Poor sleep quality and suicidal ideation among pregnant women during COVID-19 in Ethiopia: systematic review and meta-analysis

General objective

• This study aimed to provide comprehensive evidence on the prevalence and associated factors of poor sleep quality and suicidal ideation among pregnant women during COVID-19 in Ethiopia.

Specific objectives

• To estimate the pooled prevalence of poor sleep quality and suicidal ideation during COVID-19 among pregnant women in Ethiopia.

• To review associated factors of poor sleep quality and suicidal ideation during the COVID-19 pandemic among pregnant women in Ethiopia.

Protocol registration

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 recommendation was used (Moher et al., 2015). The protocol for this systematic review and meta analysis was registered by International Prospective Register of Systematic Reviews with PROSPERO registration number: CRD42023389896.

Search strategy

Searching of literature was done by two authors (AAH and AAS) individually. Disagreement was handled by discussion with third author (AAM). Databases PubMed, African Journals Online and Google Scholar, and papers published from the occurrence of the COVID-19 pandemic to February 2023 were searched. Observational studies that evaluated sleep quality and suicidal ideation during COVID-19 among pregnant women in Ethiopia were considered. Systematic searches were carried out by combining every possible predefined search term determined by Medical Subject Headings (MeSH) and keywords. The following key words were used: “COVID-19”, “2019 novel coronavirus disease”, “2019 novel coronavirus infection”,“2019 ncov disease”, “2019 ncov infection”, “covid 19 pandemic”, “covid 19 pandemics”, “covid 19 virus disease”, “covid 19 virus infection”, “COVID19”, “coronavirus disease 2019”, “coronavirus disease 19”, “sars coronavirus 2 infection”, “sars cov 2 infection”, “severe acute respiratory syndrome coronavirus 2 infection”, “SARS-CoV-2”, “2019 novel coronavirus”, “2019 novel coronavirus”, “2019- nCoV”, “covid 19 virus”, “covid19 virus”, “Coronavirus disease 2019 virus”, “SARS coronavirus 2”, “SARS cov 2 virus”, “severe acute respiratory syndrome coronavirus 2”, “Wuhan coronavirus”, “Wuhan seafood market pneumonia virus”, “mental illness”, “psychiatric problem”, “psychology problem”, “mental health effect”, “psychological disturbance”, “mental disorder”, “psychiatric illness”, “psychiatric diseases”, “psychiatric disorders”, “behavior disorders”, “severe mental disorder”, “sleeping quality”, “poor sleep quality”, “insomnia”, “suicide”, “suicidal ideations”, “pregnant women”, “prenatal”, “ perinatal”, “postpartum”, “antenatal”, “postnatal”, “puerperal”, “lactating women”, “reproductive age women”, “child bearing women” and “Ethiopia”.

Reference lists of full text articles included in the review were checked to identify any potentially eligible study for this review. Systematic procedure confirmed that the literature search encompasses all published studies on the prevalence and associated factors of poor sleep quality and suicidal ideation during COVID-19 among pregnant women in Ethiopia. From the searched results duplicates were detached by using Mendeley (Kwon et al., 2015). The search strategy for finding suitable studies from databases is presented in Supplemental File S1.

Eligibility criteria

Exclusion criteria

The following types of studies were excluded: studies focusing on the whole population, studies that do not report prevalence and associated factors of poor sleep quality and suicidal ideation among pregnant women, systematic reviews and meta-analyses, randomized controlled trials, editorials, and conference abstracts and opinions.

Outcome measures

The primary outcomes were the pooled prevalence of prevalence of poor sleep quality and suicidal ideation among pregnant women during COVID-19 in Ethiopia. The secondary outcomes were factors associated with poor sleep quality and suicidal ideation among pregnant women during COVID-19 in Ethiopia.

Selection of studies

Two researchers (AAH and AAS) reviewed studies based on eligibility criteria. Firstly, the authors evaluated titles and abstracts of the studies from the searched databases. Then, full-text screening was done to screen the full texts. We have provided a reasoning for inclusion and exclusion of studies in the PRISMA diagram. The list of suitable studies for this systematic review and meta-analysis were arranged.

Data extraction

The data were collected by two researchers (AAH and AAS) individually. Pretest on the data extraction form by using two pilot surveyed studies was done. This helps ensure the collection of all necessary data for this systematic review and meta-analysis is accurate. Disagreements were handled by deep discussion among the investigators. All relevant data were collected from the original studies. Specifically, author’s last name, year of publication, region, study design, number of cases, sample size, sampling design, instrument, study population, average age and prevalence of poor sleep quality and suicidal ideation with their associated factors were extracted.

Methodological quality assessment

Two researchers (AAH and AAM) separately evaluated the quality of the included studies by the Newcastle-Ottawa Scale (NOS) (Peterson, Welch & Losos, 2011). According to this scale, three parameters named selection, comparability, and assessment of exposure/outcome were considered. Studies with less than 5 scores were considered low quality, 5-7 scores of moderate quality, and more than 7 scores of high quality (Ssentongo et al., 2020). Studies with moderate and above quality score were considered for this study. Disagreements on scores between reviewers were resolved by discussion. The final quality score was presented in Table 1.

Data synthesis

The statistical data analysis was done by Stata 14.0 (StataCorp, Collage Station, Texas, USA). The pooled prevalence with 95% confidence interval (CI) by using random-effect model with the generic inverse variance method was computed. Evaluation of heterogeneity was checked by I² and Cochran’s Q-statistics (Bowden et al., 2011; Zhu et al., 2020). Subgroup analyses was performed by regions, year of publication and sampling methods used in individual studies to decide the source of heterogeneity. Publication bias was checked by Doi plot and the Luis Furuya Kanamori (LFK) index was applicable when the number of studies was small and used to assess asymmetry (Furuya-Kanamori, Barendregt & Doi, 2018) (Furuya-Kanamori & Doi, 2021). The LFK index value out of the interval -1 and 1 were considered asymmetry (existence of publication bias) (Furuya-Kanamori & Doi, 2021).

Study characteristics

In this systematic review and meta-analysis, we included nine cross sectional studies (Amare, Chekol & Aemro, 2022; Anbesaw et al., 2021a; Anbesaw et al., 2021b; Bogale et al., 2022; Dule et al., 2021; Jemere et al., 2021; Belete et al., 2021; Molla, Nigussie & Girma, 2022; Takelle, Muluneh & Biresaw, 2022; Anbesaw et al., 2021b) on the prevalence and associated factors of poor sleep quality and suicidal ideation among pregnant women during the COVID-19 pandemic in Ethiopia. Based on the types of mental disorders; six studies (Amare, Chekol & Aemro, 2022; Bogale et al., 2022; Anbesaw et al., 2021a; Anbesaw et al., 2021b; Jemere et al., 2021; Takelle, Muluneh & Biresaw, 2022; Dule et al., 2021) were reported about poor sleep quality; three studies (Belete et al., 2021; Anbesaw et al., 2021b; Molla, Nigussie & Girma, 2022) were reported about suicidal ideation. Furthermore, the summarized data of the key characteristics of the included studies for this review were showed in Table 1.

Quality of included studies

The quality of included studies using the modified Newcastle Ottawa scale (NOS) for cross-sectional studies quality assessment was presented in Table 1. We considered moderate quality, and high-quality studies for this review. Thus, one study (Dule et al., 2021) that was appraised as moderate quality and eight studies (Amare, Chekol & Aemro, 2022; Bogale et al., 2022; Anbesaw et al., 2021b; Belete et al., 2021; Jemere et al., 2021; Molla, Nigussie & Girma, 2022; Takelle, Muluneh & Biresaw, 2022) that were appraised as high quality were considered for final systematic review and meta analysis.

Publication bias

Publication bias was evaluated by using Doi plot (Furuya-Kanamori, Barendregt & Doi, 2018), a method used to see asymmetry and by the LFK index (Furuya-Kanamori & Doi, 2021), used to detect and quantify asymmetry of study effects. For poor sleep quality studies (LFK index = −1.49, Egger’s test p-value =0.531), i.e., there is minor asymmetry, but the Egger’s test result is not statistically significant with p-value =0.531, indicating that no significant influence was found on the pooled result and therefore no publication bias. For suicidal ideation studies (LFK index = −0.05, Egger’s test p-value = 0.674). i.e, both LFK index and Egger’s test supports no publication bias (Fig. 2).

Pooled prevalence of poor sleep quality

Six studies reported the prevalence of anxiety, and the pooled prevalence of poor sleep quality was found to be 55% (95% CI [0.42–0.69], I² = 98.1%, p < 0.001). As shown in Fig. 3 there is substantial heterogeneity among study findings on prevalence of poor sleep quality among pregnant women during the COVID-19 pandemic.

Subgroup analysis of poor sleep quality by region

From the subgroup analysis of prevalence of poor sleep quality by region, the pooled prevalence of poor sleep quality in Oromiya, Amhara and South West Ethiopia was 44%, 58% and 71% respectively (Fig. 4). The prevalence was higher in the South West Ethiopia region compared to the others. There was statistically significant heterogeneity among regions (Q = 6.39, df = 2, p = 0.032).

Subgroup analysis of poor sleep quality by year of publication

From the subgroup analysis by publication year (Fig. 5), the pooled prevalence of poor sleep quality in 2021 and 2022 was 52% and 59%, respectively. The prevalence of poor sleep quality among pregnant women was high in studies published in 2022. The heterogeneity test between group result (Q = 0.2, df = 1, p = 0.657) implies that there was no significant heterogeneity in years of publication on the prevalence of poor sleep quality.

Subgroup analysis of poor sleep quality by sampling method

From the subgroup analysis by sampling method (Fig. 6) the pooled prevalence of poor sleep quality in study using sampling methods census was 57%, systematic sampling was 63% and multistage sampling was 71%. The heterogeneity test between group (Q = 20.60, df = 3, p = 0.000) implies that there is statistically significant heterogeneity among sampling methods.

Pooled prevalence of suicidal ideation

From three studies, the pooled prevalence of suicidal ideation was 11% (95% CI [0.09–0.13], I² = 48.2%, p = 0.145) as shown in Fig. 7. No significant heterogeneity was observed. We did not do subgroup analysis to assess the source of heterogeneity.

Associated factors and pooled effect size

As shown in Table 2, the pooled adjusted odds ratio of age >= 30 years pregnant women was 1.95, 95% CI (0.85, 3.06), suggesting that the odds of developing poor sleep quality among pregnant women was 1.95 times more with age >= 30 years pregnant women than women with age <30 years. The pooled adjusted odds ratio of 3rd trimester pregnant women was 3.20, 95% CI (1.82, 4.58), implying that the odds of developing poor sleep quality among pregnant women on 3rd trimester gestational age was 3.20 times more than women in the 2nd or 1st trimester.

The pooled adjusted odds ratio of substance use among pregnant women was 2.51, 95% CI (0.99, 4.04), suggesting that the odds of developing poor sleep quality among pregnant women with substance use was 2.51 times more than women without substance use. The pooled adjusted odds ratio of pregnant women with depression was 2.97, 95% CI (0.92, 5.02) suggests that the odds of developing poor sleep quality among pregnant women with depression was 2.97 times more than women with out depression. Likewise, the pooled adjusted odds ratio of pregnant women with stress was 2.14, 95% CI (0.24, 4.03) suggests that the odds of developing poor sleep quality among pregnant women with stress was 2.14 times more than women without stress.

We also reported pooled effect size of associated factors of suicidal ideation among pregnant women. The pooled adjusted odds ratio of pregnant women with depression was 3.19, 95% CI (1.68, 4.71) suggesting that the odds of developing suicidal ideation among pregnant women with depression was 3.19 times more than women that do not have depression. The pooled adjusted odds ratio of pregnant women with intimate partner violence was 3.83, 95% CI (−0.67, 8.34) suggesting that the odds of developing suicidal ideation among pregnant women with intimate partner violence was 3.83 times more than women without intimate partner violence. Due to the limited number of studies available, pooling the adjusted odds ratio for the associated factors of poor sleep quality and suicidal ideation is not possible. Hence, we summarized the collected findings in Table 3.