On March 11, 2011, an undersea earthquake (the Great East Japan Earthquake) measuring 9.0 on the Richter scale hit the northeast coast of Honshu Island, Japan. Its epicenter was located approximately 130 km east of the Pacific coast (Figure 1). The earthquake generated devastating tsunamis with waves that reached heights of 40 m. The tsunamis destroyed the towns and villages in the coastal areas of a number of prefectures, including the Iwate, Miyagi, and Fukushima prefectures. Approximately 18 400 people died or went missing, and more than 400 000 houses and buildings were completely or partially destroyed. 1
Figure 1 Location of Miyako, Iwate Prefecture The epicenter of the Great East Japan Earthquake is schematically (cross mark) indicated.
Previous studies have suggested that maternal mental health disorders, including perinatal depression, can be influenced by the devastation caused by a natural disaster.Reference Nishigori, Sugawara and Obara 2 - Reference Chang, Chang and Lin 5 Perinatal depression can affect the mental health of postnatal womenReference Sayil, Gure and Ucanok 6 and may impact the quality of care given to the newborn.Reference Murray, Fiori-Cowley and Hooper 7 Maternal psychosocial health can affect the mother-infant relationship,Reference Kokubu, Okano and Sugiyama 8 - Reference Porter and Hsu 10 the nutritional status of the infants, illness,Reference Rahman, Iqbal and Bunn 11 , Reference Turcotte-Tremblay, Lim and Laplante 12 and children's emotional or cognitive problems.Reference Hay, Pawlby and Waters 13 , Reference Talge, Neal and Glover 14 Therefore, perinatal women should be considered to be a vulnerable population, and methods to prevent perinatal depression must be established in preparation of future devastating disasters.
The Iwate Prefectural Miyako Hospital is located on the northeastern coast of Japan (Figure 1), and its medical service area includes Miyako, Iwaizumi, Yamada, and Tanohata. According to 2010 census data, this area has a population of approximately 92 000, and approximately 550 births occur annually. Approximately 390 births annually occur at the Iwate Prefectural Miyako Hospital. Following the Great East Japan Earthquake and tsunamis, approximately 1500 people died or went missing and more than 12 600 houses and buildings were completely or partially destroyed in the service area. Following this disaster, the Iwate Prefectural Miyako Hospital served as a disaster base hospital, and it provided intensive maternal mental health care to prevent perinatal depression.
In the present study, we retrospectively examined scores on the Japanese version of the Edinburgh Postnatal Depression Scale (EPDS)Reference Cox, Holden and Sagovsky 15 , Reference Okano, Murata and Msuji 16 of perinatal women who were treated at the Iwate Prefectural Miyako Hospital to assess the influence of the Great East Japan Earthquake on their mental health.
Methods
Study Design
This retrospective study was conducted using a medical chart–based survey and a questionnaire that was completed by perinatal women who gave birth at the Iwate Prefectural Miyako Hospital. Existing medical records were reviewed in this observational study. Because this was a retrospective study, informed consent was omitted according to the Ethical Guidelines for Clinical Studies issued by the Japanese Ministry of Health, Labour, and Welfare. 17 The study protocol was approved by the ethics committee of the Iwate Prefectural Miyako Hospital on January 17, 2012, and conformed to the provisions of the Declaration of Helsinki (revised in Tokyo 2004). 18
Study Subjects
The study subjects included perinatal women who had delivered at the Iwate Prefectural Miyako Hospital between October 1, 2010, and September 10, 2011. The hospital had been assessing these patients with the Japanese version of EPDSReference Cox, Holden and Sagovsky 15 , Reference Okano, Murata and Msuji 16 at the time of postdelivery hospital discharge and their regular 1-month postdelivery checkup. Perinatal women who gave birth between February 10, 2011, and March 10, 2011, were excluded because they had undergone the hospital discharge EPDS test prior to the earthquake and the 1-month EPDS test after the earthquake.
Data Collection
We collected the following data from medical charts: age, marital status, parity, employment, mental disorder history, obstetric complications during pregnancy (ie, threatened premature labor, preterm delivery, pregnancy-induced hypertension, fetal growth restriction, hemolysis, elevated liver enzymes, low platelet count syndrome, placenta previa, placental abruption, placenta accreta, gestational diabetes, or intrauterine fetal death), type of delivery, abnormal delivery (emergency cesarean section, vacuum extraction, obstructed labor, or uterine atony), postpartum obstetric complications (vulvar hematoma, problematic surgical scaring, or breast abnormalities), and scores on the Japanese version of the EPDS at hospital discharge and at the regular checkup 1 month after delivery.
Women with vaginal deliveries took the EPDS test 4-5 days later at the time of hospital discharge, whereas those who delivered by cesarean section took the test 6-7 days later at the time of hospital discharge. At the Iwate Miyako Prefectural Hospital, women are routinely admitted for 4-5 days after uncomplicated vaginal deliveries and for 6-7 days after cesarean deliveries.
The Japanese version of EPDS is a self-reported scale that contains 10 items, each of which is graded with a 4-point Likert scale, with the total score ranging from 0 to 30. The cutoff score in the Japanese population is 9, which is considered to indicate a significant risk for postpartum depression.Reference Okano, Murata and Msuji 16 , Reference Yamashita, Yoshida and Nakano 19 The self-administered questionnaire administered at hospital discharge after delivery included questions about whether perinatal women could consult their families when they had problems, anxiety over household finances, dissatisfaction with living conditions, the death of loved ones, and worries about baby care.
We collected information on the feelings of the women who gave birth after March 11, 2011, about being a disaster victim, the destruction of their homes, residence issues, and family members who had lost their jobs. The definition of a destroyed home was in accordance with the guidelines authorized by the Cabinet Office of the Japanese Government in 2011. 20
Analysis
The percentages of women with EPDS scores ≥9 at hospital discharge and 1 month after delivery were compared among the following 3 groups: the before-disaster group, which consisted of women who gave birth before the disaster (October 1, 2010-February 9, 2011); the within-3-months group, which consisted of women who gave birth within 3 months after the disaster (March 11, 2011-June 10, 2011); and the 4-6-months group, which consisted of women who gave birth 4-6 months after the disaster (June 11, 2011-September 10, 2011). Student's t tests, χ2 tests, and Fisher’s tests were used when appropriate for statistical analyses.
Multivariate logistic regression analyses were performed after adjusting for the variables that were significantly associated with EPDS scores ≥9 at hospital discharge among the after-disaster group in univariate analysis. The after-disaster group included both the within-3-months group and the 4-6-months group. The adjusted odds ratios (AORs) and 95% confidence intervals (CIs) were calculated to estimate the risk of EPDS scores ≥9 at hospital discharge after the disaster. All statistical analyses were performed with SAS version 9.3 statistical software (SAS Institute, Inc, Cary, NC).
Results
Of the 141 women in the before-disaster group, 141 and 134 underwent the EPDS test at the time of discharge and 1 month after delivery, respectively. Of the 73 women in the within-3-months group, 70 and 67 underwent the EPDS test at hospital discharge and 1 month after delivery, respectively. Of the 90 women in the 4-6-months group, 89 and 86 underwent the EPDS test at hospital discharge and 1 month after delivery, respectively.
Comparisons of the percentages of women with EPDS scores ≥9 at hospital discharge and 1 month after delivery are shown in Figure 2. The percentage of women with EPDS scores ≥9 at hospital discharge was significantly greater in the within-3-months group (31.4%, P<.0001) than in the before-disaster group (9.9%); this difference was not significant 1 month after delivery (7.5% versus 2.2%, P=.12). There was no significant difference in the percentages of women with EPDS scores ≥9 between the 4-6-months group and before-disaster group at hospital discharge (10.1% versus 9.9%, P=.96) or 1 month after delivery (4.7% versus 2.2%, P=.44).
Figure 2 Percentages of Women With Scores ≥9 on the Japanese Version of the Edinburgh Postnatal Depression Scale (EPDS) at Hospital Discharge and 1 Month After Delivery
The background characteristics of the perinatal women who took the EPDS test at hospital discharge in the before-disaster group and in the after-disaster group are shown in Table 1. In the after-disaster group, univariate analysis showed that the percentages of women with obstetric complications during pregnancy (P=.03), dissatisfaction with living conditions (P=.0001), the death of loved ones (P=.048), the destruction of their home (P<.0001), and a change in residence after the disaster (P=.0001) were significantly different between the group with EPDS scores ≥9 and the group with scores <9 (Table 1). Because the correlation between the destruction of their home and a change in residence after the disaster was multicollinear, we chose the former variable for multivariate logistic regression analysis. Multivariate logistic regression analysis was adjusted for obstetric complications during pregnancy, the death of loved ones, dissatisfaction with living conditions, and the destruction of their home. This analysis showed that the destruction of their home (AOR, 3.02; 95% CI, 1.20-7.59) and dissatisfaction with living conditions (AOR, 3.68; 95% CI, 1.46-9.26) were significantly and independently associated with EPDS scores ≥9 at hospital discharge after the disaster (Table 2).
Table 1 Background Characteristics of Perinatal Women Who Underwent the Japanese Version of the Edinburgh Postnatal Depression Scale (EPDS) at Hospital Discharge
a The after-disaster group included both the within-3-months group and the 4-6-months group.
Table 2 Multivariate Logistic Regression Analysis of the Data for Perinatal Women With Scores ≥9 on the Japanese Version of the EPDS Score at Hospital Discharge in the After-Disaster Group
Abbreviations: EPDS, Edinburgh Postnatal Depression Scale; SE, standard error; AOR, adjusted odds ratio; CI, confidence interval.
a Adjusted for “obstetric complications during pregnancy,” “death of loved ones,” “dissatisfaction with living conditions,” and “destruction of their home.”
b The correlation between “destruction of their home” and “change in residence after the disaster” was multicollinear.
Discussion
The percentage of women with EPDS scores ≥9 at hospital discharge was significantly greater in the within-3-months group than in the before-disaster group (P<.0001). This difference was not significant (P=.12) 1 month after delivery. Several explanations for this finding have been proposed. First, an increase in the number of patients with perinatal depression after the disaster was predicted; therefore, the hospital conducted a recollection of birth (birth review)Reference Mercer 21 , Reference Sleutel 22 earlier and introduced touchReference Anderson and Taylor 23 , Reference Tanaka, Masaoka and Koiwa 24 to encourage patients to express their emotions. Second, because damage to parenting environments such as housing without essential utilities was expected, the hospital provided support to promote breastfeeding and advised new mothers about the level of damage at the time of discharge, and they received parenting support at hospital discharge and at checkups performed 2 weeks and 1 month after delivery. Although the difference of 1 month after delivery was not significant (5 of 67 versus 3 of 134, P=.12), the study was statistically limited because the number of clinical cases was small. Therefore, the results must be interpreted with caution.
There were no significant differences between the 4-6-months group and the before-disaster group at hospital discharge (10.1% versus 9.9%, P=.96) and 1 month after delivery (4.7% versus 2.2%, P=.44). The prevalence of EPDS scores ≥9 at the Iwate Prefectural Miyako Hospital differed from the findings of a study that was conducted in the coastal area of Miyagi Prefecture after the disaster.Reference Nishigori, Sugawara and Obara 2 Miyagi Prefecture is located to the south of Iwate Prefecture (Figure 1). In the coastal area of Miyagi Prefecture, which was also affected by the disaster,Reference Shibahara 25 , Reference Shibahara 26 the percentage of postpartum women with EPDS scores ≥9 was 21.3% around 6 months after the disaster.Reference Nishigori, Sugawara and Obara 2 Further investigations are required to clarify the factors correlated with this difference.
Multivariate logistic regression analyses were performed in the after-disaster group with EPDS scores ≥9 at hospital discharge. Because the number of clinical cases was small in multivariate logistic regression analyses in the within-3-months group and in the 4-6-months group separately, the after-disaster group included both the within-3-months group and the 4-6-months group for multivariate logistic regression analyses. For the risk factors for perinatal depression that were related to the disaster, previous studies have found that exposure to storm,Reference Xiong, Harville and Mattison 27 loss of resources,Reference Ehrlich, Harville and Xiong 28 earthquake experience,Reference Qu, Tian and Zhang 29 , Reference Qu, Wang and Tian 30 anxiety about earthquakes,Reference Hibino, Takaki and Kambayashi 31 , Reference Hibino, Takaki and Kambayashi 32 and exposure to tsunamiReference Nishigori, Sugawara and Obara 2 were more likely to cause depression in pregnant and postnatal women who were affected by natural disasters. In the present study, we found that the destruction of their home and dissatisfaction with their living conditions were significant risk factors for postpartum depression among women in the after-disaster group.
Limitations
This study had some limitations. First, this study was based on a single institution, which decreases the generalizability of the results. There were 2 obstetric clinics other than the Iwate Miyako Prefectural Hospital in the region, and this study did not include all perinatal women in the Miyako medical service area. Second, this study did not investigate the impact of “maternity blues” on the results. Third, the study also had statistical and epidemiological limitations because the number of clinical cases was small.
Conclusions
In conclusion, the percentage of women with EPDS scores ≥9 at hospital discharge was significantly greater in the group that delivered within 3 months after the Great East Japan Earthquake than in the group that delivered before the disaster. The destruction of their home and dissatisfaction with living conditions were more likely to cause perinatal depression in women who delivered after the disaster. An increase in postpartum depression is predicted after a large-scale disaster, and interventions focused on pregnant women should be actively initiated after the disaster.
Acknowledgments
We thank Dr Hideyuki Chida, Dr Hideo Kodama, Dr Takeshi Obara, Mrs Toshiko Kurabe, and Dr Yoshihiro Akimoto for their expertise and guidance.
This work was supported by JSPS KAKENHI (C) Grant Number 24592457
