Systematic review on the prevalence of perinatal depression in Malawi

Background Perinatal depression causes significant burden to women and their families during the perinatal period. However, there is no reliable national prevalence data on perinatal depression in Malawi. Aim This systematic review aimed at establishing the pooled prevalence of perinatal depression. Setting The study setting is Malawi. Methods Two reviewers conducted the search, selection, quality evaluation and data abstraction. Appropriate terms were used to search the CINAHL, PsychINFO, PubMed and ScienceDirect databases. The relevance and the quality of the studies were assessed. The prevalence of prenatal depression was pooled using a random-effects model, which was used to synthesise the data. Results The review included a total of eight articles of fair and good quality. This review found a pooled prevalence of antenatal depression of 17.1% (95.0% confidence interval [CI]: 12.5–22.2) and postnatal depression of 19.8% (95.0% CI: 4.6–42.1) with an overall pooled prevalence of perinatal depression of 18.9% (95.0% CI: 14.5–23.8). Conclusion This systematic review provided a pooled prevalence of perinatal depression which may be used in the absence of national prevalence data on perinatal depression. Contribution This systematic review found a high a pooled prevalence of perinatal depression in Malawi suggesting that mental health should be a key component of maternal health programmes, policies and activities in the local setting.


Introduction
Perinatal depression is a depressive mood disorder that affects women during pregnancy or in the first 12 months after the birth of a child. 1 However, perinatal depression is often underdiagnosed by clinicians in Malawi. 2 There is evidence that found a pooled prevalence of perinatal depression of 11.9% in women in lower-income countries. 3 In Malawi, previous studies found that depression was prevalent during pregnancy (25.8%) 4 and after childbirth (10.7%). 5 This is within a range of prevalence rates for antenatal depression (8.3% -41.0%) 6 and postnatal depression (16.8%) in Africa. 7 Therefore, Malawi constitutes some burden of perinatal depression in Africa. their partners and infants. 10,12,13 For instance, a systematic review found that some women who committed neonaticide, infanticide or filicide had mental health concerns. 14 This may be the case in South Africa, where rates of neonaticide (19.6 per 100 000 live births) and infanticide (28.4 per 100 000 live births) are high. 15 Furthermore, perinatal depression is linked to poor uptake of antenatal services, 16 premature birth, intrauterine growth restriction, low birth weight, 17 fatigue, poor concentration and feelings of hopelessness in a pregnant woman. 18 Additionally, perinatal depression may affect mothers' ability to provide sufficient nutritional care resulting in compromised infant growth and development. 19,20 Perinatal depression may increase the risk of anxiety, depression, attention deficit hyperactivity disorder and conduct disorder in a child. 21 Additionally, perinatal depression and human immunodeficiency virus (HIV) infection are linked in a vicious cycle, in which the symptoms of one disease exacerbate the other's condition. 22 Women who have co-morbid perinatal depression and HIV infection are less likely to adhere to antiretroviral medication, which is essential for their survival and the prevention of HIV transmission to their babies. 23 This demonstrates the clinical and public health importance of perinatal depression 22,23,24,25 in low-resource settings where the condition is highly prevalent. 24 Although prevalence is often useful as it indicates the burden of a condition in a particular population, 26 there is no reliable national prevalence data on perinatal depression in Malawi. Therefore, this systematic review aimed at establishing the pooled prevalence of perinatal depression in Malawi.

Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines were used to conduct the systematic review 27 to address the review question.

Search process
To find relevant terms in the title, abstract and subject descriptors, a limited search of PubMed and ScienceDirect was conducted. After that, search phrases and synonyms were identified for use in searching various databases for prevalence studies conducted in Malawi. No date limits were set, in anticipation that a wider period to be searched from the inception of each database will yield many relevant studies for possible inclusion in this systematic review. However, the search focused on articles that were written in English only. Search terms that were used are presented in Table 1.
Using all identified search terms, electronic databases were searched for primary studies (PubMed, CINAHL, PsychINFO and ScienceDirect), and the results were imported into EndNote. Reference lists of key articles identified were hand-searched to identify additional articles. Manual searches of indexes and 'grey' literature databases were not carried out. The preliminary searches were conducted between July and August 2020, and the final search was done in October 2021.

Data Collection
Upon completion of the search, duplicates and irrelevant articles (abstracts, conferences, congresses, editorials, commentaries, reviews, news and irrelevant records) were removed from the EndNote database, and later the search data was exported into Excel. The selection of articles for review was then conducted in three phases.

Abstract screening
This first phase involved reviewers independently scrutinising the titles and abstracts and indicating which articles were relevant in Excel. When the abstract did not provide enough information or the reviewers were undecided, the full-text articles were evaluated, and a consensus was reached between the reviewers on whether the article should be included or excluded. To determine the level of agreement for eligibility for inclusion at this step, a kappa statistic was calculated.

Screening based on participant, intervention, comparison, outcome and setting criteria
In the second phase of the selection process, articles were independently reviewed by two reviewers by applying and extracting the participant, intervention, comparison, outcome and setting (PICOS) criteria as follows: Participants (P) (pregnant women and mother of infants ≤ 12 months), Intervention (I) (any studies reporting the prevalence of perinatal depression), Outcome measures (O) (the measured prevalence of perinatal depression) and study setting (S) (Malawi). In addition, all primary quantitative studies that measured the prevalence of perinatal depression were considered for inclusion in the

Assessment of methodological quality of studies
The critical appraisal tool for scoring methodological rigour of studies 28 was used to assess the methodological quality of the included studies. However, the reliability and the validity of the tool are not documented in Malawi. The critical appraisal tool for scoring methodological rigour of studies was used in this systematic review to assess the quality of included studies. 28 The tool has nine items that each assess the quality of a study on a scale as: very poor (1), poor (2), fair (3) and good (4). The tool has a minimum score of 9 and a maximum score of 36, and it allowed each reviewer to grade the included studies independently. Full texts of the included studies were retrieved and double-blind assessed for methodological quality by the two reviewers (P.M. and K.W.). The average score achieved by each study on the tool was used to determine its quality.

Data analysis
Quantitative data analysis was conducted using MedCalc software. 29

Review process and results
The electronic search yielded 1829 published records from PubMed, CINAHL, PsychINFO and ScienceDirect (Figure 1). A total of 23 duplicates were removed, resulting in 1806 records that were scrutinised. A further 1778 records were removed because they were irrelevant (conferences, congresses, editorials, commentaries, reviews and news).
The remaining 28 articles were assessed for relevancy by the reviewers using the PICOS criteria, excluding a further 18 articles, leaving 10 articles. The reviewers' ratings for inclusion or exclusion of studies agreed with a kappa = 0.94. Of the 10 articles that remained, two were excluded 31,32 because they did not report the prevalence of perinatal depression, resulting in eight studies that were included in this systematic review.

The methodological quality of reviewed studies
All the eight included articles were rated for quality independently by G.C. and K.W. Overall, the quality was satisfactory with four articles 4,33,34,35 rated as good, and four articles 5,20,36,37 were fair ( Table 2).

Findings from studies for inclusion in the review (n = 8)
Of the eight articles that were included in this systematic review, seven articles were published in medical journals and one in a nursing journal, and their years of publication ranged from 2010 to 2021 (Table 1). Most of the articles were for cross-sectional studies (n = 6).

Instruments that were used to measure perinatal depression
The systematic review revealed that various instruments were used to assess for perinatal depression in Malawi ( Table 3). The most-used instrument of the articles was Edinburgh Postnatal Depression Scale (n = 4), whose cut-off PICO, participant, intervention, comparison, outcome and setting.  points ranged from ≥ 6 to >12. The second most-used instrument was SCID (n = 2), followed by the Self-Reporting Questionnaire 20 (n = 1), with a cut-off point ≥ 8, and the MINI (n = 1) ( Table 3).     (Table 6) were comparable. However, the pooled prevalence of postnatal depression was higher than the overall pooled prevalence.

Discussion
Depression is the most common mental health condition that affects women worldwide during the perinatal period. 9,10,38,39,40 It is necessary that women are screened for perinatal depression for early detection and treatment. 4 However, the studies and instruments that are used to measure perinatal depression globally vary, 3,7,41,42 leading to clinical heterogeneity. Clinical heterogeneity refers to variability in study population characteristics, interventions and outcomes across studies, while statistical heterogeneity includes methodological heterogeneity, biases and random error. 43 This systematic review showed that the included primary studies had substantial clinical heterogeneity regarding study designs, sample sizes, sample characteristics and assessment tools. This might have affected the magnitude of the overall pooled prevalence that was generated in this study. As indicated by Kriston and colleagues, reviewers' beliefs on the role of heterogeneity on the pooled prevalence found in this systematic review must be reflected. 44 This systematic review suggests that the prevalence of perinatal depression in Malawi is high with an overall pooled prevalence of 18.9%. This pooled prevalence for perinatal depression in Malawi is lower compared to the one that was found in Ethiopia (25.8%). 45 However, contrary evidence showed that the pooled prevalence of perinatal depression in Malawi is higher than that of women in lower-income countries (11.9%). 3 This may be explained by literature which indicated that women encounter numerous risk factors during the perinatal period in Malawi. 4,5,46 However, the need for conducting a proper national epidemiological study for perinatal depression still remains in the country. Epidemiological data will help in the implementation of targeted evidencebased interventions to protect the public 47 and perinatal women in particular.
Literature suggests that violence, anxiety, life stress, prior depression and lack of social support are some of the risk factors for perinatal depression. 48 Furthermore, prenatal depression is a risk factor for postnatal depression, 49 so that pregnant women with untreated depression are more likely to suffer from postpartum depression and suicidality. 50 This may be the case in Malawi, where this systematic review found that pooled prevalence for postnatal depression (19.8%) was higher than the pooled prevalence for antenatal depression (17.1%). Nonetheless, this result is contrary to the evidence that showed that pooled prevalence for postnatal depression (16.8%) 7 was lower than the pooled prevalence of antenatal depression (26.3%) 42 in Africa.
Despite the variations, this review agrees with existing evidence that the prevalence of perinatal depression is high in Africa and Malawi in particular.

Study limitations
This systematic review is limited in that there is a shortage of studies in Malawi to produce generalisability evidence. There is also a difference in the depression assessment tools, which may also have contributed to the detected heterogeneity in this review. Hence, caution should be taken during the interpretation of the results.

Conclusion
This systematic review generated a pooled prevalence of perinatal depression for Malawi, which may be used by clinicians, researchers and policymakers in the absence of national prevalence data on perinatal depression. The review suggests that perinatal depression is highly prevalent in Malawi, based on a few published studies with inherently heterogeneous estimates. Based on this review, maternal health programme policies and activities should incorporate maternal mental health as a core component to promote early detection of perinatal depression and prompt interventions that would save the mother and her baby from different forms of morbidity and mortality.