Authors' contributions: Study Design,
Low level of parental bonding might be a risk factor among women with prolonged depression: A preliminary investigation
Article first published online: 23 SEP 2009
© 2009 The Authors. Journal compilation © 2009 Japanese Society of Psychiatry and Neurology
Psychiatry and Clinical Neurosciences
Volume 63, Issue 6, pages 721–729, December 2009
How to Cite
Handa, H., Ito, A., Tsuda, H., Ohsawa, I. and Ogawa, T. (2009), Low level of parental bonding might be a risk factor among women with prolonged depression: A preliminary investigation. Psychiatry and Clinical Neurosciences, 63: 721–729. doi: 10.1111/j.1440-1819.2009.02018.x
- Issue published online: 19 NOV 2009
- Article first published online: 23 SEP 2009
- Received 5 February 2008, revised 24 May 2009, accepted 6 July 2009.
- parental bonding instrument;
- risk factors;
- sex difference
Aims: The aim of the present study was to determine whether or not the effect of parenting by the father and mother on outcomes for depression may be different between male and female subjects.
Methods: A total of 115 patients were involved in this investigation: 74 had states of depression that continued for more than 2 years, and 41 had symptoms that remitted within 4 months. The Parental Bonding Instrument (PBI) was used to test for gender differences in the PBI score, the level of education, and the age at which the depression began, using an unpaired t-test.
Results: It is suggested that female patients with low paternal care and low levels of education have a higher likelihood of showing symptoms of prolonged depression in a primary episode. No relationship was found among prolongation of depression, educational level, and parental care in male patients. Furthermore, comparing the PBI quadrants established by Parker showed that female patients who were exposed to paternal care as ‘Affectionless Control’, had a tendency towards a higher risk of prolonged depression than female patients who received ‘Optimal Parenting’.
Conclusion: Especially in female patients, the prolongation of depression is likely a result of low levels of paternal care and low education.
IN RECENT YEARS, depression has been considered to be more widespread than other mental disorders. Unipolar major depression is the single most common psychiatric disorder in the US adult population, with a lifetime prevalence of 16.6%.1 Similarly prolonged depression is a common disorder seen in clinical practice. The chronic course of major depressive disorder (MDD) has received attention, particularly in the past 15 years.2,3 Eaton et al. noted that MDD were unremitting in 15% of the cases in their study.4
Some risk factors of prolongation include: young5 or old age,6,7 gender (specifically being female),4,8 family history of depressive illness,8,9 low family income,10 less education,11 a psychiatric condition,10,12 comorbidity with anxiety disorders,13 neuroticism14,15 and concurrent general illness.8,16 There have also been reports indicating that a lack of social support has negative effects on recovering patients.17,18 Others report that sex difference in responses to antidepressant treatment response,19 sex factor of being female, social isolation, and psychosocial factors in childhood also affect the prognosis in depressive patients.20
On the other hand, there have always been studies on the psychosocial environment in childhood as a risk factor of depression. Tennant found that the most important factor was the quality of the subject's home environment during childhood.21 Kendler et al.22 suggest different risk factors for female depressive patients at various ages. In addition, they suggest that childhood-related factors include a disturbed family environment, genetic factors, sexual abuse, and loss of a parent.22
Most studies focusing on parental rearing methods as a risk factor using the child-rearing measurement, including PBI, conclude that inadequate child rearing methods are a risk factor of depression.23–28 One study, involving PBI, compared melancholic depressive patients and non-melancholic depressive patients. The melancholic group did not show much difference when compared with control, but the non-melancholic group showed a lower care score and a higher protection score.29 By comparison, another study found not much difference between control and patients with bipolar disorders30,31 or endogenous depression.32
In day-to-day clinical practice, we treat a variety of depressive patients. While some recover within a short period without any major conflicts, others show prolonged symptoms of depression and experience difficulty in overcoming it. Our clinical experience often leaves the strong impression that especially women in the latter category are likely to show discontent with their parents' child-rearing methods and attitudes during childhood, compared to their male counterparts. This empirical impression led us to address the following issue: the environmental factor of parenting is not only a risk factor of initiating depression but also of its prolongation. We further considered if there were any changes in the results between different sexes. To our knowledge, no study so far has shed light on male–female differences within this context. Therefore, we investigated whether or not the effect of parenting by the father and mother on outcomes for depression may be different between male and female subjects using the PBI.
From 1 April 2006 to 30 June 2006, using the Structured Interview Guide for the 21-item Japanese version Hamilton Rating Scale (HRSD),33 we evaluated depressive patients (aged 20–65) at two psychiatric clinics in two Japanese metropolitan cities. Further, the Structured Clinical Interview for DSM-IV (SCID-I; First et al. 1996)34 was administered to patients who scored 16 points or higher on the HRSD. Patients diagnosed with MDD and dysthymia (Dys) by SCID-I were included in the study. We also added minor depressive disorder (MinD) in reference to DSM-IV-TR.35 Initially, we discarded patients with histories of organic brain syndrome, manic episodes, severe anxiety disorders, psychotic features, severe physical disease, and those whose first-degree relatives had affective disorders. Also excluded were patients who refused to give consent (four cases of prolonged depressive patients, who likely believed that their privacy had been compromised), those whose symptoms may worsen as a result of the test (two cases of prolonged depressive patients), and those who could not be tested using the PBI as a result of a parent(s)' death or divorce during childhood (seven cases of prolonged depression). Due to the risk that recurrences may be a factor in prolongation, prolonged depressive patients who had recurrences were excluded, and we included 12 non-prolonged depressive patients whose depression recurred 6 months (or more) after they ended their previous treatment in the early remitted group. A total of 162 patients met the criteria for initial inclusion. PBI was applied with these patients, and we grouped them into the following categories: (i) prolonged depressive patients whose depression is a primary episode, did not show any signs of remission, and scored 16 points or higher on the HRSD more than 2 years after the first examination; and (ii) early remitted patients who scored 16 points or higher in the first examination or the first after a recurrence, but decreased to 7 points or less within 4 months and showed no sign of recurrence 6 months after the first examination or the first after a recurrence. Patients were excluded who were unable to tolerate an ordinary dose of antidepressant treatment for a minimum of 20 weeks, or who showed no response, and were then unable to take a full dose of another antidepressant for more than 8 weeks. At the end, the target group consisted of 74 prolonged depression cases (27 men, 47 women) and 41 early remitted cases (19 men, 22 women).
This study was performed in accordance with the Declaration of Helsinki. The study protocol was approved by the Institutional Review Board of the Department of Psychopathology and Psychotherapy, Graduate School of Medicine, Nagoya University. Written informed consent was obtained from all subjects.
Subjects were requested to complete the PBI.36 The PBI and the Egna Minnen Betraffande Uppfostran (EMBU)37 are reliable Japanese versions of child-rearing measurements. The EMBU consists of 81 questions and was likely to become a burden on patients; so the PBI, which is comparatively less of an imposition, was chosen.
The PBI measures parental attitudes toward child rearing using a self-report questionnaire of 25 items, developed by Parker et al. in 1979.38 The PBI test measures the child rearing that the responder received from his/her parents before the age of 16 by having him/her recall his/her memories. The questionnaire has two subscales – Care and Overprotection. The responder can score between 0 and 3 points for each question. The results measure the child rearing attitudes of the responder's father and mother separately. The care measurement is comprised of 13 questions, and the range of scores is 0–36 points. Higher scores point to parental attitudes toward their child as open and deeply caring while lower scores illustrate a lack of interest or denial. The overprotection section also consists of 13 questions with a score range of 0–39 points. Higher scores characterize parents as overprotective and controlling, while lower scores show parents' respect for the child's independence. PBI has been proven to be valid and reliable by Parker. Answers to the PBI show similar results to the evaluation of parents and siblings by a third party.38,39 The PBI indicates that the test not only provides the patient's perspective based on recollection of the past, but is highly likely to be representative of the parents' actual child-rearing methods.38,39 It also shows that results taken from the PBI are less likely to be affected by a patient's exaggeration as a result of depression.26,40,41 Furthermore, in the study conducted by Lizardi et al. the PBI was tested on 152 outpatients, and then retested three times.42 The time period between the initial testing and the retesting was 90 months. Results showed that despite long time intervals and changes in patients' phases of depression, the PBI test results remained fairly stable throughout the study. Kitamura and Suzuki validated the Japanese version in 1993.36 The original reliability of the PBI Japanese version was r = 0.76 (score of care), and r = 0.63 (overprotection), while split-half reliability was 0.88 (care score) and 0.74 (overprotection). This Japanese version is used in the present study.
The PBI quadrants,23 using the established cut-off score, are created from two factors: care and overprotection. These factors then compose the x- and y-axes in the quadrant. Figure 1 shows the four quadrants: Optimal parenting (I), Affectionate constraint (II), Absent parenting (III), and Affectionless Control (IV). For mothers, the dichotomizing care score is 27.0 and the overprotection score is 13.5; for fathers, the dichotomizing care score is 24.0 and the overprotection score is 12.5.30 For example, for mothers, patients belonging to ‘Absent parenting’ have a care score of less than 27.0 points, and an overprotection score of less than 13.5 points.
The subjects' age, age of onset, the years of education and the PBI score were compared between men and women using an unpaired t-test. We statistically analyzed the distribution into each PBI quadrant of early remitted patients and patients with prolonged depression with the χ2-test. Values were expressed as means ± SD or means ± SE (medians are in parentheses). All statistical analyses were conducted using the spss ver. 14.0J software package (spss Japan Inc., Tokyo).
Characteristics of subjects
The demographic and clinical characteristics are shown in Tables 1 and 2. No significant results were observed when the ages between men and women were compared: 42.8 ± 0.6 vs 40.4 ± 11.5, P = 0.269, or when the age of onset was compared between men and women: 39.5 ± 10.2 versus 41.4 ± 12.1, P = 0.369.
|Demographic and clinical characteristics||Early remitted patients||Patients with prolonged depression||P|
|Item||All subjects (n = 19)||MDD (n = 13) 68.42%||MinD (n = 6) 31.58%||All subjects (n = 27)||MDD (n = 17) 62.96%||Dys (n = 10) 37.04|
|Age, mean (SD), years||39.4 (8.8)||42.0 (8.7)||33.8 (6.4)||45.1 (11.3)||48.2 (11.1)||39.8 (10.0)||0.073|
|Age at onset, mean (SD), years||39.0 (8.8)||41.8 (8.6)||32.8 (6.2)||39.8 (11.2)||42.7 (11.3)||35.0 (9.8)||0.780|
|Education, mean (SD), years||15.2 (1.8)||14.8 (1.9)||16.0 (1.3)||14.4 (2.5)||14.1 (2.4)||15.1 (2.6)||0.292|
|Employment status, n (%)|
|Regular occupation||17 (89.5)||12 (92.3)||5 (83.3)||23 (85.2)||14 (82.4)||9 (90.0)||ND|
|Part-time||2 (10.5)||1 (7.7)||1 (16.7)||2 (7.4)||1 (5.9)||1 (10.0)||ND|
|Housewife||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||ND|
|Student||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||ND|
|Unemployed||0 (0.0)||0 (0.0)||0 (0.0)||2 (7.4)||2 (11.8)||0 (0.0)||ND|
|Marital status, n (%)|
|Currently married or cohabiting||11 (57.9)||9 (69.2)||2 (33.3)||16 (59.3)||13 (76.5)||3 (30.0)||ND|
|Not married||7 (36.8)||4 (30.8)||3 (50.0)||9 (33.3)||3 (17.7)||6 (60.0)||ND|
|Widowed||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||ND|
|Divorced||1 (5.3)||0 (0.0)||1 (16.7)||2 (7.4)||1 (5.9)||1 (10.0)||ND|
|No cohabitant, n (%)||4 (21.1)||1 (7.7)||3 (50.0)||6 (22.2)||3 (17.7)||3 (30.0)||ND|
|Without children, n (%)||7 (36.8)||4 (30.8)||3 (50.0)||12 (44.4)||4 (23.5)||8 (80.0)||ND|
|HAM-D, mean (SD)||19.4 (2.7)||19.4 (3.0)||19.3 (2.1)||18.7 (2.3)||19.5 (2.4)||17.4 (1.4)||ND|
|Demographic and clinical characteristics||Early remitted patients||Patients with prolonged depression||P|
|Item||All subjects (n = 22)||MDD (n = 19) 86.36%||MinD (n = 3) 13.64%||All subjects (n = 47)||MDD (n = 20) 42.55%||Dys (n = 27) 57.45%|
|Age, mean (SD), years||40.4 (12.0)||40.1 (12.2)||42.3 (12.6)||40.4 (11.3)||40.3 (10.7)||40.5 (12.9)||0.983|
|Age at onset, mean (SD), years||40.8 (12.3)||40.2 (12.3)||44.3 (14.5)||41.7 (12.1)||43.9 (11.5)||40.1 (13.3)||0.768|
|Education, mean (SD), years||14.8 (1.8)||14.6 (1.9)||16.0 (0.0)||12.9 (2.1)||12.7 (2.1)||13.1 (3.1)||0.001|
|Employment status, n (%)|
|Regular occupation||10 (45.5)||9 (47.4)||1 (33.3)||17 (36.2)||10 (50.0)||7 (25.9)||ND|
|Part-time||1 (4.6)||1 (5.3)||0 (0.0)||7 (14.9)||2 (10.0)||5 (18.5)||ND|
|Housewife||10 (45.5)||8 (42.1)||2 (66.7)||15 (31.9)||6 (30.0)||9 (33.3)||ND|
|Student||1 (4.6)||1 (5.3)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||ND|
|Unemployed||0 (0.0)||0 (0.0)||0 (0.0)||8 (17.0)||2 (10.0)||6 (22.2)||ND|
|Marital status, n (%)|
|Currently married or cohabiting||13 (59.1)||11 (57.9)||2 (66.7)||25 (53.2)||10 (50.0)||15 (55.6)||ND|
|Not married||8 (36.4)||7 (36.8)||1 (33.3)||10 (21.3)||2 (10.0)||8 (29.6)||ND|
|Widowed||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||0 (0.0)||ND|
|Divorced||1 (4.6)||1 (5.3)||0 (0.0)||12 (25.5)||8 (40.0)||4 (14.8)||ND|
|No cohabitant, n (%)||5 (22.7)||4 (21.1)||1 (33.3)||9 (19.2)||5 (25.0)||4 (14.8)||ND|
|Without children, n (%)||9 (40.9)||8 (42.1)||1 (33.3)||22 (46.8)||9 (45.0)||13 (48.2)||ND|
|HAM-D, mean (SD)||20.3 (21.0)||20.5 (2.9)||19.0 (3.6)||19.7 (3.4)||20.6 (2.9)||19.1 (4.6)||ND|
As Tables 1 and 2 show, there were no significant differences in ages and years of education between male early remitted and prolonged depressive patients. On the other hand, comparison between female early remitted and prolonged subjects revealed no significant differences between ages of the patients and their ages of onset. However, the differences were significant given the years of education: 14.8 ± 1.8 versus 12.9 ± 2.1, P = 0.001.
Comparison of PBI testing results between early remission and prolongation in male and female patients
Comparison of PBI scores between early remitted patients and patients with prolonged depression are shown in Table 3.
|Clinical characteristics||Early remitted patients||Patients with prolonged depression||P|
|Sex||Item||n||Mean ± SE||n||Mean ± SE|
|Male||Paternal care||19||17.2 ± 1.6||27||20.6 ± 1.7||0.176|
|Maternal care||19||26.0 ± 1.3||27||25.2 ± 1.5||0.699|
|Paternal overprotection||19||11.6 ± 2.0||27||12.2 ± 1.5||0.790|
|Maternal overprotection||19||13.2 ± 1.8||27||12.5 ± 1.5||0.765|
|Female||Paternal care||22||24.5 ± 1.7||47||15.8 ± 1.4||0.001|
|Maternal care||22||23.5 ± 2.0||47||19.8 ± 1.3||0.126|
|Paternal overprotection||22||11.2 ± 1.7||47||13.6 ± 1.2||0.255|
|Maternal overprotection||22||12.8 ± 2.0||47||14.3 ± 1.1||0.482|
For male subjects, the PBI scores [paternal care (PC), maternal care (MC), paternal overprotection (PO), maternal overprotection (MO)] revealed no significant differences between early remitted and prolonged patients.
For female subjects, statistical significance was proven for PC (P = 0.001).
|Parent||Group||High care–low overprotection||High care–high overprotection||Low care–low overprotection||Low care–high overprotection||Total||P|
|Sex of patient||‘Optimal parenting’||‘Affectionate constraint’||‘Absent parenting’||‘Affectionless control’|
|n (%)||n (%)||n (%)||n (%)||n|
|Male||Early remitted patients||3 (15.8)||0 (0.0)||10 (52.6)||6 (31.6)||19||0.199|
|Patients with prolonged depression||10 (37.0)||1 (3.7)||7 (25.9)||9 (33.3)||27|
|Female||Early remitted patients||10 (45.5)||3 (13.6)||4 (18.2)||5 (22.7)||22||0.036|
|Patients with prolonged depression||7 (14.9)||5 (10.6)||13 (27.7)||22 (46.8)||47|
|Male||Early remitted patients||7 (36.8)||1 (5.3)||4 (21.1)||7 (36.8)||19||0.930|
|Patients with prolonged depression||12 (44.4)||2 (7.4)||5 (18.5)||8 (29.6)||27|
|Female||Early remitted patients||10 (45.5)||2 (9.1)||2 (9.1)||8 (36.4)||22||0.249|
|Patients with prolonged depression||12 (25.5)||2 (4.3)||9 (19.1)||24 (51.1)||47|
As the results show, paternal Low Care–High Overprotection compared to High Care–Low Overprotection is likely to involve a higher risk of prolongation in female patients. (χ2-test, P = 0.036)
The present study focused on parental child-rearing attitudes using the PBI to compare the influence on male and female patients suffering from prolonged depression lasting over two years with those with an early remitted depression that lasted less than 4 months. The study results suggest that: (i) low levels of paternal care during childhood and low education (measured in years of education received) may become a risk factor for prolonged depression in female patients; (ii) women who received paternal ‘Affectionless control’ during childhood would seem to have a higher probability of facing prolongation than those who received paternal ‘Optimal bonding’; and (iii) for men, none of the factors showed any statistical significance.
Examples of studies that focus on the risk factors for depression comparing men and women that used the PBI are those by Enns and Parker, who used clinical patients as test subjects, and that by Sato et al. for non-clinical test subjects. Enns et al.43 concluded that in men, overprotection by their fathers was significantly associated with depression, while in women, lack of care by their mothers was significantly associated with depression. In contrast, Parker23 suggests in his study that depressive patients are most likely to report anomalous parenting characteristics by the subject's same-sex parent. On the other hand, the study by Sato et al.27 of 418 non-clinical subjects using the PBI showed that parental care, rather than parental overprotection, was important in predicting lifetime depression in both male and female subjects. Their study indicates that a significant risk factor for lifetime depression in male respondents was an interactive combination of low care and high overprotection (‘Affectionless control’). For female respondents, paternal child-rearing behaviors were suggested as key risk factors.
In contrast, few studies using PBI can be found on the development and prognosis of depression, and none compared the differences between men and women as we did in the present study. Sakado et al.44 explored the relationship between PBI sores and 4-month outcomes after treatment with antidepressants in 60 university hospital patients (DSM-III-R with a total of 60 MDD male and female patients). Results suggested that non-remitted patients had lower care scores than remitted patients, hence, they concluded that low paternal care is a significant risk factor for a poor response to treatment with adequate antidepressants. In our study, (i) prolonged depression (for more than 2 years) was compared to that of early remitted patients (within 4 months); (ii) the subjects were not from a university hospital but from a private practice with no inpatient facilities and showed fairly mild symptoms of depression; and (iii) our study involved MinD and Dys cases in addition to MDD cases, in contrast to the study by Sakado et al. Nevertheless, both Sakado et al. and our own studies show similar results: patients with low scores of paternal care showed a worse prognosis in comparison to patients with higher scores.
The main difference between our study and previous research is that we focused on sex differences in the relationship between prolongation and child-rearing methods. Our data results showed that women are highly likely to be affected by such factors, whereas men are not. The results indicated similarities with our clinical experiences. Moreover, there is a possibility that in women, good relationship skills based on optimal child-rearing situations are a major factor in early remission from depression.
Many studies show that child rearing characterized by ‘Affectionless control’ (using the PBI measure) is highly likely to be correlated with onset of depression (Parker 1983; Parker et al. 1987; Plantes et al. 1988; Sato 1997),23,26,27,32 but our study shows that it may be correlated with prolongation as well in female subjects.
Nevertheless, for women, it is likely that child-rearing issues have a high correlation with fast recovery from depressive states. It can be inferred that prolongation is a result of a series of life phases where suboptimal objective relationships during childhood later lead to difficulty in receiving social support. Specifically, we observed that paternal and not maternal care seems to impact the prolongation of female patients with depression.
Our results also suggest that fewer years of education lead to a higher risk of prolongation among women. In women with low educational levels, the lengthened time period between the onset of depression and having a medical examination due to a lack of medical knowledge, less compliance in using medication, and suboptimal economic conditions may be factors leading to prolongation. However, the present results show no correlation between men and low education levels, which leads to the conclusion that there is a third factor mediating the two variables.
- 1The present study had a small number of subjects (46 men, 69 women, total of 115 patients).
- 2Subjects were outpatients with fairly mild depression compared to inpatients diagnosed with depression.
- 3Patients' parental child-rearing attitudes were based on the patients' recollection of their childhood memories.
- 4Our study did not focus on the prospective development of prolongation.
- 5The present study used SCID to compare sex differences in all types of depressive outpatients excluding adjustment disorders. The results were therefore not controlled for operational diagnosis of mood disorder subgroups. The sample used in our study thus represents the general population of clinical depression in outpatient clinics.
Women who received low levels of paternal care and education are likely to have a higher risk of prolongation from the primary depressive episode. In contrast, male subjects did not show any correlation between prolongation of depressive states, child rearing, and years of education. Female subjects who received paternal parenting characterized by ‘Affectionless Control,’ a measure of PBI, were likely to show a higher risk of prolongation compared to those who received ‘Optimal Parenting.’
The authors acknowledge Drs T. Furuhashi and K. Matsuyama for their support in performing this study and providing useful discussion. The authors are grateful to Prof Dr N. Ozaki for helpful advice in the course of our study.
- 33A structured interview guide for the Hamilton Depressive Rating Scale Japanese version. Rinsyo Seishin Yakuri 2003; 6: 1353–1368 (in Japanese)., .
- 34Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). User's Guide. Biometrics Research Department. New York State Psychiatric Institute, NY, 1996 (Japanese Version)., , , .
- 35American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th edn. Text revision. APA, Washington, DC, 2000.
- 37Development of a Japanese form of EMBU Scale and study on its reliability. Seishin Igaku 1996; 38: 1065–1072 (in Japanese)., , .
- 39The parental bonding instrument: Psychometric properties reviewed. Psychiatr. Dev. 1989; 4: 317–335..
- 40The Parental Bonding Instrument: A measure of perceived or actual parental behavior? Acta Psychiatr. Scand. 1994; 82: 153–159., , .