overall (n=1,259) n(%)	pa
Individual factors
1. Teen (age 18–20)	194 (22.0)	28 (22.8)	47 (26.9)	28 (35.4)	297 (23.6)	.002
2. African American	373 (42.3)	63 (51.2)	86 (49.1)	44 (55.7)	566 (45.0)	.024
3. Household income <$15,000	180 (20.4)	29 (23.6)	51 (29.1)	24 (30.4)	284 (22.6)	.023
4. High school or less education	390 (44.2)	64 (52.0)	68 (50.3)	61 (64.6)	593 (47.1)	.002
5. Inner city residence	386 (43.8)	61 (49.6)	89 (50.9)	51 (64.6)	587 (46.6)	.002
Risk factors considered as a sum, mean (SD)b	1.7 (1.8)	2.0 (1.8)	2.1 (1.9)	2.5 (1.7)	1.8 (1.8)	.001c

^adf = 3, p value is for chi square testing differences in rates of each risk factor occurring among the groups.

^b0 = Having none of the above demographic risk factors. 5 = Having all 5 risk factors.

^cPost-hoc contrast between having no abuse versus having abuse in both periods is the only significant difference.

Mental health morbidity (Table 2) occurred in a similar dose-response pattern. In this prenatal clinic sample of women who were nulliparas, 4.6% of the 882 women never exposed to abuse met the diagnostic criteria for PTSD during the first interview due to non-abuse types of trauma exposure, such as being in a natural disaster or an accident. The rate of PTSD was 12.5% among those abused in adulthood only, 16.9% among those abused in childhood only, and 41.3% among those who experienced abuse in both periods of their life. Dose-response patterns occur in relation to depression and generalized anxiety disorder diagnoses as well. Any history of abuse was associated with nearly double the rates of having ever used individual psychotherapy (43% versus 23%). Psychotherapy use during pregnancy increased from 3.1% for those never abused to 10.1% among those abused during both periods of life. Pre-pregnancy use of medication for psychiatric conditions ranged from 11.2% of those never abused to 27.8% of those with adult and childhood abuse. There was less variance and no statistically significant difference in rates of medication use in pregnancy, from 2.7% to 3.8% across the same groups.

Table 2.

Mental Health Morbidity Among Women Exposed to Abuse in Adulthood Only, Childhood Only, or in Both Periods

	No abuse (n=882) n(%)	Adult only (n =123) n (%)	Child only (n =175) n(%)	Both periods (n =79) n(%)	Overall (n=1,259) n(%)	pa
Diagnoses
Current PTSD	36 (4.1)	14 (11.4)	28 (16.0)	31 (39.2)	109 (8.7)	<.001
Current Depression	82 (9.3)	17 (13.8)	29 (16.6)	25 (31.6)	153 (12.2)	<.001
Current Anxiety	31 (3.5)	7 (5.7)	11 (6.3)	7 (8.9)	56 (4.4)	.064
Mental health treatment
Individual psychotherapy ever	206 (23.4)	52 (42.3)	76 (43.4)	35 (44.3)	369 (29.3)	<.001
Psychotherapy in pregnancy	27 (3.1)	7 (5.7)	12 (6.9)	8 (10.1)	54 (4.3)	.004
Medication use ever	99 (11.2)	24 (19.5)	29 (16.6)	22 (27.8)	174 (13.8)	<.001
Medication use in pregnancy	24 (2.7)	3 (2.4)	6 (3.4)	3 (3.8)	36 (2.9)	.898

PTSD = posttraumatic stress disorder

^adf = 3, p value is for chi square testing differences in rates among the groups

Substance use before pregnancy was assessed by asking if the woman ever used alcohol, tobacco, or drugs to cope with emotions or problems. Rates of using substances for coping (Table 3) were significantly higher among abused women in all groups. For example, among non-abused women, 11.5% reported generally using tobacco to cope with emotions or problems, but the rate was 35.4% among those abused both as children and as adults. Rates of reported use of substances during pregnancy were also greater among abuse survivors, with 5.4% of non-abused women smoking in pregnancy, compared with 21.5% of those abused both as children and as adults.

Table 3.

Use of Substances to Cope with Difficult Emotions Generally and Use of Substances in Pregnancy as Reported by Women Exposed to Abuse in Adulthood Only, Childhood Only, or in Both Periods

	No abuse (n=882) n(%)	Adult only (n =123) n (%)	Child only (n =175) n(%)	Both periods (n =79) n(%)	Overall (n=1,259) n(%)	pa
Which, if any, of these ways have you ever used to cope with emotions or problems?
Alcohol	114 (12.9)	302 (4.4)	36 (20.6)	23 (29.1)	203 (16.1)	<.001
Tobacco	101 (11.5)	35 (28.5)	49 (28.0)	28 (35.4)	213 (16.9)	<.001
Illicit drugs	58 (6.6)	20 (16.3)	32 (18.3)	17 (21.5)	117 (9.3)	<.001
Since you have been pregnant, have you used…?
Alcohol (any use)	133 (15.5)	23 (18.7)	20 (11.4)	18 (22.8)	194 (15.4)	.089
Cigarettes (≥1/day)	48 (5.4)	20 (16.3)	32 (18.3)	17 (21.5)	117 (9.3)	<.001
Illicit drugs (any)	24 (2.7)	8 (6.5)	10 (5.7)	13 (16.5)	55 (4.4)	<.001

^adf = 3, p value is for chi square testing differences in rates among the groups.

Of the 1259 women who completed the first interview, 357 women had progressed through the study to complete the postpartum interview at the time of this analysis. This group of women’s data was used to address the second aim: to explore experiences of being screened and choosing to disclose or not. Their verbatim statements about this disclosure decision comprised the data for the content analysis. In the 6-week postpartum interview, we asked the women several questions about screening for abuse in the prenatal and labor and delivery settings. Half of the 357 women who had completed the postpartum interview reported that their midwife or doctor used a form and/or asked face-to-face about whether they had ever been “hurt physically, sexually, or emotionally.” Of the 176 women who were asked, the vast majority (97.2%) thought the provider had done a good job of asking for this information.

In labor and delivery, many fewer women reported being asked about abuse by their labor and delivery nurse (34.7%). This lower rate may be due to the decreased likelihood of speaking to the woman privately when she is being admitted in labor. Of the 124 women who were asked on the labor unit, 97.6% reported that the nurse did a good job of asking.

We also asked about whether those who were abused chose to disclose this when they were screened, at either the clinic or at labor and delivery. In order to direct this questioning to abuse survivors only, we first asked, “Have you ever been physically or sexually assaulted?” When worded in this general way with a single question, and using the word “assault” rather than using the six behaviorally descriptive questions in the first prenatal interview’s extensive trauma history section, only 39 (10.9%) of the women answered “yes.” This included one-third of those who had reported adult-only or childhood-only abuse and half of those abused in both periods. Of these, nearly half (18) reported disclosing to their maternity care provider, and somewhat fewer (15) felt it was helpful for their provider to know this information about them. It is important to note that women made their decision to disclose or not in congruence with their judgment about whether disclosing would be relevant, helpful, or too painful. All 21 who did not disclose said they thought it would not have been helpful. Of the 18 who did disclose only two said they thought it was not helpful and one could not say if it was helpful or not. Disclosers and non-disclosers were not significantly different on any sociodemographic, psychiatric, or risk behavior characteristics.

In response to an open-ended question, “Can you describe what was helpful or not helpful?” there were three main reasons given for not disclosing. Some women thought it was not relevant, saying, “It wasn’t relevant. It happened a long time ago.” Some felt it was fully behind them, saying “It was way in the past, and I had already dealt with it.” Finally, some indicated it would be too painful to mention, saying “It wouldn’t have been relevant, and it would only bring up painful feelings,” or “Because it was wrong, and it was life, and I had to block it out.”

One reason given in favor of disclosing included just needing to talk about it, with women saying, “It was something to get off my chest,” and “Just talking about it….” Others gave specific reasons, for example “…just in case any anxiety came up during delivery,” or “Just the fact she knew if something was wrong with me because of the abuse I experienced when I was younger they would be able to take care of my son.” Benefits of disclosing included being understood and being helped, as two women explained, “…they were more understanding about concerns and the examinations and everything,” and “They made sure that they followed up with me to tell me about resources that were available based on how I was feeling about these past experiences.”

Finally, at the time of this analysis, medical record abstraction completed for 275 of the women who had given birth showed notation of any abuse history in only 22 (8.0%) of the 275 women. Among these 275 women whose chart data were available were 83 of the 377 women who disclosed an abuse history on the initial survey. Clinician screening and client disclosure had thus resulted in identification of 26.5% of those who disclosed abuse histories in the context of the research interview. Rates of noting abuse history did not differ across the three medical centers.

DISCUSSION

Abused pregnant women in this analysis were more likely to have demographic risk factors associated with poor perinatal outcomes, more likely to be using substances in pregnancy, and more likely to meet diagnostic criteria for posttraumatic stress disorder, depression, and anxiety. They also were more likely to have used mental health services in the past and to be in psychotherapy at the time of the first prenatal interview. The adverse impact of abuse on their mental health was related to the timing and accumulation of abuse. Those abused only in childhood were generally more adversely affected than those abused only in adulthood. Those abused in both periods of life were consistently the most adversely affected.

The rates of childhood abuse, adult abuse, and PTSD in our sample were similar to those reported in other U.S. studies. Although we find no U.S. studies of childhood abuse prevalence in an adult pregnancy sample, the rate reported of childhood sexual abuse in a middle class gynecology practice was 19.8%, ³³ and the rate of childhood sexual and physical abuse among pregnant adolescents was 33%. ⁵³ Rates of adult abuse prevalence in pregnancy vary but the range estimated by the CDC is 2.1% to 6.3%.⁵⁴ The range of PTSD rates among U.S. prenatal clinic samples is 3.5% ⁵⁵ to 7.7%.⁵⁶ Thus, it seems reasonable to expect that the rates of abuse and PTSD found in our sample to date are generalizable. The profile of our sample likely also is generalizable since these pregnant participants came from both private practice and public sector settings and are demographically diverse.

Furthermore, the prevalence of abuse and subsequent PTSD correlates with other negative health conditions for which routine screening and prevention are a standard of care. Women with a history of abuse, whether they have disclosed to their health care provider or not, represent an at-risk population for negative outcomes related to childbearing^{15, 57–59} and parenting. ^60–63 It would seem then that it is important to screen for the more detailed aspects of an abuse history, including childhood maltreatment and adult abuse in an effort to provide necessary assessment, secondary prevention and treatment for this population of childbearing women. There are effective treatments for these mental health conditions and effective interventions for the risk behaviors, many of which can be delivered in a trauma-informed manner to match abuse-related needs. Not only might case-finding and interventions potentially improve perinatal outcomes and women’s experience of childbearing, they have the potential to reduce maternal morbidity from injury. Ten per 1,000 pregnant women experience emergency department visits for inflicted injury, with another 2 per 1,000 being seen for self-inflicted injury, ⁶⁴ a risk behavior strongly associated with PTSD.^{13, 15} Case-finding and interventions may also prevent mortality since homicide, suicide, and drug overdose account for one-third of maternal deaths in urban areas.^{37, 65–66} This makes deaths due to partner violence and potentially abuse-related mental health and risk behavioral sequelae a more common, and potentially preventable, cause of U.S. maternal death than any of the three leading obstetric causes: embolism (21.4%), hypertensive disorders (19.4%), and hemorrhage (13.4%).⁶⁷

Barriers to screening for a more detailed history of abuse and trauma in clinical practice should be addressed. Consistent with newer literature regarding provider screening, the women in this analysis had positive experiences or felt their health care providers were competent when they did conduct both face-to-face and form-based assessments of abuse histories. While prenatal screening rates for violence and abuse were relatively high, the labor and delivery period represents challenges for screening, as reflected by the much lower rate of questioning done by nurses at this time.

Regardless of the timing, women in this analysis with a history of abuse decided if they were comfortable disclosing their histories whenever they were asked. Additionally, asking at varied times is more likely to elicit a positive response if it corresponds with a time when the women feels it would be useful to have this information included in her plan of care. The open-ended survey responses affirm that women do not mind being asked and are also able to selectively determine if answering will be too painful. Thus health care providers should not be discouraged from or fearful of eliciting negative responses when they do screening. This should not be a continued barrier to conducting routine screening regarding both a violence and sexual trauma history for childbearing women.

The rates of mental health sequelae in this sample also strongly support the need to follow disclosure of past or present abuse with assessment for posttraumatic stress, depression, and substance use and to make referrals as appropriate. This is in addition to the need to respond to disclosure of current abuse with danger assessment, safety planning, and referral to domestic violence services. It is also important to note that many affected women already had used mental health services in the past. So it is important to ask about treatment history, whether the woman deemed past treatment helpful or not, and what she would consider helpful now.

There are strengths and limitations to the data presented here. The strengths of this analysis include being able to draw from a large study with extensive, longitudinal measures, including chart data. The parent study has diversity in the clinical sites and demographics of the women who participated in the study. It uses detailed standardized instruments to assess and identify trauma, depression and PTSD. The use of telephone interviews provided women with a level of privacy not afforded them in a face-to-face interview. Additionally they were able to choose the timing of the call, which meant they could also choose the location where they received a call, promoting privacy and confidentiality in disclosing the information to the research team. This methodology may be more likely to yield a reliable response to the questioning than may be possible in the prenatal care setting. Chart review processes in the parent study also were developed to a point of excellent reliability (>=90% inter-rater agreement) and on-going audits indicate that quality of abstraction remains high (Seng, unpublished data).

There also are several limitations. First, the parent study is ongoing, so the proportion of women who have completed the postpartum survey and whose charts have been abstracted is smaller than it will be once all participants have delivered (in 2009) and all data collection processes are complete. Thus, we have a smaller sample available in the postpartum analyses. A comparison of the demographic, psychiatric, and risk behavior characteristics between all of the first survey completers and those who completed the last survey shows that there is no factor which differs significantly. Only power is reduced by the smaller sample size, not generalizability. Second, the study population, while diverse in its racial/ethnic composition, was limited to nulliparous women who were over the age of 18 and spoke English. Other populations of childbearing women, in particular adolescents and recent immigrant women were excluded but may have differing experiences of abuse and different abuse sequelae. Third, women with abuse in their history having subsequent birth experiences represent an important population for consideration regarding how and why they may decide to disclose in subsequent pregnancies. That interpersonal violence is correlated with rapid repeat pregnancy among low-income young women⁶⁸ suggests the need for screening and referral for multiparous women, but this study cannot provide data specific to this question. Reasons for disclosure in subsequent pregnancies may include additional difficulties, including traumatic first births, postpartum mental health or attachment delays for which abuse survivors, especially those with pre-existing psychiatric conditions, are more at risk.^69–70 Future studies should include this population for important information about their experience of screening as well as potential differences in their risk profile. Finally, it is a limitation that we cannot yet present findings on the effects of past and current abuse and PTSD on the perinatal outcomes of these women since the project is still ongoing. Although it is very important to learn about the effects of abuse and PTSD on maternal and infant physical health outcomes, the mental health morbidity and reliance on substances in pregnant that occurred more among the abused women in this study suggest that screening and trauma- and abuse-informed interventions via maternity settings are warranted for these reasons alone.

CONCLUSION

This analysis describes the mental health morbidity and risk behaviors among a larger proportion of abuse survivors than usually disclose in clinical settings. The dose-response pattern of morbidity and risk affirms that screening for both childhood and adult abuse histories is warranted because the long-term negative effects are most severe in those abused in both periods. Providers concerned about mental health morbidity and risk behaviors in their pregnant patients, who are not screening, may miss opportunities to provide trauma-informed care. The women in this study appeared to value screening, generally perceived that providers screen well, and were able to decide for themselves whether or not they wanted to disclose.

The screening research literature suggests that barriers to screening on the part of providers have evolved from lack of education or confidence, to a sense that screening is not worthwhile because there are no evidence-based interventions. Work is now needed to develop and disseminate interventions and allocate integrated mental health, substance abuse, and domestic violence services into prenatal care settings so that providers can overcome this barrier.

Biographies

Julia Seng is a certified nurse-midwife and research associate professor at the University of Michigan Institute for Research on Women and Gender and the School of Nursing and research assistant professor in the Department of Obstetrics and Gynecology.

Mickey Sperlich is a certified professional midwife and the study coordinator for the study on which this paper is based at the University of Michigan Institute for Research on Women and Gender.

Lisa Kane Low is a certified nurse-midwife and assistant professor of Nursing and Women’s Studies at the University of Michigan.

Appendix A. Literature-Based Health Care Provider Self-Assessment for Barriers to Screening for Violence and Trauma Histories and Knowledge of Best Practice Options

Which of these are potential barriers to screening for you?

• Belief that strategies to help did not exist ¹⁶

• Absence of effectiveness studies and demonstration of perceived benefit^{18, 38}

• Lack of confidence or comfort with inquiry ^{16,17,71–72}

• Lack of education ^{2,16–17, 25}

• Later entry to prenatal care by abused women leaves little time ¹⁹

• Perception that abuse is not a problem for their patients²

• Lack of time to deal with abuse ⁷³

• Question appropriateness of hospital-based screening given acute presentations²¹

• Perception that abuse is not a medical problem ²³

• Perception that abuse is a private issue ²³

• Perception that if a woman wants help she will ask for it ²³

• Language barriers and cultural differences ^{22, 27, 74}

• Lack of private opportunities to screen ²⁷

• Inadequate resources or referrals to help ^{20, 24}

• Concern they might offend or anger ²⁰

• Belief in low prevalence due to patient non-disclosure ^{24–25, 72}

• Concern that patient has fear of repercussions ²⁵

• Greater barriers for doctors if male, in private practice, not an OB/GYN ²⁶

Did you know this about women’s perspectives on screening?

• Provider characteristics matter, with women preferring to be screened by a woman, of the same race, 30–50 years old, and with nobody else present ²⁸

• A sense of trust, support, and nurturing matter ²⁹

• Concerns about confidentiality are barriers ²⁹

• Some women do not perceive childhood beatings as abuse, but as punishment ⁷⁵

• Some feel it is not relevant to their care ³¹

• Stigma is a barrier, sometimes complicated by fear or perception of discrimination ^{32, 76}

• Previous negative response from a provider deters future disclosure (e.g.: silence, expression of shock with no other reaction, doubt, or expressing the view that the abuse had no relevance to the current medical care). ^33,34

• African American women exposed to IPV are significantly more likely to report using prayer as a coping strategy⁷⁷ and less likely to seek mental health counseling. ^77,78

• Women experiencing IPV who have medical insurance coverage are more likely to seek mental health treatment than those without ⁷⁸

Are you aware of these best practices?

• Asking directly about abuse, using behaviorally-specific wording in a standardized manner detects more cases – although indirectly asking with a “safety” instrument is also effective. ^75,79

• Developing simple, time-efficient responses to disclosure, (e.g., wallet-sized referral cards, 20-minute nurse case-manager protocol) ⁸⁰

• Using a screening instrument (e.g., the Abuse Assessment Screen) ^81–82

• Providing training for providers that targets attitudes, beliefs, and communication skills increases screening rates. ^{16, 23, 83–85}

• Normalize inquiry by stating that “all patients are asked these questions” and use follow-up questions when appropriate ³⁸

• Use empathy, “I’m sorry that happened to you”, and support,“ This is not your fault”, “You did not deserve this” ³⁸