The relationship between postpartum depression and lower maternal confidence in mothers with a history of depression during pregnancy

Abstract

Background:

The objective of the present study is to evaluate the association of postpartum depression and low maternal confidence in a sample of women who had depression during pregnancy.

Methods:

Cross-sectional study performed from 2013 to 2015 with 346 postpartum women who had participated in an intervention to treat their depression during pregnancy. This study used the Maternal Confidence Questionnaire and the Patient Health Questionnaire 9-item scale. The prevalence ratio, adjusted and non-adjusted, and the 95% CI were calculated using Poisson regression with robust variance. Multivariate models estimated the Prevalence Ratios between postpartum depression and low maternal confidence adjusted for socio-demographic variables and maternal characteristics. Statistical analysis was performed with the STATA12.

Results:

Among a sample of women who were depressed during pregnancy, only 19% had probably moderate to severe depression and nearly half, 48%, reported high maternal confidence in the postpartum period. In the fully adjusted model, women with moderate/severe probable depression showed increased risk of lower maternal confidence in comparison to women without probable depression Prevalence Ratio=1.37 (95% CI 1.10–1.71).

Conclusions:

The results reinforce the importance of the evaluation of maternal confidence feelings in primary care particularly for women with more severe forms of depression.

Background

Maternal mental health is an important factor for postpartum health and infant health in the first year after giving birth. Maternal confidence might be a protective factor for women and infants. Maternal self-confidence is defined as the mother’s perception of her own ability to take care of the child and to correctly interpret the child’s signals¹. Many factors determine a mother’s feelings of self-confidence, including contextual characteristics such as social support, infant temperament, and maternal mental health^2–4. In theory, the nature of the experiences with caregivers, especially mothers, during childhood, influences the cognitive, emotional and social development of the child^5–8. Thus, the evaluation of maternal confidence level is critical to understand the attachment between mother and child, especially among depressed women⁹. At the same time, studies show that individuals with depressive symptoms or depression tend to have a lower level of confidence in their own abilities, including in the relationship with their offspring¹⁰.

Maternal depressive symptoms have considerable negative impact on the mother’s ability to hold onto and care for her baby. Maternal depression has been considered the most important predictor of difficulties in the establishment of a bond between mother and child¹¹. Proper establishment of a bond or attachment between mother and infant from the first days after delivery is essential for the cognitive and emotional development of the child and for the mother’s own well-being^11–13. Studies indicate an association between maternal mood disorders and impaired cognitive development in children^12,14. The effects of maternal depression are not limited to a delay in early-life development. It can also cause changes in mother-child interaction in early childhood and harm language development in the long term, both cognitive and behavioral, which can be minimized by a process of early detection of risks to child development¹⁴. Depressed mothers are insecure of their capabilities and maternal performance, which can lead to a lessening of attention towards the child and to gaze aversion characteristic of micro-rejection from mother to infant^14,15.

More recently, primary care clinics have started to screen for depression given the tremendous risks for maternal and infant adverse outcomes. The prevalence of postpartum depression (PPD) in high income countries is approximately 13% of adult women^16,17. The majority of mental health services research in primary care comes from high-income countries, but the number of investigations is increasing globally. In low and middle income countries (LMICs) the prevalence rates for PPD in primary care range from 13% in Kenya¹⁸, to 15.9% in Brazil¹⁹ and as high as 30% in Nepal²⁰. Despite the high prevalence of PPD in primary care settings in LMICs, few investigations have examined depressive symptoms as a potential impairment to maternal confidence.

Research studies on the development of and changes in parental confidence during the postpartum period indicate increased maternal self-confidence increases during the first week postpartum. In contrast, studies with women who had received a PPD diagnosis did not show an increase in maternal self-confidence during the same period^21–23. Reck and colleagues demonstrated a negative impact of depressive episodes after the postpartum period on maternal confidence²⁴. The worse the emotional state of postpartum women, the greater the chances of developing low maternal confidence (LMC)^23,25. The aim of this study is to evaluate the relationship between postpartum depression and lower maternal confidence among women with a history of depression during pregnancy.

Methods

This is a cross-sectional study of 346 low-income postpartum women, who had depression during pregnancy, and gave birth in a public health clinic. Women were recruited for the current study, PROAPEGO, from a previous clinical trial, PROGRAVIDA. Accordingly, all study participants originally took part in the PROGRAVIDA study, a cluster randomized trial that aimed to evaluate the impact of a Depression Management Program (PMD) on pregnant women with symptoms of depression. The methodology for PROGRAVIDA has been described elsewhere in further detail²⁶. The postpartum women included in PROAPEGO were 18 years of age or older and had had their prenatal care at a primary care public health clinics. The study excluded postpartum women whose children were over 9 months old and those who could not be reached for the postpartum evaluation. Of the 506 eligible postpartum women, 154 could not be reached either by telephone contact or personal contact (home visit). Thus, 352 (69.5%) women were interviewed, but six women were subsequently excluded due to missing data either in the main outcome variable or in the exposure variable.

All women were interviewed at home by a previously trained field researcher between 6 and 9 months after giving birth. During the in-home interviews, the participants signed a consent form and answered all questionnaires and instruments. The interviews were conducted between 6 and 9 months after delivery. This study was performed with the assistance of the Family Health Association, at municipality of São Paulo. The study procedures were approved by the Institutional Review Board at the University of São Paulo.

Measures

The main outcome variable is maternal confidence. The Portuguese version of Maternal Confidence Questionnaire (MCQ) consists of thirteen items with a Likert response scale of 5 points (from never to always). It presents a version in Portuguese developed by³, who found good internal consistency (Cronbach’s alpha of .88) and good construct validity, both convergent and discriminant. The total score was divided into two groups (higher or lower) split at the 50% percentile, which was used to classify women with Lower Maternal Confidence (LMC). This questionnaire has been used in several previous studies^1,27,28.

The main exposure variable is postpartum depression (PPD). The Patient Health Questionnaire 9-item scale (PHQ-9) is a specific instrument for depression that enables diagnosis according to the DSM-IV criteria for depressive disorder. The presence and intensity of each of the nine items in the two weeks preceding the interview are investigated. The scores range from 0 to 27. Scores of 10 or more are considered cases of major depression, while scores 5–9 represent cases of mild depression, in accordance with the original instrument validation²⁹. The PHQ-9 was recently evaluated in Brazil and presented a great capacity to discriminate between cases and non-cases of depression³⁰. This study makes use of the Brazilian version of the PHQ-9, translated from English into Portuguese by Pfizer Inc. (educational services). We used three categories of PPD: no depression (totals score below 5), mild depression and moderate/severe depression

Control variables

Covariates were identified a priori based on previous research on PPD and maternal confidence. A questionnaire was used to obtain socio-demographic and socioeconomic information from the participants (age, self-reported skin colour, education, occupation, personal income, family income, marital status). The questionnaire also includes information on the number of children at home, planning of previous pregnancy and randomization group (control or intervention) in the PROGRAVIDA study.

Statistical analysis

All variables were categorized. The prevalence ratio (PR) of 95% was used to evaluate the association between LMC and PPD controlling for covariates. We used the chi-square test or the chi-square test for linear trend when the categories are ordered. The PR, adjusted and unadjusted, with 95% CI was estimated using Poisson regression with robust variance. We examined the effects of PPD on LMC accounting for potential confounders by using three different models: (1) model 1 adjusted for the randomization in the cluster study; (2) adjusting for model 1 plus socioeconomic variables (schooling, family monthly income in tertiles, self-reported skin colour, marriage status); (3) adjusting for model 2 plus maternal characteristics (mother’s age, number of children, planning of the previous pregnancy). The adjustment performed through three different models aimed to control for distal (socioeconomic) and proximal (maternal characteristics) factors associated with the main outcome variable (LMC), considering a possible hierarchic model. In order to be kept in the adjusted model the variable had to show a 5% level of significance. Statistical analysis was performed using the STATA 12 software.

Results

Table 1 reports the sample characteristics. In this sample 83.1% had a partner (married or living together), 163 (47.1%) were White and 241 (70%) had at least 9 years of education. A hundred and fourteen (33%) had a monthly family income of less than USD 396.68. Mean maternal age was 27.5 years ranging from 18.9 to 46.7 years of age. Two hundred and one (58.1%) women were between 20 and 30 years old. Almost one third of the women had three or more children at home. Only 98 (28.4%) women classified their previous pregnancy as planned. Taking into account the randomization group during pregnancy in the cluster study, there was no difference between participants and non-participants in to the following variables: PPD, self-reported skin color, number of children, and marital status. Regarding the educational level variable, there was a statistically significant association among the women in the intervention group: participants had a higher level of education compared to non-participants (p =0.001). Regarding PPD, 199 (57.6%) women were classified as non-depressed while 81 (23.4%) were classified as mildly depressed. Sixty-six (19.1%) postpartum women had moderate/severe depression. In the bivariate analysis, being single and having moderate/severe depression were associated with LMC. On the other hand, having three children was associated with higher maternal confidence (Table 1).

Table 1:

Bivariate analysis of LMC and explanatory variables (number and percentage of LMC cases, Prevalence Ratio, Confidence Interval of 95% and significance level)

	High Maternal Confidence	Low Maternal Confidence	PR	CI 95%
Explanatory variable	N= (%)	N= (%)			P value
Marital Status					0.05
Married	144 (50.2%)	143 (49.3%)
Single	21 (36.2%)	37 (63.8%)	1.28	1.02:1.61
Educational Level					0.64
0–8 years	35 (53.0%)	31 (47%)
9–11 years	112 (46.7%)	129 (53.5%)	1.14	0.85–1.51
12+ years	18 (47.8%)	20 (52.6%)	1.12	0.75–1.66
Family Income (monthly USD1)					0.31
0 – 397	49 (43%)	65 (57%)
398 −677	59 (47.9%)	64 (52.0%)	0.91	0.72–1.15
> 677	58 (53.2%)	51 (46.8%)	0.82	0.63–1.06
Skin Colour					0.25
White	84 (51.2%)	80 (48.8%)
Non white	82 (45.5%)	100 (54.95%)	1.12	0.92–1.38
Number of children					0.04
1 child	52 (41.6%)	73 (58.4%)
2 children	57 (48.3%)	61 (51.7%)	0.88	0.70–1.11
3 or more children	57 (55.3%)	46 (44.7%)	0.76	0.59–0.99
Planned Pregnancy					0.50
Yes	50 (51%)	48 (48.9%)
No	116 (47%)	131 (53.0%)	0.92	0.73–1.16
Mother’s age					0.06
18 to 24 years	53 (47%)	62 (53.9%)
25 to 29 years	48 (41.4%)	68 (58.6%)	1.08	0.86–1.36
30 years or more	65 (56.5%)	50 (43.5%)	0.80	0.61–1.05
Depression					0.03
No depression	101 (50.7%)	98 (49.4%)
Mild depression	43 (53.1%)	38 (47%)	0.95	0.72–1.24
Moderate to severe depression	22 (33.3%)	44 (66.7%)	1.35	1.08–1.68

¹Based on December 12-31-2014 exchange rates.

Table 2 reports the multivariate analysis, even after adjustment for covariates the association between moderate/severe depression and LMC remained. As shown in Table 2, after adjusting for all covariates, there was no significant risk between probable minor depression and LMC (PR=0.94; 95% CI 0.72–1.23). After controlling for all factors there is a significant risk for LMC among women with probable moderate or major depression (PR 1.37; 95% CI 1.10–1.71).

Table 2.

Multivariate analysis of PPD and LMC in three adjustment models

	Model 1	Model 2	Model 3
	PR (95% CI)	PR (95% CI)	PR (95% CI)
Depression
No	1.0	1.0	1.0
Mild	0.95 (0.72–1.25)	0.94 (0.71–1.23)	0.94 (0.72–1.23)
Moderate/Severe	1.32 (1.08–1.68)**	1.33 (1.06–1.65)*	1.37 (1.10–1.71)**

Model 1: crude (adjusted for randomization group on the PROGRAVIDA study)

Model 2: adjusted for model 1 plus social characteristics educational level, family income, ethnicity and marital status

Model 3: adjusted for model 2 plus maternal characteristics mother’s age, number of children and planned pregnancy.

Discussion

In the present study, LMC showed to be associated with moderate/major PPD even after controlling for sociodemographic variables and maternal characteristics among a sample of women who had depression during pregnancy. The women who showed severe depression symptoms presented a 37% increased risk of LMC in comparison to women with no depression symptoms. On the other hand, LMC was not associated to PPD in its mild form.

In this sample, 42.5% of postpartum women had no symptoms of depression while 19% had probable moderate/severe depression. This high prevalence can be explained by the fact that these women had presented depression during pregnancy³¹. In a study of 828 low income pregnant women in the western region of São Paulo, showed that antenatal depression is an important predictor of PPD Women with depression during pregnancy have a 2.5 times higher chance of presenting PPD. Nevertheless, both antenatal and PPD are frequently underdiagnosed and undertreated^19,32,33.

A study similar to the present one was undertaken where mothers were screened for maternal confidence³⁴. The authors evaluated 48 clinically depressed mothers and 38 non-depressed ones, showing that maternal self-efficacy relationship between maternal competence and PPD. Thus, maternal confidence may have a crucial role in the development parenting behavior and infant psychosocial risk. In these authors’ views, it is possible to raise the hypothesis that postpartum women with depressive symptoms tend to present lower levels of maternal confidence that in turn could compromise the type and quality of the attachment to be established with the baby and future child. In this sense, the evaluation of the level of maternal confidence is fundamental to understand the attachment between mother and baby^34,35.

A study with Japanese and Vietnamese mothers screened for maternal confidence in primary care settings verified an association between LMC and mothering a single child, being single, and being young³⁶. According the authors, such women may benefit from both social support and their partner’s support. It necessary to highlight that in the present study, on the contrary, the association verified between PPD and LMC lost significance after being adjusted for confounding variables in the multivariate model. The proportion of women without secure feeling of confidence was 22% in Japan and 66% in Vietnam. The same authors carried out another study in Japan where women were screened for maternal confidence in primary care and identified risks factors for women with low-maternal confidence such as interpersonal conflict in the home³⁷.

Some limitations of the present study must be mentioned. This study is cross-sectional in design therefore we cannot establish causal and temporal inferences between PPD and maternal confidence. If on the one hand severe postpartum depression increases the risk of LMC, on the other hand it is possible to suggest that women with LMC may come to develop or present symptoms of depression. Thus, the cross-sectional design does not exclude the risk of reversed causality. Another limitation was the follow-up loss. Of the 506 eligible women that originated our sample, only 346 (68.4%) provided complete data to evaluate the association between depression and LMC. Sample loss occurred either due to the impossibility of locating the research subjects through both telephone contact and home visits, or due to the mothers’ refusal to participate. Nevertheless, considering the randomization of the cluster trial, such losses were not differential. Another possible limitation is information bias. It is possible to suppose that the women felt pressured to reply affirmatively to the questions on the different baby-caring capabilities. Finally, the present study evaluated different socio-demographic variables and maternal characteristics that may have confounded the association between LMC and PDD. However, variables such as social support and intimate partner violence were not evaluated. A previous study showed that PPD is associated with violence by an intimate partner and that social support is an independent protection factor for PPD³⁸.

Among the strengths of the present study, it is worth mentioning the unprecedented nature of investigating the association between postpartum depression and maternal confidence in low-income postpartum women who had previously shown symptoms of depression during the pregnancy. The present study uses a dichotomous measure of maternal confidence, which has clinical implications for screening during maternal health and well-child visits. In previous studies, the ability to screen for maternal confidence can indicate additional risk factors such as violence. Future studies can screen for maternal confidence in the context of postpartum depression screening protocols.

The present study reinforces the need to investigate both maternal confidence and depression among postpartum women. As for depression in the pregnancy-postpartum cycle, Tran et al.³⁹ defend that it is essential for every country to have local evidence on the nature and prevalence of the problem for interventions are developed. According to authors, poorer countries need approaches that promote the integration of mental health care to primary perinatal health care, as well as strategies to reduce poverty and domestic violence, promote income-generation opportunities for women, and provide access to education. Finally, they conclude by stating that approaches to the issue must be multidisciplinary and include the investigation of the maternal history, education, community-based interventions, the development of health services, and the strengthening of health systems and political and social formation systems.

Conclusions

The evaluation of the maternal confidence, which is essential for the parental role, is warranted in primary care settings. Improving the mental well-being of these women can be an important strategy in the promotion of maternal confidence and childcare capability. Taking into consideration that low levels of parental confidence may undermine mother-baby interactions, promoting a feeling of maternal competence is crucial. This study suggests that maternal confidence evaluation should take place soon after the neonatal period. Screening for maternal confidence in primary care settings is appropriate alongside postpartum depression screening efforts.

Author Bios

Flavia Orante is a researcher in the area of mental disorders during pregnancy and the postpartum period at the Preventative Medicine Department at the University of São Paulo in Brazil

Karen Tabb Dina, PhD, MSW is associate professor in the University of Illinois at Urbana-Champaign School of Social Work and faculty affiliate at Carle Foundation Hospital, the Lemann Institute for Brazilian Studies, and the Women and Gender in Global Perspectives. Her research agenda focuses identifying risk factors for morbidity and mortality among perinatal women and clinical factors to improve minority health.

Yang Wang is a lecturer at Remin University in Bejing, China specializing in adult health and mental health.

Alexandre Faisal Curry, MD, PhD is a gynecologist/obstetrician who is a medical researcher at the Preventative Medicine Department at the University of São Paulo in Brazil since 2009. Dr. Curry is a researcher in the area of mental disorders during pregnancy and the postpartum period. He also conducts teaching activities with graduate and postgraduate students.