Impact of infant health problems on postnatal depression: Pilot study to evaluate a health visiting system


Dr Keiko Yoshida, MD, PhD, Department of Neuropsychiatry, Kyushu University Hospital, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. Email:


Abstract  This pilot study aimed to elucidate whether or not mothers who receive neonatal home visits tend to suffer from depressive disorders during the first postpartum year and to examine what kinds of factors were related to the postnatal depression observed in these mothers. The subjects consisted of 70 Japanese mothers who received neonatal home visits and completed the 1-year study. At their first visit, the health visitors asked about their personal backgrounds and early postpartum mood, and depression was assessed by a Japanese translation of the Edinburgh Postnatal Depression Scale (EPDS). At 12 months postpartum, a diagnostic interview using the Structured Clinical Interview for DSM-IV (SCID) was performed to confirm the maternal diagnosis by a psychiatrist. An assessment of infant development using the Denver Developmental Screening Test (DDST) was administered to each infant by a psychologist. Nineteen mothers (27%) were categorized as having had a new onset of depression (12 major depressive disorders and seven depressive disorders not otherwise specified). In comparison to the non-depressed mothers, infant-related health problems that required either outpatient treatment or hospitalization were significantly related to postnatal depression. Pediatric diseases in infants did not always precede their mother's depression. The extra burden of caring for a sick infant tends to increase the risk of developing postpartum depression. As a result, neonatal home visits by health visitors may therefore be a potentially crucial opportunity to provide emotional support, particularly for mothers with the extra burden of having to care for sick infants.


Postnatal depression (PND) is not an uncommon condition, and it is experienced by 10–15% of mothers regardless of cultural differences, including in Japan.1–4 Recent studies have reported that unrecognized and untreated depressive symptoms may result in significant psychological, social, and occupational disabilities for mothers while also placing their children at risk for developing serious developmental, behavioral, and emotional problems.5–7 When the duration of PND is protracted, its effects are intensified.8 Therefore, it is crucial to detect and treat mothers with PND during the critical early postpartum phase. Psychiatrists, however, are not in the best position,9,10 because such women often minimize their condition by ignoring or denying symptoms and not seeking help.11 Most mothers with PND are unlikely to access psychiatric care, even though their depressive symptoms are serious.12

Therefore, neonatal home visits by health visitors could be a potentially useful opportunity for detecting mothers with PND.13 However, health visitors tend to focus on the babies’ physical health rather than on any potential psychological problems that the mothers may have. At present, their routine work is to observe the mother's infant care practices and to check the infant's weight and physical condition. Such a lack of concern for the mothers’ mental health also seems to be common among many of the other countries that have health visiting systems.14

In Fukuoka city, the neonatal home visit service is provided for mother–baby dyads when (i) a baby's birthweight is <2500 g; (ii) first-born babies have a birthweight <2800 g; (iii) babies have perinatal or pediatric health problems. In addition, mothers who request a home visit, mostly due to difficulties in baby care or parenting problems, can also utilize this service. Usually at approximately 6–10 weeks after birth, health visitors visit mothers at their residence. Utilizing such neonatal home visits, the aim of the presents study was to elucidate whether or not mothers who needs health visitors’ support due to their infant care tend to suffer from postnatal depression. And if so, what kinds of factors were related to the postnatal depression observed in these mothers.



Our subjects consisted of mothers and infants who received a neonatal home visit by health visitors from one health center. The health visitors explained the aims and the process of the study sufficiently on the first home visit, then the mothers participated with written consent. Recruitment was consecutively continued until the number of mother–baby dyads reached 100.

Our final subjects consisted of 70 mothers who completed the Structured Clinical Interview for DSM-IV (SCID) interview at 12 months postnatally. Of the 100 women, we were unable to contact two of the mothers. In addition, nine had moved out of the area and could not be traced in time. One mother who suffered from a dissociative disorder was in a psychiatric ward. Two refused to be interviewed. Sixteen declined to take part because they thought they had no problems or had already returned to work. As a result, a total of 70 mothers with 73 infants (three pairs of twin babies) maintained contact with us throughout the study.


The symptomatic features of mood disturbance were measured by two diagnostic modalities.

Diagnostic interview

Each subject's past psychiatric history and present mental state were assessed by an interview using the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, Axis I, fourth edition (SCID-IV),15 a Japanese translated non-patient version. Portions of the mood episodes section (module A in the SCID) were used to determine the diagnoses of current major depression (major depressive disorder) and minor depression (depressive disorder not otherwise specified).

Self-report questionnaires

The Edinburgh Postnatal Depression Scale (EPDS)16 is a 10-item self-report scale asking about depressive symptoms during the past week. Each item is scored on a four-point scale (from 0 to 3), with the minimum and maximum total scores being 0 and 30, respectively. The EPDS has been shown to be a valid and reliable screening instrument of postpartum symptoms. Among Japanese women, a score of ≥9 has been established as the cut-off point.17

Life stresses covering the preceding 12 months before birth were measured using the self-report Life Event Scale.18

Assessment of infant development

Infant development was assessed by a version of the Denver Developmental Screening Test (DDST)19 adapted for Japanese children.20 The test covers four functions: gross motor, language, fine motor–adaptive, and personal–social. The DDST score is considered questionable when the child has a delay in one test item or fails to pass through at least one item in each sector through which his chronological age line passes. It is considered abnormal if the child fails at least two items in any one sector.


At approximately 6–10 weeks after birth, health visitors visited mothers at their residence. The mother completed the EPDS while the health visitor checked the infant's weight and physical condition. During the visit, health visitors asked mothers about sociodemographic information (age, marital status, education level, maternal and paternal employment state, and support from a partner and her own mother), obstetric history (parity, risk in pregnancy, gestational age at delivery, mode of delivery, and perinatal complications), and infant factors (sex, birthweight, health, ordinal number). The data on perinatal complications were obtained from medical records. The health visitor also asked questions about the mother's mood change after delivery and whether any mood change had occurred. Next, they asked them about its duration, and time of onset.

At 12 months postpartum, a psychiatrist, a psychologist and a health visitor attended mothers at their home. The SCID interview was performed to confirm the maternal diagnosis by a psychiatrist (MU). The past psychiatric history and personal history details were also collected during the SCID interview. Concurrently the DDST was administered to each infant by a psychologist, who was blind to the maternal demographic background and mental state. All mothers also completed the Life Event Scale.

Statistical analysis

Statistical analysis was carried out using spss 10.0J for Windows (SPSS, Chicago, IL, USA). The demographic and clinical data were compared between the mothers who became depressed and those who did not. To analyze the categorical data we used the χ2 test or Fisher's exact test.


Characteristics of subjects

The mean age of the 70 women was 29.7 years (range 22–42 years), and 83% were primiparous. The average number of children was 1.20 (range 1–3) and all but one mother were married. The average gestational age at delivery was 38.5 weeks. Most subjects (98%) had completed high school, and approximately half (43%) had a junior college or university education. Thirty-three women (46%) had been employed when they became pregnant and 24 had left work. None of the husbands had lost their job or had any serious financial difficulties either before or after the birth.

The demographic characteristics and the EPDS scores between the 70 subjects and the 30 mothers who dropped out were compared, and all the characteristics mentioned here were not significant different, except for the number of births. The women who dropped out tended to be multiparous (12/70 vs 11/30, P = 0.033).

Incidence, onset and duration of postnatal depression during the first postnatal year

Nineteen of the 70 mothers (27%) were categorized as having a new episode of depressive disorders during the first postpartum year (we define this episode as postnatal depression). The course of depression is illustrated in Fig. 1, where a horizontal line shows the onset and the duration of each episode of depression and a dotted line shows the timing of pediatric disease of their babies. Twelve mothers (subjects 1–12) had major depressive disorders and seven (subjects 13–19) had depressive disorders, not otherwise specified (minor depression). All of the disorders occurred within 3 months postpartum. In eight out of 19 cases the duration was ≤1 month, in seven it was ≤3 months, and in the remaining two it was between 5 and 6 months. There were two cases of recurrence among the mothers with major depressive disorder at 5 months, and at around 11 months postpartum. None had received any psychiatric treatment.

Figure 1.

Maternal postnatal depression and pediatric problems: timing of onset and duration. inline image, time of the first home visit by health visitors; inline image, time of the admission or operation of infant; inline image, LBW (low birthweight); inline image, duration of episode; inline image, duration of pediatric disease; (a), admission; (o), operation; (p), phototherapy.

Postpartum mood change when interviewed by health visitors

The first home visit was carried out between 15 and 109 days (an average of 51.7 days) after birth. Of the 19 mothers with depression, two mothers received the home visit before the episodes started. When the health visitor asked about the mother's mood change after delivery at the first home visit, 10 of the remaining 17 mothers reported that they had been feeling low for over 2 weeks. In addition four mothers reported that they had been feeling low for between 1 and 2 weeks. The subsequent SCID interview revealed that in all cases, their mood change lasted for a few more weeks. The remaining three mothers had not been feeling low when they were interviewed.

In contrast, none of the 51 non-depressed mothers reported that they had been feeling low for more than 2 weeks to the health visitors. Forty-three mothers did not complain of mood change. Five felt this way for <1 week and three had been feeling low for between 1 and 2 weeks.

Edinburgh Postnatal Depression Scale scores at the first home visit

In two out of 19 cases, the EPDS was administered before the onset of depressive episodes. In 10 of the remaining 17 cases, the EPDS was administered after the end of the depressive episode as is illustrated in Fig. 1. All of these women had depressive disorders that were of short duration and with an early onset. Because of the discrepancy between the timing of the EPDS and the depressive episode, eight out of 17 mothers (47%) had an EPDS score of ≥9. In five of seven cases (71%) where the visit was carried out during a depressive episode, the mother had an EPDS score of ≥9. In spite of timely home visits, two mothers were not identified as potential cases by the EPDS.

Comparisons between depressed mothers and non-depressed mothers

Demographic and psychosocial characteristics

Comparisons between the mothers who became depressed and those who did not, regarding demographic and psychosocial characteristics, are summarized in Table 1. There were no differences in the age, gestational age at delivery, education level, parity or support from their partners and their own mothers. The presence of a past psychiatric history did not differ substantially between the two groups.

Table 1. Subject characteristics
 PND mothers (n = 19) n (%)Non-PND mothers (n = 51) n (%)P
  • PND, postnatal depression.

  • Fisher's exact test, or t-test.

Age (years)30.6 ± 4.329.3 ± 4.4NS
Gestational age (weeks)38.5 ± 1.338.5 ± 1.7NS
Education level: college or university 6/19 (32)24/51 (47)NS
Parity (proportion of primipara)17/19 (89)41/51 (80)NS
Lack of support from partner12/19 (63)27/51 (53)NS
Lack of support from their own mothers 3/19 (16)4/51 (8)NS
Past history of psychiatric illness 3/19 (16) 6/51 (11)NS
Life events (excluding pediatric diseases)13/19 (68)38/51 (75)NS

Obstetric complications and pediatric disease

Obstetric complications and pediatric diseases found in depressed mothers and non-depressed mothers are shown in Table 2. There were no differences in the mode of delivery and the proportion of low-birthweight infants between the two groups.

Table 2. Obstetric difficulties and pediatric problems
 PND mothers (n = 19)Non-PND mothers (n = 51)P
  1. PND, postnatal depression.

Delivery by forceps or cesarean section 515NS
Twin births 0 4NS
Low birthweight (<2500 g) 510NS
Pediatric diseases1210<0.001
 Baby's admission due to pediatric diseases  (during 12 months after birth) 8 3<0.001

Pediatric diseases, which needed prolonged care including inpatient or outpatient treatment rather than transient symptoms such as vomiting and febrile seizure, were found in 12 infants out of 19 mothers with PND (63%). These included three with inguinal hernia, two with jaundice requiring phototherapy, two with pneumonia, one with bronchitis, one with urinary tract infection, two with atopic dermatitis or dermatitis, one with Kawasaki disease and one with cardiac anomalies. As illustrated in Fig. 1, the period of the infant's disease and their mother's depression overlapped in nine out of these 12 cases, while in seven cases the pediatric disease preceded or occurred simultaneously with the maternal depression.

In contrast, 10 infants out of 51 non-depressed mothers (20%) suffered from pediatric diseases, which included one with inguinal hernia, three with jaundice requiring phototherapy, two with pneumonia, two with dermatitis, one with hydronephrosis, one with hemangioma and one with hypothyroidism.

Pediatric diseases that required either outpatient treatment or admission were found significantly more frequently in the infants of the mothers with PND than the non-depressed mothers. The proportion of babies who needed hospitalization was much higher in depressed mothers (42%) than in non-depressed mothers (6%).

Denver Developmental Screening Test scores

The DDST scores of the infants were compared between the two groups (Table 3). Each pair of twin babies had the same score, so we looked at the scores of the twin babies as one. There were no differences and no obvious developmental delay was found in either group.

Table 3. DDST scores
DDST scorePND mothers (n = 19)Non-PND mothers (n = 51)P
  1. DDST, Denver Developmental Screening Test; PND, postnatal depression.

Abnormalinline imageinline imageNS


Higher incidence, early onset and detection of postnatal depression

In the present study the incidence of depression during the first postnatal year was 27%, which was almost twice as high as the level reported in a previous study on the incidence of PND using diagnostic interviews in Japan.4 Even when limited to major depression, the incidence was 17%, which was also relatively high.21 One of the possible reasons for this may be due to the sampling bias. The mothers who comprised the subjects of the present study were those who needed to be supported by the home visiting system. Approximately half of them delivered low-birthweight infants or had some other difficulty in terms of baby care, which are recognized as risk factors for higher EPDS scores or clinical postnatal depression.22–24

Our results showed that postnatal depression could occur early in the postnatal period. Cox et al. reported that 50% of the episodes occurred within the first 5 weeks after delivery.2 O’Hara et al. found an even earlier onset and demonstrated that half of the women developed symptoms within 1 week of delivery.1 The fact that a diagnostic interview was carried out 1 year after birth might suggest that the mothers’ recollections after this period of time were not very reliable. However, the records of the mother's mood when interviewed by the health visitors at the time of the first home visit supported the data obtained 12 months later.

Because of the early onset and relatively short duration of the cases and the late timing of the neonatal home visit when the EPDS was first performed, the EPDS scores showed lower detection rates compared to other studies in Japan, in which the EPDS was performed around the fourth postnatal week.4,17

As for the false negative cases (subjects 4 and 16 in Fig. 1), which were not identified as potential cases by the EPDS regardless of the timely home visits, we further looked at the details. Both of the mothers confirmed to the health visitors that they had been feeling low when they were asked face to face about their mood change, which was confirmed by the record obtained by the health visitors. Both of them experienced social relationship difficulties especially in relation to their own mothers and parents in law. They found it difficult to ask for support from other people in a positive way. Guedeney et al. also reported the false-negative cases and noted that postnatal depression with a predominant profile of psychomotor retardation seems not to be detected by the EPDS.25 One of the reasons for the false negatives therefore may be psychomotor retardation, which does not explain the present two cases. As regards these two subjects, those did not present with features of psychomotor retardation. Psychosocial factors such as luck of support or limited expression for asking for support are risk factors for onset of depression, but those do not explain the reason for the false-negative cases. Therefore there is certain limitation of the EPDS, especially for mothers who could not express their need for support.

Relationship between pediatric disease and maternal depression

Regarding the correlation between PND and infant-related factors, several studies have been reported: taking an infant to a primary care or an emergency department;26 premature infants;23,27 infant temperament or behavior;24,28,29 and infant sleep problems.30 We did not survey all those infant-related problems. However, from our survey, pediatric disease, especially when it required either outpatient treatment or hospitalization, seemed to be an important related factor.

Particularly, Hopkins et al. summarized these aforementioned infants as being difficult to care for.24 As for the mechanism, they explained that those infants may be an especially potent source of stress in the postpartum period, contributing to onset of postnatal depression. In the present study, pediatric diseases were found more frequently with mothers with PND than the non-PND mothers. One of our explanatory hypotheses is that caring for sick infants most likely puts an extra burden on postnatal mothers both physically and psychologically. It would not be surprising that such an infant-related extra care burden provoked an onset of PND. As mentioned here, most of the present infants’ diseases were not life-threatening ones. However, for example, the mother of the baby with an inguinal hernia (subject 12 in Fig. 1) was told by her pediatrician not to let her baby cry as much as possible. She thus became nervous and anxious to prevent her baby from crying. She thereafter carried her baby in her arms all day long and she did not sleep well. As a result, she became exhausted and depressed. After her baby's operation, she recovered spontaneously. In seven of 12 mothers with PND whose infants suffered from diseases, the infants’ disease either preceded or occurred simultaneously with maternal depression.

We also considered whether most subjects were the mothers who looked after low-birthweight infants, who might be at risk of developmental delay. When comparing between the PND cases and non-PND cases, no differences were observed in the rate of low-birthweight infants and the 1-year development assessed by DDST.

Conversely, maternal depression may alter the way in which the mother perceives the demands of infant care so that mothers tend to see their infants’ care as being more troublesome.25,31 A mother with PND might be even more upset by her infant's physical problems, especially those diagnosed in early infant life or those needing hospitalization or operation, and feel the extra child-care burden due to an ill baby even more heavily. The extra burden of caring for a sick infant can worsen or prolong the state of maternal depression. In the present PND mothers, some maternal depression occurred prior to the infants’ diseases. In two cases (subjects 4 and 13 in Fig. 1), maternal depression preceded and overlapped the infants’ disease. Those pediatric diseases were relatively severe and needed hospitalization. Naturally, those mothers were upset by their infant's disease. However, as they stayed with her sick baby in the hospital and were assisted with their babies’ care, their state of depression improved. By the time their babies were discharged, most of those mothers had recovered from depression.

Contrary to our expectations, we failed to find any relationship between psychosocial risk factors reported by previous studies and PND. Hopkins et al. found that the infants of depressed mothers had a significantly greater incidence of neonatal complications, and failed to find any relationship between general life stress factors and PND.24 Interestingly, both the Hopkins et al. subjects and the present subjects were married, and consisted of a homogeneous sample of middle-class women. They also mentioned that in middle-class, stably married women, infant-related stressors may be additional sources of stress in the neonatal period that should thus be considered in research on psychosocial factors in postnatal depression.

Optimal timing and detection of mothers with PND utilizing the neonatal home visit service

As we showed in our results, PND occurred in the early postnatal period, but most home visits by health visitors were carried out too late to be of sufficient help. We suggest that home visits should be carried out during the critical early postpartum period, preferably within the 2 postnatal weeks. However, in reality it is impossible to make such early visits due to a lack of a sufficient number of health visitors. In addition, in Japan, many mothers with newly born babies are away from their normal residence because they stay at their parents’ home to get sufficient support and physical rest for up until 1 month after delivery due to the traditional Japanese support system Satogaeri bunben.32 Even now, more than half of all pregnant women return to their parents’ home. As a result, it is difficult for a health visitor to contact such mothers during this period. This support from families or health professionals differs in various cultures and countries.

Therefore, from a practical point of view, we propose to detect postnatal depression using an opportunity of a nationwide check-up system either for screening infant developmental problems or postnatal maternal physical conditions. Because for both check-ups mothers will attend. In Japan, women typically have one visit at 1 month postpartum with their obstetricians or midwives, which focuses on maternal physical recovery from delivery and a physical check-up for the infant. We thus consider that the most practical method would be routine screening using a simple self-report questionnaire, such as the EPDS. A mother can complete a self-report questionnaire during the infant health check-up by obstetricians. Then obstetricians can feed information about mothers’ depressive symptoms to health visitors by means of the EPDS. Subsequently, mothers who have high EPDS scores and/or some problems regarding their infant's health can thus be shifted to health visiting care and afterward the health visitors would then pay as much attention to the mothers to observe and monitor maternal mental condition as their infants, utilizing the conventional neonatal home visits. Whenever babies have been confirmed to have pediatric health problems, it is prudent to check the mothers regarding mood changes for as long as the extra child-care burden has been continued.

Limitations and clinical implications

The most obvious limitation of the present study is that we could not prove that the infant health problems were the etiology or the direct trigger of onset of PND. Infant care problems, those causing adding an extra burden to postnatal mothers, comprised not only the pediatric diseases but also several other factors, for example low birthweight, occurrence of twins, infant behavior, and infant sleep problems. Babies had some of these factors at one time and no distinction was made between infant physical problems and those other factors in the babies. However, even within the present small sample, the results indicate that it is useful to be aware of the higher risk of onset of PND among the mothers with infant problems. The clinical implication therefore is to advise health visitors or midwives who carry out postnatal visits to be aware and sensitive of a mother's mental health, which could be at risk of possible PND due to her infant's physical health problems.

Children's health problems are of course the primary concern for mothers, but we found that mothers may suffer from their own mental health problems. Whitton and his colleagues interviewed women who were diagnosed as having suffered from postnatal depression, about their symptoms, help-seeking behaviour and treatment. They reported that over 90% of women recognized that there was something “wrong”, but only one-third of them believed that they were suffering from postnatal depression. Over 80% had not reported their symptoms to any health professional.33 Taking this important information into consideration, neonatal home visits should be utilized to the full extent. It is important not to overlook the possibility of PND. To detect maternal depression, routine screening by means of the EPDS is useful. However, as we described, there are also false-negative cases in which the mother needs emotional support but cannot express it. So, it is prudent not only to perform the EPDS but also to ask mothers personally about their mood change, via health visitors.


This project was supported by the Japanese Ministry of Health and Welfare (chief organizer: Professor Hitoo Nakano, Department of Gynecology and Obstetrics, Graduate School of Medical Science, Kyushu University, Japan). We thank the postnatal mothers who participated in this study; Dr Hiroko Suzumiya, the head of the community health center in Fukuoka City; and the ex-chief midwife Michiko Makimoto and the midwives and health visitors for their cooperation. We also thank Professor Emeritus Nobutada Tashiro and Professor Shigenobu Kanba for valuable advice throughout our work.