A cross-sectional study was conducted with an aim to screen for depression and associated risk factors in pregnant women attending antenatal care services in eight public antenatal clinics (four urban and four rural in January to May 2016) in the Blantyre district, Malawi, using the Edinburgh Postnatal Depression Scale (EPDS) and the Pregnancy Risk Questionnaire (PRQ). In a sub-sample, depression diagnosis was determined using the Mini International Neuropsychiatric Interview (MINI) and the EPDS was validated.

To determine the prevalence of depression in this setting, a sample size of 480 pregnant women was calculated using the formula: N = (TP + FN)/ (1-P),⁹ with N = 1593, sensitivity 96%, specificity 57%,¹⁰ confidence interval 95% and p = 21%.² Inclusion criteria were: attending an antenatal service with a pregnancy of any stage, being 18 years old and above, agreeing voluntarily to participate in the study, giving a written consent before joining the study and being able to speak and understand Chichewa (a local language). Exclusion criteria were: having complications of pregnancy or a known mental disorder. To determine Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV) criteria for major depressive disorder (MDD) and to validate the EPDS as a gold standard, a sub-sample of 100 pregnant women was calculated as adequate to validate the EPDS against the MINI with sample parameters of p = 21%,² N = 480, error level 7% and confidence interval 95%.

This study used a standardised depression screening instrument, namely the EPDS, and the validated PRQ to collect data on risk factors for pregnancy. Although a previous study in Malawi used the SRQ,² this study was performed in a single clinic with a convenience sample. In this study, following a systematic review of accurate instruments for screening depression in low-resource settings,¹¹ the EPDS was selected. The EPDS is a 10-item self-reporting questionnaire for screening postnatal depression¹² which can also be used to screen for antenatal depression.¹¹ The instrument has a maximum total score of 30 with a standard cut-off score of ≥ 10 for depression caseness.¹³ The EPDS targets depressive symptoms that an individual has experienced in the past 7 days.¹² The MINI, a brief structured diagnostic interview for DSM-IV,¹⁴ was used as the ‘gold standard’ in generating psychiatric diagnoses using only the MDD module. The PRQ is an 18-item scale which is designed to assess psychosocial factors for depression during pregnancy. The instrument was designed to assess psychosocial risk factors for depression during pregnancy and used to predict antenatal or postnatal depression.¹⁵ The PRQ has a maximum total score of 90. Previously validated Chichewa language versions of the EPDS were used in this study.⁵ The PRQ and the MINI were translated into Chichewa by the researcher and a bilingual social worker through forward and backward translations.¹⁶

Research assistants trained in administration of the EPDS and PRQ collected the data from 480 pregnant women. They systematically picked every other third pregnant woman from the queue after randomly picking the first one. A sub-sample of 100 pregnant women drawn from the 480 respondents volunteered to undergo further interview using the MINI. A mental health nurse administered the MINI and was blind to the respondents’ initial screening outcomes.

The International Business Machines (IBM) Statistical Package for Social Sciences (SPSS) version 22.0 was used to analyse data. Significance level for all tests was set at 95%. Descriptive statistics were used to summarise data. A diagnosis of MDD based on the MINI was assigned. An EPDS ‘case’ of depression, that is, being screened positive, was computed based on a cut-off score of ≥ 10. The EPDS cases were validated against the MDD diagnosis using standard sensitivity analysis. Differences in demographics and psychosocial risk factors between screen positives and negatives were compared using Pearson chi-square test. Associations between psychosocial risk factors and EPDS depression cases were tested using odds ratios (OR). Direct logistic regression was performed to assess the impact of demographic variables and psychosocial risk factors in screening positive for depression in pregnant women.

A total of 480 out of a possible 496 respondents had the EPDS and PRQ administered (response rate of 96.8%). The age of respondents ranged from 18 to 43 years (mean 25.2 ± 5.5). The mean number of pregnancies per respondent was 2.4 ± 1.3 (range = 1–6 pregnancies), with a current mean gestation period of 26.7 weeks ± 7.4 (range = 5–40 weeks). More than half of the respondents were unemployed (52.5%, n = 252), had more than primary school education (53.8%, n = 256) and were from an urban area (65.2%, n = 313). Nearly all the respondents were supported by a partner (92.9%, n = 446). No significant differences were found between demographic data of the depression cases and non-depression cases except for marital status and occupation (approached significance) and no difference by clinical pregnancy variables (Table 1).

A total of 3 out of the 100 respondents in the sub-sample refused to be interviewed, resulting in a sample size of 97 (97%) – 48 screen positives and 49 screen negatives (mean = 27.7 ± 7.9 years). The prevalence of current MDD using the MINI was 25.8% (95.0% CI = 17.5% – 34.0%, n = 25), with no significant differences between demographic and clinical data of the MDD cases and non-MDD cases. The EPDS had a sensitivity of 68.0%, specificity of 88.0% and AUC = 0.85, using the MDD diagnosis as the gold standard (MINI), confirming its validity in measuring risk for antenatal depression.

Using the EPDS, this study found rates for antenatal depression cases (screen positives) of 19.0% (95.0% CI 15.5% – 22.5%, n = 91) in pregnant women attending antenatal clinics in Blantyre district. A significantly higher proportion of EPDS cases reported that were not supported by a partner (12.1%, n = 366) compared to 5.9% (n = 80) non-cases (p = 0.01) (Table 1). The demographic characteristics of the EPDS cases were similar to the sub-sample MDD cases (urban areas [62.8% vs. 60.0%], secondary education or over [49.5% vs. 56.0%], not supported by a partner [12.1% vs. 12.0%]), except for unemployment which was higher in MDD cases (58.2% vs. 80.0%). Similarly, the mean number of pregnancies in the EPDS versus MDD cases were 2.3 versus 2.5 and mean gestations of 27.7 weeks for both.

The PRQ was used to assess psychosocial risk factors associated with antenatal depression. There were significant differences between EPDS positive and negative screened cases in 12 of the psychosocial risk factors measured by PRQ (Table 2).

Edinburgh Postnatal Depression Scale screen positives were significantly associated (p < 0.05) with being ‘distressed by anxiety or depression for a period of 2 weeks or more during this pregnancy’ (χ² = 101, p < 0.001) (OR = 4.2 [2.2–8.2], p < 0.001), experiencing ‘major stresses, changes or losses in the course of this pregnancy’ (χ² = 81.7, p < 0.001) (OR = 3.5 [1.7–6.9], p < 0.001), were ‘physically abused when they were growing up’ (χ² = 19.8, p < 0.001) (OR = 1.7 [82–3.4], p = 0.15), had a history of feeling ‘miserable or depressed for 2 weeks or more before this pregnancy’ (χ² = 28.5, p < 0.001) (OR = 2.3 [1.2–4.3], p = 0.01), ‘feeling that father was critical of her when growing up’ (χ² = 4.2, p = 0.04) (OR = 3.5 [0.97–12.8], p = 0.06), ‘having trouble finishing jobs because of wanting to get it exactly right’ (χ² = 7.8, p = 0.01) (OR = 1.8 [0.92–3.4], p = 0.09), ‘feeling that will have no people to depend on for emotional support after giving birth’ (χ² = 23.9, p < 0.001) (OR = 1.9 [0.86–4.1], p = 0.12), ‘feeling that a relationship with partner is not an emotionally supportive one’ (χ² = 43.3, p < 0.001) (OR = 3.8 [1.5–9.5], p = 0.004), ‘feeling that pregnancy has not been a positive experience’ (χ² = 16.9, p < 0.001) (OR = 1.6 [0.64–4.1], p = 0.3), ‘feeling that mother was not emotionally supportive of her when growing up’ (χ² = 7.9, p = 0.01) (OR = 3.9 [0.84–18.1], p = 0.08), ‘previously told by health professional that she was depressed or needed antidepressants’ (χ² =8.2, p = 0.004) (2.3 [0.62–8.2], p = 0.22) and ‘sexually abused when growing up’ (χ² = 13.6, p < 0.001) (OR = 4.1 [0.56–30.2], p = 0.17) compared to screen negatives (Table 2).

A direct logistic regression model with 14 variables (2 demographic and 12 psychosocial risk factors with significant differences) was constructed. The model with the 14 variables was statistically significant (χ² = 153.9, p < 0.001) and it correctly classified 87.7% of screen positives. Furthermore, the model showed that there were only five predictors of caseness for antenatal depression, namely: (1) ‘being distressed by anxiety or depression for more than 2 weeks during this pregnancy’ (OR = 4.1 [2.1–7.9], p £ 0.001); (2) ‘feeling that a relationship with partner is not an emotionally supportive one’ (OR = 3.5 [1.4–8.4], p = 0.01); (3) ‘having major stresses, changes or losses in the course of this pregnancy’ (OR = 3.2 [1.7–6.2], p = 0.01); (4) ‘feeling that father was critical of her when growing up’ (OR = 3.2 [1.4–7.6], p = 0.01); and (5) ‘having history of feeling miserable or depressed for ≥ 2 weeks before this pregnancy’ (OR = 2.4 [1.3–4.4], p = 0.01). This showed that respondents who had been distressed by anxiety or depression for 2 weeks or more during pregnancy had the highest likelihood (four times) of screening positive for antenatal depression in this study.

This study found that depression is significantly associated with being alone, unemployment, major stresses, poor relationships, physical or sexual abuse, lack of support and prior history of anxiety or depression. There is evidence which shows that domestic violence, maternal anxiety, life stress, prior depression and lack of social support are psychosocial risk factors of antenatal depression.⁶ This study confirmed these factors. A personal history of depression and experiencing stress are risk factors that are associated with antenatal depression.⁶ The risk factor which was the strongest predictor of depression was ‘being distressed by anxiety or depression during pregnancy’, with respondents reporting this being four times more likely to screen positive for antenatal depression. Another risk factor that was associated with screening positive for antenatal depression was ‘having major stresses, changes or losses in the course of this pregnancy’ with respondents who had experienced major stresses, changes or losses during pregnancy being three times more likely to be depressed.

It is documented that antenatal depression is more likely to occur among pregnant women with a recent history of stressful life events.^17,18 Major stresses such as the death of a relative or intimate partner violence may have contributed to their depressive symptoms. In addition, it has been postulated that childhood physical abuse is associated with depressive symptoms in early pregnancy.¹⁹ ‘Having history of physical abuse when growing up’ was significantly associated with antenatal depression in this study, although this was not a significant factor in the overall prediction model.

The thought of having an unsupportive partner was significant in this study, with ‘feeling that relationship with partner is not an emotionally supportive one’ being found to be a risk factor that predicted antenatal depression (OR = 3.5 [1.4–8.4], p = 0.01). This is supported by other studies that showed that a pregnant woman may suffer from depression if she lacks support from her partner.²⁰ Most pregnant women attending antenatal clinics in low-resource settings depend on their partners for financial support² and are at risk of depression if they do not receive adequate psychosocial support.⁶ Psychosocial support serves as a buffer from stressful life events by providing resources, support and strength during pregnancy.²¹

The relationship of children and their fathers during childhood may reduce or increase the children’s vulnerability to emotional problems. This study found that pregnant women who indicated that their fathers were critical of them when growing up were three times more likely to screen positive of antenatal depression. Consistent with this finding is Rosenberg and Wilcox who asserted that girls who had a good relationship with their fathers develop a stronger self-esteem and are less likely to experience depression.²² Therefore, this study suggests that poor relationship between father and girl child during childhood is a risk factor for antenatal depression.

This study may have been affected by selection bias because pregnant women who did not present themselves at antenatal clinics were not represented. Secondly, the interviewer administration of screening instruments may have influenced respondents to give answers that they deemed as socially acceptable in the presence of the interviewer.