Depression and insomnia among healthcare professionals during COVID-19 pandemic in Ethiopia: a systematic review and meta-analysis

Protocol registration

This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and registered in the International Prospective Register of Systematic Reviews with PROSPERO registration number: CRD42022314865.

Search strategy

PubMed, Cochrane Library, CrossRef, African Journals Online and Google Scholar databases were searched for articles published from the occurence of the pandemic to June 2022. To assess the mental health impact of COVID-19 among healthcare professionals, observational studies were considered. Systematic searches were conducted by combining every possible combination of medical subject headings (MeSH) terms and keywords. Reference lists of key full text articles included in the review were checked to recognize any potentially eligible studies. The systematic procedure verifies that the literature search comprises all published studies on the impact of COVID-19 among healthcare professionals in Ethiopia. The search results were exported to Mendeley and duplicates were removed (Kwon et al., 2015). Two authors (AAH and AAS) independently screened titles and abstracts of the studies, and any disagreement between the authors was resolved by discussin with third author (AAM). The search strategy of PubMed database is presented (Table 1). The search strategy is considered as adequate to reduce the risk of selection and detection bias. For this study only observational studies (cohort, case-control and cross-sectional) focus the impacts of COVID-19 on depression and insomnia among healthcare professionals during the pandemic in Ethiopia were included.

Setting/context: Studies conducted in Ethiopia was the main concern of this review.

Population: All categories of healthcare professionals in Ethiopia.

Study design: Observational studies (cohort, case-control and cross-sectional studies) that reported the prevalence and associated factors of mental disorders among healthcare professionals during the COVID-19 pandemic.

Language: English language reported studies were considered.

The following types of studies were excluded: studies on whole populations; studies with very small sample size (n < 30); studies that did not have enough statistical information to be extracted and descriptive reviews, randomized controlled trials, systematic review, meta-analysis, editorials, comments, conference abstracts, and expert opinions, not exactly reported the prevalence and the determinants of mental health problems among healthcare professionals were excluded.

Outcome measures

There are two main outcomes in this systematic review and meta-analysis. The first outcomes were the prevalence of depression and insomnia on healthcare professionals during the COVID-19 pandemic. The second outcome of the study was factors related to the prevalence of depression and insomnia among healthcare professionals during the COVID-19 pandemic in Ethiopia.

Selection of studies

Two authors (AAH and AAS) assessed the studies based on inclusion and exclusion criteria. Firstly, the authors assessed both the titles and abstracts of the studies identified from the searched databases. Then full-text screening was done to screen the full texts selected in the previous stage. Moreover, we have a rationale for inclusion and exclusion of studies in the PRISMA flow diagram. Lastly, the final list of articles for data extraction for systematic review and meta-analysis was prepared.

Data extraction

The following data were extracted from each article by two authors independently: author’s name, study type, total number of participants, year of publication, region, study design cases, sample size, instrument used, mental disorders, prevalence of mental disorders, and significant associated factors of mental disorders with their effect size. There was pretest the data extraction form to ensure effective, facilitates the collection of all necessary data required for the valuable systematic review and meta-analysis. Disagreements were resolved by deep argument among authors.

Methodological quality assessment

Two authors (AAH and AAS) separately assessed the quality of included studies using the Newcastle-Ottawa Scale (NOS) (Stang, Jonas & Poole, 2018). The scale scores observational studies based on three parameters: selection, comparability and exposure/outcome assessment. Studies with less than five scores were considered low quality, five to seven scores of moderate quality, and more than seven scores of high quality (Ssentongo, Ssentongo & Heilbrunn, 2020). Only studies with moderate and above quality score were included in this systematic review and meta-analysis.

Data synthesis

The extracted data was entered into a Microsoft Excel and then imported in to Stata version 16.0 (StataCorp. 2019. Stata Statistical Software: Release 16. College Station, TX: StataCorp LLC) software for the analyses. We calculated pooled prevalence and pooled adjusted odds ratios (AOR) with 95% confidence interval (CI) by the generic inverse variance method. Heterogeneity among included studies was assessed using the I² test. If I² > 0.5 or P < 0.1 it is considered that there is a significant heterogeneity among the included studies (Zhu et al., 2020) and random-effect model with the inverse variance method was used. To determine the source of heterogeneity subgroup analyses was performed by regions and instruments used in individual studies, the difference between the subgroup was assessed by Cochran’s Q-statistics (Bowden et al., 2011).

Characteristics of included studies

In this systematic review and meta-analysis, we included 8 studies (Jemal, Deriba & Geleta, 2021; Wayessa et al., 2021; Jemal et al., 2021; Yadeta, Dessie & Balis, 2021; GebreEyesus et al., 2021; Asnakew, Amha & Kassew, 2021; Yitayih et al., 2021; Habtamu et al., 2021) focusing on the impact of COVID-19 on depression and insomnia among healthcare professionals in Ethiopia. Regarding the regional distribution one study (GebreEyesus et al., 2021) is from SNNP, three studies (Jemal, Deriba & Geleta, 2021; Wayessa et al., 2021; Yadeta, Dessie & Balis, 2021) are from Oromiya, one study (Asnakew, Amha & Kassew, 2021) is from Amhara, one study (Jemal et al., 2021) is from Addis Ababa and Oromiya, one study (Habtamu et al., 2021) is from Addis Ababa. Furthermore, the key characteristics of the included papers was summarized and showed in table (Table 2).

Quality of included studies

The methodological quality score of the eight included studies using the modified Newcastle Ottawa scale for correctional studies quality assessment tool was presented (Table 2). Accordingly, two studies were rated as moderate quality (Asnakew, Amha & Kassew, 2021; Yitayih et al., 2021) and six studies were rated as high quality (Jemal, Deriba & Geleta, 2021; Jemal et al., 2021; Yadeta, Dessie & Balis, 2021; GebreEyesus et al., 2021; Habtamu et al., 2021; Wayessa et al., 2021) were considered for final systematic review and meta analysis.

Publication bias

Detection of publication and related biases is vital for the validity and interpretation of meta-analytical findings. The test power is usually too low to distinguish chance from real asymmetry when there are less than 10 studies in the meta-analysis (Furuya-Kanamori, Barendregt & Doi, 2018). Accordingly, the number of included studies for depression and insomnia are less than ten we do not apply the asymmetry test.

Pooled prevalence of depression

A total of seven studies reported the prevalence of depression, and the pooled prevalence of the depression was 40% (95% CI [0.23–0.57]; I² = 99.00%; p = 0.00) (Fig. 2). From the heterogeneity test, there is significant heterogeneity is observed among individual studies on the prevalence of depression among healthcare professionals during the pandemic in Ethiopia.

Subgroup analysis of depression by region

To handle the variability in studies the subgroup analysis by region is done. From the forest plot (Fig. 3), the pooled prevalence of depression in Addis Ababa & Oromiya, Oromiya, SNNP, Addis Ababa and Amhara is 60%, 35%, 26%, 27% and 63% respectively. The heterogeneity test (Q = 226.85, p = 0.000) indicates that there is significant variability among regions. The prevalence of depression is higher in Amhara region compared to the others.

Subgroup analysis of depression by instrument

Based on the instrument used in individual included studies, subgroup analysis is done. From the forest plot (Fig. 4), the pooled prevalence of depression by DASS-21 and PHQ-9 is 48%, and 34% respectively. The heterogeneity test (Q = 0.79, p = 0.374) indicates that there is no significant variability on a study finding between measurements. The prevalence of depression measured in DASS-21 is higher than that measured by PHQ-9 among healthcare workers during the pandemic. This might be due to the difference in sensitivity and specificity of the assessment instruments.

Pooled prevalence of insomnia

Three studies reported the prevalence of insomnia, and the pooled prevalence of the insomnia was 35% (95% CI [0.13–0.58]; I² = 98.20%; p = 0.00) (Fig. 5). In the test of heterogeneity, we have seen that there is considerable variation among individual included studies on the prevalence of insomnia among healthcare professionals during the pandemic in Ethiopia.

Subgroup analysis of insomnia by region

Subgroup analysis by region is done. From the forest plot (Fig. 6), the pooled prevalence of insomnia in Oromiya and Addis Ababa is 24% and 41% respectively. The heterogeneity test (Q = 22.01, p = 0.000) indicates that there is significant variation in the prevalence of insomnia between regions. The prevalence of insomnia is higher in Addis Ababa than Oromiya.

Subgroup analysis of insomnia by instrument

Based on the instrument used in individual included studies, subgroup analysis is done. From the forest plot in (Fig. 7), the pooled prevalence of depression measured by ISI and PSQI is 24% and 41% respectively. The heterogeneity test (Q = 22.02, p = 0.000) indicates that there is significant variation on a study finding between measurements. The prevalence of depression measured in PSQI is higher than that measured by ISI among HCWs during the pandemic.

Pooled adjusted odds ratio of associated factors of depression and insomnia

The pooled adjusted odds ratio of the factors associated with prevalence of depression among healthcare professionals during COVID-19 in Ethiopia presented (Table 3).

The pooled adjusted odds ratio on female healthcare workers is 2.09, 95% CI [1.41–2.76], implies that the odds of female healthcare workers is two times more to develop the depressive symptom than males during the pandemic. Similarly for marital status (being married) the pooled adjusted odds ratio is 2.95, 95% CI [1.83–4.07] indicates that the odds of married healthcare workers is nearly 3 times more to develop depression than not married. Whereas, in this study working unit (isolation center) and with medical illness are not statistically significant variables on affecting the prevalence of depression during the pandemic. Since the findings on the factors associated with the prevalence of depression and insomnia were heterogeneous and limited, it is impossible to pooling. We explore and present these factors systematically as summarized (Table 4).