Prevalence of depressive symptoms and associated factors among people with Tuberculosis in Ethiopia: meta-analysis

Tuberculosis (TB) is a chronic infectious disease caused by mycobacterium tuberculosis. Co-morbid depression among individual with TB is common public health concern that adversely affect the psychological well being of the patient. Thus, the aims of this study were to determine the magnitude of depression and its associated factors among people with TB in Ethiopia. Methods: Databases including: PubMed/Medline, SCOPUS, and HINARI were searched. Weighted inverse variance random-effects model was used to calculate the pooled prevalence of depression. The heterogeneity between studies was measured by the index of heterogeneity (I2 statistics) test. Funnel plots and Egger’s test were used to determine publication bias. Sensitivity test and subgroup analysis were also performed to identify influential study and to account for heterogeneity. Results: In this study, a total of 6430individuals with TB were included from 15 studies. The pooled prevalence of depression was 46% (95% CI; 40% -53%). Tests of heterogeneity was significant (I2= 96.33% and p <0.001). However, funnel plots and Egger’s test showed no evidence of publication bias. Being female, low income, substance use, perceived stigma, extra-pulmonary TB and treatment phase were factors associated with depressive symptoms. Conclusion: In Ethiopia, around half of the TB patients had depressive symptoms. Thus, authors’ suggest the need of mental health screening for people with TB particularly for substance users and socioeconomic disadvantage TB patients. Authors’ also suggest the needs of improving awareness of TB.

publication bias. Being female, low income, substance use, perceived stigma, extrapulmonary TB and treatment phase were factors associated with depressive symptoms.
Conclusion: In Ethiopia, around half of the TB patients had depressive symptoms. Thus, authors' suggest the need of mental health screening for people with TB particularly for substance users and socioeconomic disadvantage TB patients. Authors' also suggest the needs of improving awareness of TB. Background Tuberculosis (TB) is a chronic infectious disease caused by mycobacterium tuberculosis (1). TB affects one-third of the world's population, of which majority are from developing countries like Ethiopia (2, 3).According to WHO 2018 data, Ethiopia was among 3 the top 14 countries with burden of TB, TB-HIV co-infection and multi-drug resistant TB (3).
Mental illness such as depression, anxiety, and somatic symptom are growing global public concerns with an estimated lifetime prevalence of one in three individuals (4).Comorbid mental illness among individual with TB is also another common public concern (5-7).Previous studies revealed that more than 50% of people with TB had one or more of mental illness such as depression, anxiety, somatic symptoms or their combinations (5,7).
Because depression mimics and magnifies the symptoms of TB such as poor appetite, loss of weight and easy fatigability (8). Studies showed that depression is the silent driver of global TB epidemic and responsible for the emergence of Multi Drug Resistant TB (6,8)..
Although, these individual studies are important, the variation in reported prevalence of depression and its determinant factors need comprehensive up to date information.
However, to the authors' knowledge, there is no systematic review and meta-analysis that address this need. Therefore, the present systematic review and meta-analysis aims (i) to determine the pooled magnitude of depression and (ii) to synthesize the factors associated with depression among people with TB in Ethiopia.

Selection of studies
All studies retrieved through search strategy were imported in to EndNote X7 (Thomson Reuters, New York, USA) and duplications were removed. The title and abstracts of the remaining articles were assessed independently by two reviewers (BBB and YDG) and disagreements were resolved discussion.

Eligibility Criteria
Setting: This review included studies carried out in Ethiopia.
Participants: This review included data from people with TB.
Outcome measure: This systematic review and meta-analysis have two main outcomes.
These are (i) the pooled prevalence of depression and (ii) factors associated with depression. The prevalence of depression was calculated by dividing the number of people with co-morbid TB-depression to the total number of people with TB included in the study (total sample size) multiplied by 100. For the associated factors, the reported odds ratio, 95% confidence interval and p-value were used.
Study design: Observational studies (cross-sectional and cohort/longitudinal) were included. Studies that focused on case reports, conference and abstracts were excluded.

Data extraction
Data were extracted from the eligible studies using a pre-conceived data abstraction form by two independent reviewers (BBB and YDG). The extracted data include: name of the first author's last name, year of publication, study area/region, study design, assessment tool, sample size, number of cases/prevalence and associated factors.

Quality assessment
The Newcastle-Ottawa quality assessment tool, adapted for cross-sectional studies (24) was used for quality assessment. This tool had three main parts (selection, comparability and outcome). The first part (selection) has five stars and assesses the methodological quality of the study. The second part of the tool evaluates the comparability of the study.
The third part of the tool assesses the quality of the original article's outcome with respect to the statistical analysis. Individual paper was graded with score ranged from zero to ten stars. The overall quality of each study was determined using the sum score of each star of the three parts. If the overall score of a paper was ≥ 6 out of 10, it is categorized as good quality and if the score fulfilling 50% of quality assessment criteria, it is medium and for score ≤ 4, it is defined as poor quality.

Data synthesis and statistical analysis
The extracted data were entered into a Microsoft Excel Database and then imported into STATA version 14 (Stata Corp LLC, Texas, USA) software with packages of Meta-analysis for analysis. Meta-analysis was performed using STATA 14. "Metaprop" command with random-effects models was used to calculate the pooled effect size at 95% confidence intervals (25) with the Dersimonian and Laird method based transformed values and their variance (26). The Freeman-Tuckey variant of the arcsine square root transformation of proportions was fitted to avoid variance instability when handling proportions close to one (27). Heterogeneity between studies was measured by the index of heterogeneity (I 2 statistics) test (28). I 2 statistics values of 25%, 50% and 75% was used as low, medium and high heterogeneity respectively. Publication bias was examined by the visual inspection of funnel plot (29) and egger test (30) (p-value < 0.05). Sensitivity analyses was conducted to examine influential studies (31). Sub-group analysis was performed by sample size, study design, publication year, quality of studies and study setting. For the analysis of associated factors, the reported odds ratio with 95% confidence interval and pvalue were used. Meta-analysis of the associated factors was performed for those factors with at least two studies reported the same associated factors.

Results
The initial database literature search resulted in 619 published articles. Additional, 23 studies were found from other source and resulted in 642overall articles. Of these articles, 424 were excluded during the initial assessments based on their title and abstract reading and found to be irrelevant. Of the remaining 218 studies, 119 articles were excluded because of different outcome and unclear results. Finally, 84 papers were excluded as results of eligibility criteria. Then, the remained 15 studies were included in the 7 systematic review and meta-analysis (Fig. 1).

Study characteristics
In this study, a total of 6430 study participants were included from four different regions: Addis Ababa (n=5), Oromia (n=5) and SNNP (n=5). Majority (73%) of included studies were cross-sectional and published between 2010 and September 2019. The sample sizes of included studies ranged from 97 to 817, with the response rate of 78-100%. All studies utilized a validated depression assessment tool. Of this, majority (46%) was used PHQ-9 (n=7) and 40% K-10 (n=6) ( Table: 1).

Quality of included studies
The overall quality score of included studies ranged from 4 to 9. Of these, eleven studies had good quality and the remaining four studies had fair quality (Supplementary Table: 1).

Test of heterogeneity, publication bias and subgroup analysis
The tests of heterogeneity demon strated significant heterogeneity (I 2 = 96.33% and p <0.001). But, there was no evidence of publication bias from the visual inspection of the funnel plot (Fig. 2) and Egger's test (P =0.06)]. The results of sensitivity analysis also showed that none of the point estimates was outside of the overall 95% confidence interval confirming that there was no influential study. Thus, in this study, the pooled characteristics of subgroup analysis by study design, sample size, study quality, publication year and study setting were significant, which implies the presence of other characteristics that may contribute for the heterogeneity ( Table 2).

Factors associated with depression
Based on this study, factors such as: socio-demographic factors (sex and income), social factors (perceived stigma), behavioral factors (substance use) and clinical factors (extrapulmonary TB and TB treatment phase) were associated with depression (Table: 3).

Socio-demographic factors and depression:
In this study, the pooled effect of four studies (12,16,19,32) showed that females were around two times (AOR=1.6;95%, 1.10-2.23) more likely experienced depression as compared to males. Moreover, those TBpatients with low socio-economic status or income (9,11,14)were three times (AOR=3.0;95%, 1.16-4.92) more likely to have depression symptoms compared with those with high income category.

Social factors and depression:
In this study, those participant with history of perceived stigma were more than two times (AOR=2.2;95%, 1.67-6.40) (11,12,14, 15, 32) more likely experienced depression compared with those who did not experience perceived stigma.

Behavioral factors and depression:
In this study, use of substance such as: alcohol and cigarette (9,11,12) were the risk factors for depression with the pooled estimate of 3.2 (AOR=3.2; 95%, 1.80-4.45).

Discussion
This systematic review and meta-analysis estimated the epidemiological evidence of depression and its determinant factors among people with Tuberculosis. In this study, the prevalence of depression among people with Tuberculosis was found to be 46% (95% CI; 40% -53%). Even though similar study is scarce, this result is higher than the reported prevalence of depression in the general community of Ethiopia (32). For example, a previous systematic review and meta-analysis using data from 13 studies reported 20.5% (95% CI; 16.5% -24.4%) of depression in the general community (33). This showed that people with tuberculosis is more than two times more likely to suffer with depression as compared to the general population. In line with the present study, a comparative community based study using data from 48 low-and middle-income countries revealed that higher prevalence of depressive episode among those people with tuberculosis as compared to the general population (5).The increased prevalence of depression among people with TB may be due to the negative psycho-social effect of TB and its associated drug side effect (citation). In consistent with our study, another similar systematic review and meta-analysis among people with TB using data from 31 studies in 11 different countries reported a prevalence of 48.9% (95% CI 48.3%-49.6%) (8). Another metaanalysis also reported a pooled prevalence of 50% (19% -81%) using data of base line report (34).
In Ethiopia, there is a strong political commitment to control TB and the burden of mental illness, and result in decreasing trend in TB mortality and improved experiences in implementation of mental health integration; however the priority focus is given to curative services rather than mental, neurological and substance use disorders.
In this study, we found that being female was around two times (AOR=1.6; 95%, 1.10-2.23) more likely experienced depression as compared to male. This is consistent with previous studies. Similarly, those individual with low socio-economic status were three times (AOR=3; 95%, 1.16-4.92) more likely reported depression. This may be associated with individual's inability to fulfill their needs like food, drug and other, which is the cause for stress that in turn to depression. Studies showed that people living in low socioeconomic status have a greater risk of exposure to precipitating factors of depression such as violence, stigma, discrimination and lacking adequate health care information or education (6,8).
In this study, those individual with history of perceived stigma were more likely experience depression. This may be due to the fear associated with the knowledge of TB transmission, that in turn for self-stigma which is an important risk for depression. In this study, use of substance was associated with depression. This may be due to people's use of substance as a coping mechanism of depression and TB burden like poverty. Type of TB (extra pulmonary) and intensive phase of TB-treatment were also associated with depression. This may be related to individual's perceived severity of the illness and anti Tb drug side effect. Evidence showed that The associations of several treatment drug with psychiatric illness such as: cycloserine (CS), isoniazid (INH), ethionamide, ciprofloxacin, ethambutol, rifampicin (RMP), and the fluoroquino-lones (5,6,35). That is why the WHO End TB Strategy 2015-35 calls for the integration TB-mental health treatment for the management of co-morbid psychological distress (36). Thus, the results of the preset study imply high magnitude of depression among people with TB. Moreover, this study also implies the adverse effect of TB on an individuals' socio-economic status, behavior like substance use and individuals' perception of the illness for the psychological well being TB patients. This may suggest the needs of evaluating the effectiveness of health care system in addressing the comorbid mental illness.

Strengths and limitations of the study
To our knowledge, this is the first systematic review and meta-analysis about the prevalence of alcohol use and its associated factors among Ethiopia pregnant women. To manage heterogeneity and bias; First, we used appropriate model (a random effect model) to control the effects of the observed heterogeneity. Secondly, we conducted a leave one out sensitivity analysis to identify influential study. Third, we conducted sub-group analysis to account for heterogeneity. Forth, funnel plot and egger test was performed to identify small study effect. However, limitations such as: use of reference lists and Google Scholar may have the possibility of having some overlooked articles.

Conclusion
In Ethiopia, around half of people with TB had an experience of depressive symptom.
Being female, low income, substance use, perceived stigma, extra-pulmonary TB and treatment phase were factors associated with depressive symptom. Thus, authors' suggest the need of screening people with TB particularly for those with low income, negative perceptions of TB and history of substance use during the initial phase of the treatment.
Authors' also suggest the needs of improving awareness of TB.

Ethics approval and consent to participants
Not applicable.

Consent for publication
Not applicable.

Availability of data
Not applicable.

Competing interests
None declared.

Funding
This research received no grant from any funding agency in the public, commercial or notfor-profit sectors.

Authors' contributions
BBB designed the systematic review and meta-analysis in collaboration with BAD, BTT and YDG. BBB developed the search strategy and drafted the protocol. BAD, BTT and YDG improved the drafted systematic review and meta-analysis. BBB, YDG and BAD provided their expertise to the section of mental illness (depression) and methodological section.
BBB and YDG performed search strategy and conducted data selection and extraction. All authors were involved in data analysis and interpretation of the results and write up the manuscript. All authors have read and approved the final manuscript.    Forest plot presenting the prevalence of depression using random effect models with 95% CI.
23 Figure 3 Funnel plot with pseudo 95 % confidence interval that investigated the publication bias of the pooled prevalence of depressive symptoms.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.