Development of a screening tool for pregnant women who need social support at obstetric institutions: A prospective cohort study


 Background
 To develop an efficient interview sheet and the screening tool to detect pregnant women in need of social support at obstetric institutions.
Methods
 A prospective cohort study was conducted at three medical institutions that were familiar with social issues. The pregnant women were assessed using an interview sheet at the first visit (n = 678), at approximately 28 weeks gestation (n = 495), 36 weeks gestation (n = 296), and the postpartum period (n = 822). We investigated the considered important items on the interview sheet for women in need of social support who were selected based on conferences at three medical institutions. The primary outcome was to develop the available interview sheet during pregnancy and the screening tool to detect pregnant women in need of support to require collaboration with a public health center. The items on the interview sheet were scored by multiple linear regression analysis, and the cutoff values were calculated using the receiver operating characteristic curve. The secondary outcome was to investigate the perinatal factors associated with the Edinburgh postnatal depression scale (EPDS). The association between perinatal factors and EPDS was assessed using multiple logistic regression analyses.
Results
 One hundred sixty-six cases for which all interview sheets for all periods were available. Interview sheets and stepwise screening tools during the first and second trimester were developed in which 2.5% of pregnant women were identified as requiring social support, respectively. The factors associated with EPDS ≧ 9 were ‘Women who felt confused/troubled or did not feel anything to be pregnant’ (adjusted odds ratio [aOR]: 6.51, 95% confidence interval [CI]: 1.62–26.15), ‘Maternal mental disorder’ (aOR 4.38; CI 1.06–18.10), ‘Consultation request at first visit’ (aOR 3.22; CI 1.09–9.45), and ‘Women who have difficulty or anxiety about pregnancy during the second trimester’ (aOR 3.14; CI 1.29–7.67).
Conclusions
 We created the interview sheets and screening tools for three steps during the first and the second trimester. Future studies are needed to validate these screening tools.


Abstract Background
To develop an e cient interview sheet and the screening tool to detect pregnant women in need of social support at obstetric institutions.

Methods
A prospective cohort study was conducted at three medical institutions that were familiar with social issues. The pregnant women were assessed using an interview sheet at the rst visit (n = 678), at approximately 28 weeks gestation (n = 495), 36 weeks gestation (n = 296), and the postpartum period (n = 822). We investigated the considered important items on the interview sheet for women in need of social support who were selected based on conferences at three medical institutions. The primary outcome was to develop the available interview sheet during pregnancy and the screening tool to detect pregnant women in need of support to require collaboration with a public health center. The items on the interview sheet were scored by multiple linear regression analysis, and the cutoff values were calculated using the receiver operating characteristic curve. The secondary outcome was to investigate the perinatal factors associated with the Edinburgh postnatal depression scale (EPDS). The association between perinatal factors and EPDS was assessed using multiple logistic regression analyses.

Results
One hundred sixty-six cases for which all interview sheets for all periods were available. Interview sheets and stepwise screening tools during the rst and second trimester were developed in which 2.5% of pregnant women were identi ed as requiring social support, respectively. The factors associated with EPDS ≧ 9 were 'Women who felt confused/troubled or did not feel anything to be pregnant' (adjusted odds ratio [aOR]: 6.51, 95% con dence interval [CI]: 1.62-26.15), 'Maternal mental disorder' (aOR 4.38; CI 1.06-18.10), 'Consultation request at rst visit' (aOR 3.22; CI 1.09-9.45), and 'Women who have di culty or anxiety about pregnancy during the second trimester' (aOR 3.14; CI 1.29-7.67).

Conclusions
We created the interview sheets and screening tools for three steps during the rst and the second trimester. Future studies are needed to validate these screening tools.
Background Among the veri ed child abuse cases which resulted in death [1], children under 1 year of age accounted for about 50% among the deaths due to childhood maltreatment, and half of them are within the rst month of life. It suggests the importance of early preventive efforts against child maltreatment during pregnancy. It is reported that the background factors related to child maltreatment were unexpected/unplanned pregnancies, failure to receive pregnancy health checkups, failure to receive the Maternal and Child Health Handbook, and teenage pregnancies [1,2]. Other factors such as unmarried mothers, maternal mental disorder, low educational achievement, a history of childhood abuse, deprivation, and low-birth-weight children have also been reported to be associated with child maltreatment [2][3][4]. It is important to understand such social risk factors as well as medical risk factors. Identifying social risk factors at the time of pregnancy and initiating support for pregnant women who need social support may help reduce child maltreatment.
Although there are many public health centers that detect pregnant women who need social support through questionnaires and interviews when issuing the Maternal and Child Health Handbook, it is limited due to the few opportunities for contact with pregnant women. On the other hand, pregnancy health checkups at obstetric institutions are scheduled at least 14 times in Japan. In pregnancy health checkups, the mother's health condition is assessed, and a medical examination and health guidance are conducted. Health guidance has the role of taking care of the mental health of pregnant women and relieving their anxiety about pregnancy, childbirth, and childcare. However, about half of the delivery facilities in Japan are private obstetric clinics, and some facilities are inadequate for resolving social and psychological problems. Therefore, the aim of this study is to develop a screening tool for detecting pregnant women in need of support that can be used in health guidance at all facilities, including private obstetric clinics.

Methods
This was a prospective cohort study conducted at three medical institutions that were familiar with social issues. We created new interview sheets at the rst visit, at approximately 28 weeks gestation, 36 weeks gestation, and the postpartum period. Table 1 shows the items of interview sheets for each period. These interview sheets were conducted at three medical institutions (Osaka Women's and Children's Hospital, Seibo Hospital, and Showa University Hospital) that use their own methods to identify pregnant women in need of support and collaborate with a public health center. In addition, the Edinburgh postnatal depression scale (EPDS) was administered together at one-month postpartum. At the three medical institutions, information was collected from face-to-face interviews, impressions based on behavior and actions, medical information, and original questionnaires. Pregnant women in need of administrative cooperation were selected by conferences at Osaka Women's and Children's Hospital and Seibo Hospital and by nurses specializing in maternity nursing at Showa University. The members of the conference at Osaka Mothers and Children were nurses, midwives, public health nurses, caseworkers, and obstetricians, while the members at Seibo Hospital were nurses, midwives, caseworkers, and pediatricians. EPDS; the Edinburgh postnatal depression scale We investigated the considered important items on the new interview sheet for women in need of social support who were selected based on conferences at three medical institutions. The primary outcome was to develop an available interview sheet during pregnancy and the screening tool to detect pregnant women in need of support in collaboration with a public health center. A secondary outcome was to investigate the perinatal factors associated with EPDS.
Standardized partial regression coe cients were calculated by multiple linear regression analysis to re ne and score each item on the interview sheet. Scores were calculated by multiplying the standardized partial regression coe cient by 100 and rounding it off to the closest whole. The cutoff value was calculated by the receiver operating characteristic (ROC) curve. The variance in ation factor was calculated using regression analysis to con rm multi-collinearity in the multivariate analysis. The association between perinatal factors and EPDS was assessed using multiple logistic regression analysis. All statistical analyses were performed using SPSS Statistics ver. 27 software. Two-sided P-values of < 0.05 were considered statistically signi cant.
This research related to human use is in compliance with all the relevant national regulations, institutional policies and the tenets of the Helsinki Declaration. This research was approved by the Ethics Review Board of the University of Yamanashi (approval number 1663, approval date July 10, 2017). Written informed consent was obtained from all study subjects. At the time the interview sheets from the three medical institutions were mailed to the University of Yamanashi, the data center, research IDs were assigned, and names and IDs for each hospital were deleted. The matching forms of research IDs and hospital-speci c IDs were managed at each institution. Table 2 showed the number of responses to the interview sheet and the number of pregnant women in need of social support who were judged to need cooperation with the public health center at each of the periods. We received 678 responses at the rst visit, 495 during the second trimester, 296 during the third trimester, and 822 for postpartum. Figure 1 showed that the change in support status during pregnancy and postpartum for the 166 cases for which all interview sheets for all periods were available. In this study, we investigated socially high-risk pregnant women who were supported by their respective obstetric institutions as well as collaborating with a public health center. In the rst trimester, the number of pregnant women in need of support to require collaboration with a public health center was 2%, and the number of socially high-risk pregnant women was 17%. During the second and third trimesters, the number of pregnant women in need of support to require collaboration with public health centers was 6%, which was higher than that in early pregnancy, and it was the highest at 23% in the postpartum period. There was a change in the support status during pregnancy. There were some cases in which cooperation with public health centers was started but later changed to support at the respective medical institutions or some cases in which support was no longer needed. There were also 13 cases where there were no problems during pregnancy, but support was needed after delivery.  Table 3 shows the results of the multiple linear regression analysis at the rst visit. Compared to the results of the single regression analysis (Model-1) for each item, in the multiple regression analysis in which all factors were input, the results of the items " partner's feelings toward being pregnant " and " consultation request " were reversed, so they were excluded from the study. Furthermore, among the items on the interview sheet, "the use of mother's and partner's illegal drug" were excluded from the screening tool because they clearly required cooperation with a public health center. The questionnaire after exclusion of the four items was used as Model-2, and the score of each item was calculated. The cutoff value for the detection of pregnant women in need of social support was calculated to be 22 points from the ROC curve, and scores of 22 points or more were selected as STEP-1, with an area under the curve (AUC) of 0.96, the sensitivity of 0.92, and speci city of 0.92. Of the 558 subjects analyzed at the rst visit, 54 cases were detected for STEP-1. In order to further improve the selection rate, we checked the applicable percentage of each item in the group that actually had cooperation with a public health center in this study. The factors that accounted for more than 30% of the total were " no family and social support," "maternal mental disorder," "worry about the older child," and "multiple pregnancies," and those who had at least one of these four factors were selected as STEP-2 in addition to STEP-1. As a result, 20 cases were selected, of which six cases (30%) were actually involved in cooperation with a public health center. In addition, those with at least two of the following were selected as STEP-3: "depression symptoms," "maternal mental disorder," and " consultation request." The number of pregnant women in need of social support based on the screening by three STEPs was 14, of which 5 (36%) actually required administrative cooperation. The number of positive screening results was 14 (2.5%) out of 558. The screening tool is shown in Fig. 2.  Table 4 shows the results of the multiple linear regression analysis during the second trimester. Compared to the results of the single regression analysis (Model-1) for each item, in the multiple regression analysis in which all factors were entered, the item " worries about the older child " was reversed, so it was excluded from the study. In addition, "alcohol use" and " status of pregnancy health checkups " were excluded because they could not be analyzed due to missing data. The interview sheets after the exclusion of the three items were used as Model-2, and the scores for each item were calculated. The cutoff value for pregnant women in need of social support was determined to be 14 points from the ROC curve, and scores of 14 points or more were extracted as STEP-1, with an AUC of 0.77, sensitivity of 0.82, and speci city of 0.62. Of the 483 cases analyzed during the second trimester, 110 cases were extracted by STEP-1, of which 34 cases (31%) were had cooperation with public health centers in this study. In order to increase the selection rate, we examined the factors on the interview sheets of the women who scored more than 14 points. However, since the percentage of the risk group was more than 30% for all items, we calculated the Spearman correlation coe cient for the subjects with 14 points or more, weighted the correlation coe cient by a factor of 10, and considered variables with 2 points or more to be factors with higher risk. These items were " no talking with partner," "partner violence," and "smoking," and if any of these items were applicable, they were narrowed down as STEP-2 in addition to STEP-1. Twenty-one people were selected for STEP-1 and STEP-2, of which 12 (57%) required administrative cooperation in this study. In addition, STEP-3 was selected for those who had at least one of the following: "no talking with partner," "partner violence," "depression symptoms," and "consultation request. As a result, 12 cases were selected, and all of them were the subjects of the actual administrative collaboration in this study. The number of positive cases by the screening by three STEPs was 12 out of 483 (2.5%). The screening tool is shown in Fig. 3. Table 5 shows the results of the multiple linear regression analysis during the third trimester. Each item on the interview sheets was scored, and a cutoff value of 7 points was calculated using the ROC curve. The AUC, sensitivity, and speci city were 0.46, 0.43, and 0.57, respectively, which were inappropriate for screening. In addition, we examined the relationship between the EPDS at the one-month postpartum and each item on the interview sheets. The factor "no family and social support" and "mother or partner's use of illegal drugs" at the rst visit, and "no family and social support," "alcohol use," and "less or excess pregnancy health checkups" during the second trimester were excluded due to their small number. trimester' (aOR 3.14, CI 1.29-7.67). The factor "less or excess pregnancy health checkups" during the third trimester was excluded due to the small number of cases, and no factor associated with EPDS was found.

Discussion
The interview sheets and the stepwise screening tool by three STEPs tools during the rst and second trimester were developed in which 2.5% of pregnant women were identi ed as requiring social support, respectively. The screening tool was created with the goal of being able to detect women, even in obstetric clinics that were not familiar with social issues, automatically. The screening tool was created with the intention of narrowing down the number of targets because if there are too many targets, the health center may not be able to handle them.
According to a survey in Osaka Prefecture, cases of deliveries of little or no receiving for pregnancy health checkups accounted for about 0.3%. (Unpublished data) In addition, the frequency of speci ed expectant mothers was reported to be 2-5% [5,6]. Speci ed expectant mothers are de ned in the Child Welfare Act as "pregnant women who are recognized as being in particular need of support during pregnancy with regard to postpartum care." A speci ed expectant mother is determined by the health center. In this study, we aim to identify pregnant women in need of social support or those who are expected to consider speci ed pregnant women. This study aims to detect pregnant women in need of support in collaboration with health centers, that is, cases corresponding to speci ed pregnant women. The detection rate using these screening tools is about 2.5% during the rst and the second trimester, respectively, which is considered to be a reasonable number. In the future, it will be necessary to verify whether these interview sheets and screening tools can appropriately identify pregnant women in need of social support, and we are planning to create a system that can automatically identify pregnant women in need of social support by inputting the results of the interview sheets.
The frequency of postpartum depression is reported to be 5-10% in a prospective study in Japan [7] and 9% in the nal report of the "Healthy Parents and Children 21" [8]. Furthermore, postpartum depression may lead to maternal suicide as well as child maltreatment [9][10][11]. In Japan, a suicide prevention program (Nagano model) [12] has been developed to improve postpartum mental health and reduce maternal suicide, and its effectiveness has been reported. In the present study, the factors associated with EPDS ≧ 9 at one-month postpartum were 'Women who felt confused/troubled or did not feel anything to be pregnant,' 'Maternal mental disorder,' 'Consultation request at rst visit,' and 'Women who have di culty or anxiety about pregnancy during the second trimester.' Although postpartum depression may develop unexpectedly, it has been suggested that it is related to maternal mental disorder, unwanted pregnancy, and anxiety during pregnancy. According to the reports of the Japan Environment and Children's Study, compared with women who felt very happy to be pregnant, those whose pregnancy was unintended but happy, unintended and confused, those who felt troubled, and those who felt no emotion toward being pregnant had increased risks of postpartum depression [13].
The strength of this study is a prospective survey conducted in several obstetric institutions that are specialized in addressing social issues, and various criteria were used to determine the detection of pregnant women in need of social support. In addition, considering the possibility of changes in the mother's and family's condition during pregnancy and background factors that may become apparent through the establishment of a relationship between the mother and the obstetric institution, interview sheets and screening tools were developed that included items tailored to each period and items common to all periods.
Several limitations should be addressed. First, less than half of the cases had a complete questionnaire for all time periods. This was due to the fact that two of the three facilities in this study handle a large number of medically high-risk pregnancies, which may result in rst visits at various times. Second, although it had been reported that a late rst visit is associated with pregnant women in need of social support, it was not included in this study. This was because the study includes many cases in which the rst visit was made after the second trimester due to medical risks. Finally, this study was conducted up to one month after childbirth, and the relationship with the need for support at a public health center after that was unknown. Although it would be desirable to collate the data with information from the public health center if there was an actual fear of child maltreatment several months after the birth, it was di cult to obtain the data from the public health center from the viewpoint of personal information protection. In a survey such as this one, it is considered necessary to create an environment where medical institutions and government agencies can collaborate to collect information in Japan. Figure 1 The change in support status during pregnancy and postpartum

Figures
There was a change in the support status during pregnancy. †Hospital support: socially high-risk pregnant women who were supported by their respective obstetric institutions ‡Administrative cooperation: pregnant women in need of support to require collaboration with public health center Page 19/20

Figure 2
The screening tool at rst visit Pregnant women with score of 22 points or more were selected as STEP-1. Then, pregnant women are narrowed down as STEP 2 and 3 according to the items on the interview sheet at rst visit.