- Top of page
Aim: Postnatal depression is one of the life-threatening events faced by women. As the factors associated with postnatal depression have not been investigated fully in Japan, we studied the factors associated with postnatal depression.
Methods: One-hundred-and-sixty-nine women who visited the health center of a city in Aomori Prefecture, Japan, at 4 months after childbirth for regular examination fulfilled the selection criteria and completed self-reporting questionnaires on postnatal depression using the Edinburgh Postnatal Depression Scale (EPDS) and a life and social events scale.
Results: The primiparas showed a significantly higher EPDS score than the multiparas. The EPDS score decreased with the frequency of delivery in the groups of mothers in their twenties and thirties. For the multiparas, the number of participants who suffered obstetric events was lower, the number of participants who went back to their home was higher, and the number of participants who were taken care of by their mother was lower than for the primiparas. General health abnormalities, sociability, and worries about baby care were significantly associated with the EPDS for both the primiparas and multiparas. The cooperation of the husband was associated with a decreased EPDS score, both for the primiparas and multiparas, irrespective of the family structure.
Conclusion: The EPDS score decreased with an increased frequency of delivery, suggesting that the experience of delivery would impact on postnatal depression, partly because of decreased obstetric events. However, a social assistance system is needed for women with general health abnormalities, less sociability, worries about baby care, and limited cooperation of the husband for both primiparas and multiparas.
- Top of page
In a woman's life cycle, the puerperal period is a crisis period because of physiological factors (e.g. rapid changes in the levels of hormones in the body), psychological factors (e.g. worries about childbirth and child care), and social factors (e.g. support systems for pregnant and parturient women). This period has been examined as one associated with the likely onset of mental disorders. In particular, post-partum depression is common and, from a public health point of view, identifying the risk factors for postnatal depression is one of the major challenges because of the high prevalence of this disorder and its devastating consequences for the mother.
Combinations of risk factors are likely to contribute to any individual woman's vulnerability to post-partum depression. Maternal prenatal depression or anxiety, a history of depression, child-care stress, a poor marital or partner relationship, low social support, and life stress are the most consistently identified predictors of post-partum depression (Horowitz & Goodman, 2004). Parity and age have not been predictors of post-partum depression, but younger mothers (Troutman & Cutrona, 1990) and first-time mothers (Hobfoll, Ritter, Lavin, Hulsizer & Cameron, 1995) might experience higher rates of post-partum depression, suggesting that maternal age and parity could be risk factors worthy of ongoing evaluation. Other factors, such as socioeconomic status and infant temperament, are identified less consistently and tend to have smaller effect sizes in relation to post-partum depression (Beck, 2001; Nielson, Videbech, Hedegaard, Dalby & Secher, 2000; O'Hara & Swain, 1996). A low socioeconomic status and chronic stress, such as from maternal health problems, infant difficulty, a lack of money, and poor social support, also have been associated with more long-lasting post-partum depression (Seguin, Potvin, St-Denis & Loiselle, 1999).
Most of these factors of post-partum depression have been analyzed in Western countries. Horowitz, Chang, Das and Hayes (2001) pointed out that the cultural context shaped the women's interpretation of their post-partum experiences and suggested the investigation of what is helpful in relieving distressing symptoms within the women's particular cultural context. An international study that explored the levels of post-partum depressive symptomatology revealed that European and Australian women had the lowest levels of post-partum depression, women in the USA fell at the midpoint, and women from Asia and South America had the highest depressive symptom scores (Affonso, De, Horowitz & Mayberry, 2000). Japan was not included in the study. So far, a few studies have been undertaken on the factors contributing to post-partum depression in Japan. Tamaki, Murata and Okano (1997) reported that post-partum depression in Japanese women correlated with primiparous premature delivery, difficult labor, experience of life events, and worries about baby care. Yamashita, Yoshida, Nakano and Tashiro (2000) demonstrated a similar incidence of post-partum depression in Japanese women as found in women in England (Yoshida et al., 1997) in spite of the difference in the sociocultural environment, but they found few differences in the demographic characteristics between the depressed and non-depressed participants.
As those Japanese data were not reported consistently and might not be similar to those of Western women, we studied the factors contributing to post-partum depression in women living in a rural city of Aomori Prefecture, in the northern part of Japan, where a more traditional Japanese lifestyle exists; that is, about half of the post-partum women stayed with their extended families and more than half of the post-partum women had their parents' support, or Satogaeri. Satogaeri means a pregnant or parturient woman who returns to her family to receive help from her own mother, so that the new mother can spend this physically and emotionally unstable time under less stress.
- Top of page
We studied 169 women who gave birth in a rural city in Aomori Prefecture, in the northern part of Japan (population: 174 000; number of births in 2002: 1300), between January and March 2002, and who brought their infant for a check-up at 4 months after delivery at a health center. We could not contact the mothers who suffered from psychotic disorders and who consulted with a psychiatrist. Answers were obtained using self-reporting questionnaires from 196 women who agreed to participate in the study out of 205 women who visited the health center. The questionnaires were directly distributed and collected at the health center. The mothers who had babies with a birthweight of <2000 g directly consulted pediatricians and were not included in the health center. One-hundred-and-sixty-nine complete questionnaires were filled out and adopted in the present analysis.
The Edinburgh Postnatal Depression Scale (EPDS) has been used in screening for post-partum depression (Cox, Holden & Sagovsky, 1987). The EPDS is a 10-item self-reporting scale that seems to be especially valid for use with child-bearing women and it is highly effective from the viewpoint of sensitivity and specificity (Cox, 1994). Each item is scored on four-point scales from 0–3, with the minimum and maximum total scores being 0 and 30 points, respectively. A Japanese version of the EPDS was used in the present study (Okano et al., 1996).
In order to ascertain the sociopsychological state of child-rearing burnout of post-partum depression, five different kinds of questionnaire were developed: physical characteristics, obstetric factors, baby care, cooperation of the husband, and family structure (Sasaki, Takanashi & Hongoh, 1991). We followed the same parameters used for child-rearing burnout by Sasaki et al. in the present study as follows.
The physical characteristics were: previous childbirth experience (primiparas and multiparas), age, occupation, general health abnormalities, and the sociability of the participants. The economic status, social class, and educational level were not analyzed in the present study.
The general health abnormalities were adopted from the modified Cornell Medical Index (CMI) (Brodman, Erdman, Lorge, Gershensen & Wolff, 1952; Nott & Cutts, 1982). The CMI was simplified to seven items: shoulder pain, leg fatigue, yawning, lumbago, eye fatigue, headache, and appetite loss. Each item was scored on four-point scales from 0–3, with the minimum and maximum total scores being 0 and 21 points, respectively.
Sociability was adopted from the modified Maudsley Personal Inventory (MPI) (Eysenck, 1959). The MPI was simplified to four items: activities in order to communicate with friends, activities in order to communicate with anybody, activities in order to talk with anybody, and having friends to ask about problems. Each item was scored on four-point scales from 0–3, with the minimum and maximum total scores being 0 and 12 points, respectively.
The obstetric factors were the frequency of delivery, experience of abortion, abnormalities during pregnancy, abnormalities during delivery, cesarian section, and abnormalities during the hospital stay. Baby care included the baby's weight at delivery, worries about baby care, and abnormalities at 1 month post-partum.
The child-rearing anxiety score was simplified to eight items: a lack of confidence or anxiety associated with child-rearing, a feeling of fear associated with child-rearing, a feeling of difficulty in raising the baby, a comfortable feeling with the baby, one's purpose in life is to nurture the baby, some problems in staying with the baby, feeling happy to have a baby, and proud of having a baby. Each item was scored on four-point scales from 0–3, with the minimum and maximum total scores being 0 and 24 points, respectively.
The cooperation of the husband was asked in four items: the cooperation of the husband with caring for the baby, a good relationship with the husband, a tense relationship with the husband, and different principles from the husband in caring for the baby. Each item was scored on four-point scales from 0–3, with the minimum and maximum total scores being 0 and 12, respectively.
In addition to those questionnaires, in the present study, the family structure was determined; that is, if it was a nuclear or extended family and, in the case of extended families, whether the new mother lived at her parents' house or husband's house, whether the new mother went back to her home or her parents' home after delivery (satogaeri), and who were the main helpers.
The study protocol adhered to the recommendations of the Declaration of Helsinki (World Medical Association, 2000). Verbal informed consent was obtained and the participants' anonymity was preserved by the use of a coding system. Ethical approval was obtained from the Ethical Committee at Akita Nursing and Welfare University, Japan.
The statistical differences were determined by the two-sided Mann-Whitney's U-test. The statistical significance of the differences among the three or four age groups was calculated using the Kruskal-Wallis test. The statistical differences in relation to the number of participants were calculated by the χ2-value for independent tests. The statistics were calculated by one-way anova for the EPDS versus the frequency of delivery and, in the case of significance with the anova, Tukey's test was used for the individual parameters. Multivariate regression was used for the EPDS versus general health abnormality, sociability, frequency of delivery, abnormality during pregnancy, abnormality during the hospital stay, worries about baby care, and cooperation of the husband. The data were expressed as the means ± SD. The differences were regarded as statistically significant at P < 0.05.
- Top of page
Those participants with an EPDS score >9 were 23% in total, and 29% for the primiparas and 16% for the multiparas. The primiparous group showed a higher frequency of participants with an EPDS score >9 compared to the multiparous group (χ2 = 3.8, P < 0.05). The average EPDS score was 6.6 ± 4.0 for the primiparas and 5.5 ± 4.5 for the multiparas. The primiparous group showed a higher EPDS score than the multiparous group (z = −2.5, P < 0.05) (Table 1).
Table 1. Physical characteristics and the Edinburgh Postnatal Depression Scale
| Primiparas||3.5 ± 2.0||7.2 ± 4.2||5.3 ± 3.1||–||6.6 ± 4.0||*†|
| N||4||72||23||–||99|| |
| Multiparas||0||5.5 ± 5.0||5.0 ± 4.0||4.5 ± 3.5||5.5 ± 4.5||NS|
| N||1||31||44||2||78|| |
| Primiparas||–||7.0 ± 4.1||‡||6.0 ± 3.8||†||NS|
| N||–||54||–||45||–|| |
| Multiparas||–||5.4 ± 3.5||–||5.1 ± 4.9||–||NS|
| N||–||50||–||28||–|| |
|General health abnormalities||–||Low||Medium||High||–|| |
|(maximum score is 21)||–||(0–5)||(6–11)||(>12)||–|| |
| Primiparas||–||2.5 ± 2.5||5.8 ± 3.1||8.0 ± 3.3||–||***|
| N||–||13||48||37||–|| |
| Multiparas||–||2.5 ± 3.4||5.6 ± 4.6||8.2 ± 5.2||–||***|
| N||–||13||34||28||–|| |
|Sociability|| ||Good||Medium||Poor||–|| |
|(maximum score is 12)||–||(0–2)||(3–5)||(>6)||–|| |
| Primiparas||–||4.6 ± 3.7||6.9 ± 3.2||11.2 ± 4.4||–||***|
| N||–||33||50||14||–|| |
| Multiparas||–||3.6 ± 3.3||6.4 ± 4.9||7.7 ± 6.5||–||*|
| N||–||32||40||3||–|| |
The general health abnormality scores of 0–5, 6–11, and >12 corresponded to low, medium, and high amounts of general health abnormalities, respectively. The general health abnormality factor was strongly associated with the EPDS score, both for the primiparas (χ2 = 19.3, P < 0.001) and multiparas (χ2 = 19.4, P < 0.001). The sociability scores of 0–2, 3–5, and >6 corresponded to good, medium, and poor sociability, respectively. Sociability was strongly associated with the EPDS scores for the primiparas (χ2 = 12.9, P < 0.001) and it was slightly associated with the EPDS scores for the multiparas (χ2 = 7.2, P < 0.05).
There were significant differences in the EPDS score among four frequencies of delivery (χ2 = 9.2, P < 0.05) (Table 2). The relationship between the EPDS score and the age and frequency of delivery is described in Figure 1. In the group of mothers aged in their twenties and thirties, there was a significant frequency dependency of delivery (z = −2.4, P < 0.05).
Table 2. Obstetric factors and the Edinburgh Postnatal Depression Scale
|Frequency of delivery||6.6 ± 4.0||5.5 ± 4.6||5.3 ± 4.1||2.2 ± 2.6||*|
| N||99||54||19||5|| |
|Experience of abortion||No||–||Yes||–|| |
| Primiparas||6.3 ± 5.3||–||7.7 ± 5.1||–||NS|
| N||80||–||19||–|| |
| Multiparas||5.3 ± 3.7||–||4.9 ± 3.4||–||NS|
| N||49||–||29||†|| |
|Abnormality during pregnancy||No||–||Yes||–|| |
| Primiparas||5.9 ± 3.8||–||7.5 ± 4.1||–||*|
| N||57||–||42||–|| |
| Multiparas||5.0 ± 5.1||–||5.4 ± 3.8||–||NS|
| N||38||–||42||–|| |
|Abnormality during delivery||No||–||Yes||–|| |
| Primiparas||6.0 ± 4.3||–||7.3 ± 4.0||–||NS|
| N||53||–||45||–|| |
| Multiparas||5.3 ± 4.6||–||4.4 ± 3.3||–||NS|
| N||67||–||11||‡|| |
|Cesarian section||No||–||Yes||–|| |
| Primiparas||6.5 ± 4.1||–||6.8 ± 2.6||–||NS|
| N||73||–||25||–|| |
| Multiparas||5.1 ± 4.6||–||5.8 ± 3.8||–||NS|
| N||70||–||8||†|| |
|Abnormality during the hospital stay||No||–||Yes||–|| |
| Primiparas||6.0 ± 4.1||–||8.0 ± 5.1||§||*|
| N||69||–||30||–|| |
| Multiparas||5.4 ± 4.5||–||4.3 ± 3.8||–||NS|
| N||67||–||11||†|| |
Figure 1. Edinburgh Postnatal Depression Scale (EPDS) versus frequency of delivery were plotted for four age ranges of the mothers. Each mark shows the number of mothers. A small number of participants <20 years old (n = 5) and >40 years old (n = 2) were not involved in the analysis of the frequency dependency of delivery on the EPDS. (○) 10–19 years; (●) 20–29 years; (▵) 30–39 years; (□) ≥40 years.
Download figure to PowerPoint
The obstetric factors, such as abnormalities during pregnancy (z = −2.5, P < 0.05) and abnormalities during the hospital stay (z = −2.3, P < 0.05), were significantly associated with the EPDS scores for the primiparas, but not for the multiparas. The number of participants who suffered abnormalities during delivery (χ2 = 18.8, P < 0.001), cesarian section (χ2 = 5.6, P < 0.05), and abnormalities during the hospital stay (χ2 = 5.5, P < 0.05) were significantly lower in the multiparous group compared to the primiparous group. The number of participants with abortion experience was higher in the multiparous group than the primiparous group (χ2 = 7.1, P < 0.05).
For baby care, the baby's weight at delivery was not associated with the EPDS scores. The scores in relation to worries about baby care were 0–5, 6–11, and >12 and corresponded to low, medium, and high levels of worries about baby care, respectively. Worries about baby care strongly associated with the EPDS scores, both for the primiparas (χ2 = 17.6, P < 0.001) and the multiparas (χ2 = 15.7, P < 0.001) (Table 3).
Table 3. Baby care and the Edinburgh Postnatal Depression Scale
|Baby weight at delivery (g)||≥2000||≥2500||≥4000||P-value|
| Primiparas||6.1 ± 3.4||6.6 ± 4.0||–||NS|
| N||7||92||0|| |
| Multiparas||3.5 ± 5.7||5.3 ± 4.4||8.0 ± 0.0||NS|
| N||5||74||2|| |
|Worries about baby care||Low||Medium||High|| |
|(maximum score is 24)||(0–5)||(6–11)||(>12)|| |
| Primiparas||4.6 ± 3.6†||7.4 ± 3.5†||10.3 ± 5.1||***|
| N||34||55||7|| |
| Multiparas||2.2 ± 2.9||5.7 ± 3.9||8.8 ± 6.0||***|
| N||19||45||11|| |
|Abnormality at 1 month post-partum||No||–||Yes|| |
| Primiparas||6.6 ± 4.1†||–||5.5 ± 1.9||NS|
| N||95||–||4|| |
| Multiparas||5.3 ± 4.5||–||3.0 ± 0||NS|
| N||75||–||2|| |
The scores regarding cooperation of the husband were 0–3 and >4 and corresponded to positive and negative cooperation of the husband, respectively. Cooperation of the husband was strongly associated with the EPDS scores, both for the primiparas (z = −4.1, P < 0.001) and for the multiparas (z = −3.9, P < 0.001), irrespective of the family structure (Table 4).
Table 4. Cooperation of the husband and the Edinburgh Postnatal Depression Scale
|Family structure||Positive score (0–3)‡||Negative score (>4)‡||P-value|
|In the total family|| || || |
| Primiparas||5.3 ± 3.9†||8.8 ± 3.2†||***|
| N||63||33|| |
| Multiparas||3.8 ± 4.4||7.1 ± 3.2||***|
| N||46||30|| |
|In the nuclear family|| || || |
| Primiparas||5.5 ± 3.8||8.6 ± 2.7||*|
| N||39||16|| |
| Multiparas||3.8 ± 4.5||7.0 ± 4.4||*|
| N||30||14|| |
|In the extended family|| || || |
| Primiparas||5.2 ± 4.3||9.0 ± 3.6||**|
| N||23||17|| |
| Multiparas||2.5 ± 3.1||7.1 ± 1.8||***|
| N||16||16|| |
The family structure showed no differences in the EPDS scores between nuclear and extended families, between living in the woman's parents' home and in the husband's home in extended families, nor among the three different main helpers of husband, mother, and mother-in-law (Table 5). The number of participants who returned to their home, without satogaeri, was significantly higher in the multiparous group (χ2 = 6.9, P < 0.05). The number of participants with a mother as the main helper was significantly lower in the multiparous group (χ2 = 1.8, P < 0.05).
Table 5. Family structure and the Edinburgh Postnatal Depression Scale
|Family structure||Nuclear|| ||Extended||P-value|
| Primiparas||6.5 ± 3.7†|| ||6.5 ± 4.4||NS|
| N||58|| ||40|| |
| Multiparas||5.2 ± 4.8|| ||5.2 ± 4.0||NS|
| N||44|| ||33|| |
|In the extended family||Living in her home|| ||Living in her husband's home|| |
| Primiparas||6.9 ± 5.5†|| ||6.2 ± 3.3||NS|
| N||12|| ||28|| |
| Multiparas||4.3 ± 2.7|| ||6.4 ± 3.6||NS|
| N||9|| ||23|| |
|At home after the delivery||At her home|| ||At her parents' home (satogaeri)|| |
| Primiparas||6.8 ± 3.5‡|| ||6.5 ± 4.5||NS|
| N||29|| ||69|| |
| Multiparas||4.7 ± 4.0|| ||5.7 ± 4.9||NS|
| N||36§|| ||37|| |
|Main helper (overlapped)||Husband||Mother||Mother-in-law|| |
| Primiparas||6.4 ± 3.8||7.1 ± 4.5†||5.6 ± 3.3||NS|
| N||82||44||38|| |
| Multiparas||5.4 ± 4.6||5.0 ± 3.6||4.5 ± 4.2||NS|
| N||64||24§||32|| |
The multivariate regression coefficients were calculated to assess the relative impact for the EPDS versus general health abnormality, of which the low, medium, and high scores corresponded to 1, 2, and 3, the sociability scores of good, medium, and poor corresponded to 1, 2, and 3, primiparas and multiparas corresponded to 1 and 2, abnormality during pregnancy of “No” and “Yes” corresponded to 1 and 2, abnormality during the hospital stay of “No” and “Yes” corresponded to 1 and 2, worries about baby care scores of low, medium, and high corresponded to 1, 2, and 3, and cooperation of the husband as “positive” and “negative” corresponded to 1 and 2, respectively. The standardized multivariate regression coefficients of general health abnormality, sociability, number of children, abnormality during pregnancy, abnormality during the hospital stay, worries about baby care, and cooperation of the husband were 0.297 (P < 0.001), 0.290 (P < 0.001), −0.191 (P < 0.01), 0.019 (not significant), 0.014 (not significant), 0.265 (P < 0.001), and 0.140 (P < 0.05), respectively (R2 = 0.451) (Table 6).
Table 6. Relative impact for the Edinburgh Postnatal Depression Scale score versus seven factors that were calculated by the size of the multivariate regression coefficients
|Factor||Standardized multivariate regression coefficients||P-value|
|General health abnormality||0.297||***|
|Number of children||−0.191||**|
|Abnormality during pregnancy||0.019||*|
|Abnormality during the hospital stay||0.014||NS|
|Worries about baby care||0.265||***|
|Cooperation of the husband||0.140||*|
- Top of page
In Western populations, a score of 12/13 on the EPDS is regarded as indicative of postnatal depression (Cox et al., 1987). Among Japanese women, a score of 8/9 has been established as a cut-off point for depression (Okano et al., 1996; Tamaki et al., 1997; Yamashita et al., 2000). The study's participants showed a relatively higher frequency of postnatal depression: 23% compared with the previously reported frequency in Japan of 17% of participants during the first 3 months postnatally (Yamashita et al.), 12% during the first 3 months postnatally (Yoshida et al., 1997), and 12% during the first 3 months postnatally, decreasing to 7% during the fourth postnatal month (Tamaki et al.). With the support of a Health and Labor Science Research Grant, a large number of postnatal depression screenings all over Japan were studied by Suzumiya, Yamashita and Yoshida (2004). A total of 3370 participants were analyzed for postnatal depression in 2002 at 4 months postnatally in 12 districts of the health center, including Aomori Prefecture. The participants with a EPDS score of ≥9 in areas of a representative large number of participants varied from 9% in Fukuoka Prefecture, in the southern part of Japan (1990 participants) to 14% in Miyagi Prefecture (245 participants), in the middle-northern part of Japan, and 21% in Aomori Prefecture (435 participants), in the northern part of Japan. The present study area, a city in Aomori Prefecture, showed similarly high EPDS scores of 21%. An international study reported different levels of post-partum depression in the USA, Europe, and Asia. The EPDS scores at 10–12 weeks post-partum ranged from 5.0 for Sweden to 10.95 for Taiwan, with an overall mean of 7.43 (Affonso et al., 2000). It has been suggested that Asia has the highest depression symptom scores. The present study population is in the northern rural area of Japan and more traditional Japanese customs are performed. However, the previous Japanese EPDS scores were analyzed in urban areas of Japan (Okano et al.; Tamaki et al.; Yamashita et al.) and in Japanese women who gave birth in England (Yoshida et al.). The EPDS scores might not be stable or homogeneous among Japanese women, although we do not know the exact reasons for such variation. In the present study, instead of cut-off EPDS scores for “depressed” or “not depressed”, continuous values of the EPDS scores were analyzed in order to find associations of the factors to postnatal depression.
The present study confirmed that the EPDS scores were higher for the primiparas than for the multiparas. The frequency dependency of the EPDS score on delivery in the twenties and thirties age groups suggests that the experience of delivery might decrease the depression state of mothers. Less obstetric complications in the multiparas might contribute to their decreased EPDS scores. Obstetrical complications have been reported to be a risk for postnatal depression (Austin, Hadzi-Pavlovic, Leader, Saint & Parker, 2005; Verdoux, Sutter, Glatigny-Dallay & Minisini, 2002; Yoshida et al., 1997). However, some reports showed no significant contribution of the obstetric complications on postnatal depression, although they did not analyze the obstetric complications in primiparas and multiparas separately (Okano et al., 1996; Tamaki et al., 1997; Yamashita et al., 2000). The exploration of the consequences of obstetric complications on the body should take into account not only the somatic impact of the complication, such as the induction of brain lesions, but also the psychological state of the mother, especially for primiparas. The number of mothers who went back to their home instead of their mother's home (i.e. satogaeri) was higher for the multiparas than for the primiparas and the number of mothers who were taken care of by their mother was lower for the multiparas than for the primiparas, which might suggest the more stable behaviors of the multiparas compared to the primiparas after delivery. The experience of delivery would be a primary determinant of the development of postnatal depression.
The present study showed that the cooperation of the husband was important in reducing the risk of postnatal depression for both the primiparas and multiparas, irrespective of the family structure. The cooperation of the husband has not been reported consistently. Some report it to be associated with postnatal depression (Horowitz & Goodman, 2004; O'Hara, 1986). However, Tamaki et al. (1997) reported that there were no significant contributions regarding the cooperation of the husband to postnatal depression. We do not know the reasons for different reports on the contribution of the cooperation of the husband. It was an unexpected finding that the extended family did not contribute to reducing the risk of postnatal depression. About half of the participants lived in an extended family, which might be a benefit in the support of mothers. On the contrary, a clear contribution of the cooperation of the husband in an extended family, rather than in a nuclear family, might suggest the difficulties that the mothers, both primiparas and multiparas, experience in an extended family, probably in order to accommodate the parents, which might counterbalance the benefit of the support of mothers by their parents. Yoshida et al. (1997) suggested that, for Japanese mothers who gave birth in England, depressed and well mothers were equally likely to have had grandmothers traveling to England to be with their daughter during childbirth. In the present study, the multiparas were cared for less by their mother and they went back to their home instead of their parents' home (satogaeri), which suggests that the experience of delivery does not lead to the need for help by the women's mother.
Yoshida et al. (1997) suggested that Japanese women tend to express emotional complaints by referring to physical problems or worries about child care rather than by expressing their feelings when they are depressed. In the present study, general health abnormalities and worries about baby care were significantly associated with the EPDS scores for the primiparas and multiparas. A depressive state was usually accompanied by general health abnormalities. Arimoto and Murashima (2007) reported that child-rearing anxiety correlated with less child-care satisfaction, more depressive symptoms, worries about the child, less support from the husband, and less social support. General health abnormalities and worries about baby care might reflect a similar basic depressive state.
A history of depression was reported to be one of the factors of postnatal depression (Horowitz & Goodman, 2004). In the present study, sociability was significantly associated with postnatal depression, both for the primiparas and the multiparas. Sociability means the kind of character of the participants and has been reported to be associated with general depression. All the women in the present study were married and did not suffer from psychiatric illness. Postnatal depression would be caused by both inherited characteristics and the circumstances of life.
As there is a high prevalence of postnatal depression in the general population, it is necessary to establish a support system for pregnant and parturient women to combat post-partum depression. It would be helpful to consult with multiparas with a large number of children to reduce anxiety about baby care for primiparas.
Social support for these women is needed to decrease the EPDS score. Depressed mothers with less sensitivity to attending to their infant and disturbances in early mother–infant interactions were found to be predictive of poorer infant cognitive outcomes (Murray, Fiori-Cowley, Hooper & Cooper, 1996). There are various proposed types of care for post-partum depression (Gjerdingen, Katon & Rich, 2008). The factors associated with post-partum depression simultaneously should be focused on and eliminated as much as possible. The present study suggests that the combined support of the husband and social support should contribute to reducing the risk of postnatal depression.