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

Introduction Healthcare professionals play a great role in the struggle against COVID-19. They are highly susceptible to COVID-19 due to their responsibilities. This susceptibility directly affects their mental health status. Comprehensive evidence on prevalence of depression and insomnia during this pandemic is vital. Thus, this study aims to provide the pooled prevalence of depression and insomnia, and their associated factors during the COVID-19 pandemic. Materials and methods This systematic review and meta-analysis follow the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines. Studies were searched from PubMed, Cochrane Library, CrossRef, African Journals Online and Google Scholar databases from the occurence of the pandemic to June 2022. Study selection, data extraction and methodological quality assessment were done by two authors independently. The I2 statistics was used for testing heterogeneity. A random effect model was used. Stata version 16.0 was used for statistical analysis. Results Eight studies were incorporated for this systematic review and meta-analysis. From seven studies the pooled prevalence of depression was 40% (95% CI [0.23–0.57]; I2 = 99.00%; P = 0.00). From three studies the pooled prevalence of insomnia was 35% (95% CI [0.13–0.58]; I2 = 98.20; P = 0.00). Associated factors of depression on healthcare workers (HCWs) were being female pooled AOR: 2.09; 95% CI [1.41–2.76], been married (pooled AOR = 2.95; 95% CI [1.83–4.07]). Due to limited studies available on the factors associated with insomnia, it is impossible to pool and associated factors were presented in narrative synthesis. Conclusion COVID-19 is highly associated with the prevalence of depression and insomnia among healthcare professionals in Ethiopia. The pooled prevalence of depression and insomnia were significantly higher among healthcare professionals. Appropriate psychological counseling package should be realized for healthcare workers (HCWs) in order to recover the general mental health problems. Trial registration. This review was registered PROSPERO with registration number CRD42022314865.


INTRODUCTION
The pandemic disease caused by severe acute respiratory syndrome coronavirus 2 (SARS CoV-2), first reported by officials in Wuhan City, China, in December 2019, and spread worldwide (World Health Organization, 2020). Healthcare professionals are forefront in the struggle against COVID-19. They are particularly susceptible to this disease due to their clinical task in the healthcare settings. This affects their mental health status. Studies have started investigating the mental health condition during COVID-19 pandemic. Study findings on Spanish healthcare workers (HCWs) imply that COVID-19 has impact on the mental health of HCWs (García-Fernández et al., 2022). In Iran, more than half of the nurses had depression in response to the COVID-19 outbreak (Pourteimour, Yaghmaei & Babamohamadi, 2021). In Africa, the prevalence of depression is higher compared to those reported elsewhere (Chen et al., 2021). The UNICEF situation report points Ethiopia had 96,169 confirmed cases of COVID-19 on 31 October, 2020 about 1,876 healthcare workers had tested positive and 77 had died (UNICEF, 2020).

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. 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 I 2 test. If I 2 > 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).

RESULTS
A PRISMA flow diagram illustrating the steps of data search and refining process for the study on depression and insomnia among healthcare professionals during the COVID-19 pandemic period (Fig. 1). We have got 30 papers from the searched databases. Eight studies duplicated were removed, we examined the titles and abstracts and three papers were removed. By examining the full text, we removed four that did not meet inclusion criteria. Seven Studies were removed due to not reporting about depression and insomnia. Finally, eight studies were relevant to the systematic review and meta-analysis.

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;

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]; I 2 = 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]; I 2 = 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).

DISCUSSION
This study aims to investigate the pooled prevalence and associated factors of depression and insomnia among healthcare professionals during the COVID-19 pandemic in Ethiopia. The result shows a high prevalence depression and insomnia among healthcare professionals during the COVID-19 pandemic in Ethiopia. Consequently, there is a major concern for the mental health of HCWs during the COVID-19 pandemic, as well as in potential future public health crises. There are studies at the single level, but to our knowledge, this systematic review and meta-analysis study is the first of its kind that assessed the pooled prevalence of depression and insomnia and their associated factors.
Among regions, there is heterogeneity on the prevalence of depression. We found the prevalence of depression in Amhara regions is 63%, is higher compared to the others. This might be due to the difference in the availability of prevention equipment's for the COVID-19 and the levels of awareness on the pandemic. Also subgroup analysis by instrument, the pooled prevalence of depression by DASS-21 is 48%, this is higher than study in global 34. 83% (Mahmud et al., 2021) and pooled result by PHQ-9 is 34%, lower than study meta analysis result 38. 11% (Mahmud et al., 2021).
The pooled adjusted odds ratio on female healthcare workers is 2.09, 95% CI [1.41-2.76], implies that odds of female healthcare workers is two times more to develop the depressive symptom than males during the pandemic. This is inline with the study finding in Kenya females HCWs experiencing more symptoms of all the mental health disorders than males (Ali, Shah & Talib, 2021) and Egypt (Elgohary et al., 2021). Similarly for marital status (being married) the pooled adjusted odds ratio is 2.95, 95% CI [1.83-4.07] indicate odds of married healthcare workers are nearly 3 times more to develop depression than non 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 COVID-19 pandemic. The pooled prevalence of insomnia among HCWs during the COVID-19 pandemic was 35%. This is in line with the pooled prevalence of insomnia among healthcare workers in China which is 34.5% (Zhang et al., 2021) and in Kenya which is 37.0% (Ali, Shah & Talib, 2021). On the other hand the prevalence is higher compared to meta analysis results 28% (Chen et al., 2021), and 27.8% (Batra et al., 2020) however lower than the global meta analysis result 43.76% (Mahmud et al., 2021). Up on subgroup analysis by region, the heterogeneity test indicates that there is significant variation in the prevalence of insomnia between regions. The prevalence of insomnia in Addis Ababa is 41% higher than Oromiya. The pooled prevalence of depression measured by ISI is 24%, this is lower than study in China 36.1% (Zhang et al., 2020) similarly using PSQI is 41%, this is higher than study in Kenya 24.2% (Kwobah et al., 2021). The heterogeneity test indicates that there is significant variation on a study finding between measurements. This might be due to the sensitivity and specificity on the measurement tools. For insomnia, pooling the adjusted odds ratio for associated factors was not performed due to the limited data available.
This study is with strengths and some limitations. Study selection, data extraction and quality assessment were performed by two authors independently. Newcastle-Ottawa Scale used to assess the quality of the included studies were the strengths. Whereas, the absence of sufficient studies investigating the impact of COVID-19 on depression and insomnia among healthcare professionals in Ethiopia and heterogeneity among studies were the limitations of this systematic review and meta-analysis.

CONCLUSION
The COVID-19 pandemic caused a variety of mental health impacts among healthcare professionals in Ethiopia. Due to this pandemic, the prevalence of depression and insomnia among healthcare professionals became high in Ethiopia. The prevalence varied among regions according to the instruments used. The suitable programs that offer awareness on the COVID-19 pandemic, psychological counseling and intervention should be implemented for HCWs to improve the general mental health problems including depression and insomnia.

AOR
Adjusted