Healthcare-associated infection and its determinants in Ethiopia: A systematic review and meta-analysis.

BACKGROUND: Healthcare-associated infection is a global threat in healthcare which increases the emergence of multiple drug-resistant microbial infections. Hence, continuous surveillance data is required before or after patient discharge from health institutions though such data is scarce in developing countries. Similarly, ongoing infection surveillance data are not available in Ethiopia. However, various primary studies conducted in the country showed different magnitude and determinants of healthcare-associated infection from 1983 to 2017. Therefore, this systematic review and meta-analysis aimed to estimate the national pooled prevalence and determinants of healthcare-associated infection in Ethiopia.
METHODS: We searched PubMed, Science Direct, Google Scholar, and grey literature deposited at Addis Ababa University online repository. The quality of studies was checked using Joanna Brigg's Institute quality assessment scale. Then, the funnel plot and Egger's regression test were used to assess publication bias. The pooled prevalence of healthcare-associated infection was estimated using a weighted-inverse random-effects model meta-analysis. Finally, the subgroup analysis was done to resolve the cause of statistical heterogeneity.
RESULTS: A total of 19 studies that satisfy the quality assessment criteria were considered in the final meta-analysis. The pooled prevalence of healthcare-associated infection in Ethiopia as estimated from 18 studies was 16.96% (95% CI: 14.10%-19.82%). In the subgroup analysis, the highest prevalence of healthcare-associated infection was in the intensive care unit 25.8% (95% CI: 3.55%-40.06%) followed by pediatrics ward 24.16% (95% CI: 12.76%-35.57%), surgical ward 23.78% (95% CI: 18.87%-29.69%) and obstetrics ward 22.25% (95% CI: 19.71%-24.80%). The pooled effect of two or more studies in this meta-analysis also showed that patients who had surgical procedures (AOR = 3.37; 95% CI: 1.85-4.89) and underlying non-communicable disease (AOR = 2.81; 95% CI: 1.39-4.22) were at increased risk of healthcare-associated infection.
CONCLUSIONS: The nationwide prevalence of healthcare-associated infection has remained a problem of public health importance in Ethiopia. The highest prevalence was observed in intensive care units followed by the pediatric ward, surgical ward and obstetrics ward. Thus, policymakers and program officers should give due emphasis on healthcare-associated infection preventive strategies at all levels. Essentially, the existing infection prevention and control practices in Ethiopia should be strengthened with special emphasis for patients admitted to intensive care units. Moreover, patients who had surgical procedures and underlying non-communicable diseases should be given more due attention.

Introduction

According to the Communicable Diseases Control (CDC), healthcare-associated infection (HCAI) is defined as the acquisition of infectious agent(s) or its toxin(s) which occurs after 48 hours of hospital admission, or up to 3 days after discharge, or up to 30 days after the operation when someone was admitted for reasons other than infection [1, 2].

Globally, according to the World Health Organization (WHO) 2019 HCAI fact sheet report, a hundred million patients were affected each year [3]. The point prevalence of HCAI ranged from 3.5%-12% and 5.7%-19.1% in developed and Low-and Middle-Income Countries (LMICs), respectively [3, 4]. Though data is scarce, the burden of HCAI was found to be high in Sub-Saharan Africa (SSA) countries [5]. Specifically, the prevalence of HCAI was noted in Botswana (13.4%) [6], South Africa (8%) [7], and Ethiopia (13% to 22%) [8-10].

Healthcare-associated infection increases the occurrence of antimicrobial resistance [11], long-term disability [4], and mortality among individual patients [12]. The additional financial burden to the healthcare system, patients, and families due to HCAI is also significant [4]. Hence, the “clean care the safer care” program has been launched in 2004 with the WHO patient safety directive, which was aimed to reduce HCAI through improving hand hygiene practice at the center of achieving its aim [13]. The aforementioned infection prevention program and the WHO initiative about infection prevention and control policy recommendations have been implemented in developing countries, including Ethiopia. Despite these efforts, studies conducted at different settings of the globe revealed that admission to the surgical ward and hospital type [8], chest tube placement, prolonged hospital stays, patient on mechanical ventilation, previous hospitalization [9], pediatric patients, malnutrition, and length of staying in hospital >5days [10] were contributing factors of HCAI.

Various studies were conducted to determine the prevalence of HCAI in Ethiopia, but it showed great variation across geographical setting and variant periods. Based on this fact, there was a need for nationally representative data on HCAI in the country. Moreover, the pooled effect sizes of the determinants of HCAI weren’t explored nationwide. Consequently, this systematic review and meta-analysis was aimed to address the following research questions: (1) what is the national pooled prevalence of HCAI in Ethiopia; and (2) what are the determinants of HCAI in the country?

Materials and methods

Reporting

The study results were reported based on the Preferred Reporting Items for Systematic Review and Meta-analysis statement (PRISMA) guideline [14] (S1 File). The protocol was registered on the PROSPERO database with a registration number (CRD42020166761), and available on https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020166761.

Inclusion and exclusion criteria

We included cross-sectional, case-control and cohort studies, but case-control studies weren’t used to estimate the pooled prevalence of HCAI. These studies were included when the prevalence, incidence, and/or at least one determinant was reported. All studies published in the English language were considered. There was no restriction of the study period, age group, and study setting. All citations without abstract and/or full-text, anonymous reports, editorials, and qualitative studies were excluded.

Search strategy and information source

PubMed, Science Direct, Google Scholar, and grey literature deposited at Addis Ababa University online repository were searched. The core search terms and phrases were “prevalence”, “incidence”, “epidemiology”, “proportion”, “magnitude”, “burden”, “associated factors”, “risk factors”, “predictors”, “determinants”, “healthcare-associated infections”, “healthcare-acquired infections”, and “nosocomial infections”, “hospital acquired infections” and “Ethiopia”. The search strategies were developed using different Boolean operators. Notably, to fit the advanced PubMed database, the following search strategy was applied: [(prevalence) OR incidence[MeSH Terms]) OR epidemiology[MeSH Terms]) OR proportion[MeSH Terms]) OR magnitude[MeSH Terms]) OR burden[MeSH Terms]) AND associated factors) OR risk factors[MeSH Terms]) OR predictors[MeSH Terms]) OR determinants[MeSH Terms]) AND healthcare-associated infections) OR healthcare-acquired infections[MeSH Terms]) OR nosocomial infections[MeSH Terms]) OR hospital acquired infections[MeSH Terms]) AND (Ethiopia)]. Then, we retrieved 611 articles using this PubMed searching strategy.

Study selection

Duplicate studies were removed using Endnote version 8 (Thomson Reuters, London) reference manager software. The two independent reviewers (AYA and WAB) screened the titles and abstracts. The disagreements were handled based on established article selection criteria. Then, two independent authors (AE and BMB) conducted the abstracts and full-texts review.

Quality assessment

The two independent authors (DKM and DMB) appraised the quality of the studies. The Joanna Briggs Institute (JBI) quality appraisal checklist was used [15]. The disagreement was resolved by the involvement of a third reviewer (AE). To appraise cohort studies, the following items were used: (i) similarity of groups;(ii) similarity of exposure measurement;(iii) validity and reliability of measurement;(iv) identification of confounder;(v) strategies to deal with confounder;(vi) appropriateness of groups/participants at the start of the study;(vii) validity and reliability of outcome measured;(viii) sufficiency of follow-up time;(ix) completeness of follow-up or descriptions of reason to loss to follow-up;(x) strategies to address incomplete follow-up; and (xi) appropriateness of statistical analysis. The items used to appraise case-control studies were: (i) comparable groups;(ii) appropriateness of cases and controls;(iii) criteria to identify cases and controls;(iv) standard measurement of exposure;(v) similarity in the measurement of exposure for cases and controls; (vi) handling of confounders;(vii) strategies to handle confounder;(viii) standard assessment of outcome;(ix) appropriateness of duration for exposure; and (x) appropriateness of statistical analysis. Cross-sectional studies were appraised based on (i) inclusion criteria;(ii) description of study subject and setting;(iii) valid and reliable measurement of exposure; (iv) the objective and standard criteria used;(v) identification of confounder;(vi) the strategies to handle confounder; (vii) outcome measurement; and (viii) appropriate statistical analysis. All the studies which got 50% and above on the quality assessment scale were considered as low risk.

Data extraction

Two independent reviewers (AYA and AE) extracted data using a structured data extraction form. Whenever variations of extracted data were observed, the phases were repeated. If discrepancies between data extractors continued, the third reviewer (WAB) was involved. The name of the first author and year, study region, study design, target population, diagnostic methods, sample size, the prevalence of HCAI, and adjusted odds ratio (AOR) of associated factors were collected.

Outcome measurement

HCAI was considered when reported as infection(s) acquired while receiving medical care based on culture-confirmation [10, 16–18], or clinical and laboratory methods [8, 9, 19–31].

Statistical analysis

Publication bias was checked visually by the funnel plot, and objectively using Egger’s regression test [32]. Heterogeneity of studies was quantified using the I-squared statistic, in which 25%, 50%, and 75% represented low, moderate, and high heterogeneity, respectively [33]. Pooled analysis was conducted using a weighted-inverse variance random-effects model [34]. The subgroup analysis was done by region, study design, diagnostic method, sample size and ward type. Sensitivity analysis was employed to see the effect of a single study on the overall estimation. Besides, the time-trend analysis was conducted to check the variation through time. STATA version 11 statistical software was used for meta-analysis.

Ethics approval and consent to participate

Not applicable because no primary data were collected from patients.

Results

Literature search

The search strategy retrieved 611 articles from PubMed, 133 from Science Direct, 19 from Google Scholar, and 3 grey literatures from Addis Ababa University online repository. After duplicates were removed, 740 studies remained. Then, sixty studies were screened for full-text review. Finally, 19 studies were used in the systematic review and/or meta-analysis ().

Characteristics of the included studies

Six studies were found in Addis Ababa [19–21, 23, 25, 31], five studies in Amhara region [8, 16, 26, 27, 29], five studies in Oromia [9, 17, 18, 28, 30], one study both in Addis Ababa and Southern Nation Nationalities and People Region (SNNPR) [22], one each in Tigray [24] and SNNPR [10]. Nine studies were conducted across all age groups. Eight studies were done on the adult population and one study was on pediatric patients. Fourteen studies used clinical and laboratory methods for the diagnosis of HCAI while the remaining were culture-confirmed. Four studies were conducted using cohort study design, fourteen were cross-sectional and only one was a case-control study. Only six studies had >1000 sample size ().

Quality of studies

The JBI quality appraisal criteria established for cross-sectional, case-control, and cohort studies were used. The studies included in this systematic review and meta-analysis had no considerable risk. Therefore, all the studies were considered [8–10, 16–31] ().

Meta-analysis

Publication bias

The funnel plot showed symmetrical distribution (). Egger’s regression test p-value was 0.328, which indicated the absence of publication bias.

Funnel plot for publication bias, LN of proportion (X-axis) with its standard error of LN of proportion (Y-axis).

The prevalence of healthcare-associated infection

A total of 18 studies were used and 14,240 patients participated in the prevalence estimation. The estimated overall prevalence of HCAI is presented in a forest plot (). The overall prevalence of HCAI was 16.96% (95% confidence interval (CI): 14.10%-19.82%).

Subgroup analysis

The subgroup analyses based on study region, study design, diagnostic method, and the sample size were done. Accordingly, the prevalence of HCAI was found 27.6% in Tigray region, 18.2% diagnosed by clinical and laboratory methods, 17.83% in the cross-sectional studies, 18.15% in studies using < 1000 study samples ().

The prevalence of HCAI was reported across various wards too. A study conducted at Jimma University Hospital showed that the incidence of HCAI was the highest in the Intensive Care Unit (ICU) (207.55/1000 patient-days) followed by the pediatric ward (69.16/1000 patient-days), and surgical ward (28.87/1000 patient-days) [9]. In two studies, Yallew WW. et al. [8] and Ali S. et al. [9], HCAI was the lowest in the ophthalmology ward.

Besides, in this meta-analysis, HCAI was estimated in different wards based on the pooled effect of two or more studies. As estimated from the effect of two studies [25, 30], the prevalence of HCAI was the highest in ICU (25.8%) followed by pediatrics (24.16%) [8, 10, 31], surgical (23.78%) [8, 19, 21, 23, 25, 26, 28, 30, 31] and obstetrics ward (22.25%) [19, 26] ().

Sensitivity analysis

The studies of Endalafer N. et al. [25] and Sahile T. et al. [28] had shown an impact on the overall estimate of HCAI ().

Time-trend analysis

The time-trend analysis showed that the prevalence of HCAI was increased from 13.4% in 1983 to 19.8% in 2017. However, the pooled prevalence was not increasing significantly from year to year (p-value: 0.620) ().

Determinants of healthcare-associated infection

In this systematic review and meta-analysis, HCAI in the Ethiopian context is associated with socio-demographic, patient health condition, and healthcare-related risk factors. Thus, based on the report of a single study, the age range of the patient 18–30 years was found to be protective (AOR = 0.54; 95% CI: 0.22–0.85) [9] (). On the contrary, based on the reports of individual studies included, HCAI had shown a positive association with the following healthcare-related factors: taking prophylaxis (AOR = 1.76; 95% CI: 1.21–2.3) [27]; admission to the surgical ward (AOR = 2.86; 95 CI: 1.33–4.38) [8]; admission at Felege Hiwot Referal Hospital (FHRH) (AOR = 1.99; 95% CI: 1.2–2.77) [8] and chest tube insertion (AOR = 4.14; 95% CI: 1.57–6.71) [9].

Moreover, in this meta-analysis, the determinants of HCAI were identified based on the pooled effect of two or more studies. Hence, as estimated from the pooled effect of two studies [25, 29], HCAI was 3.37 times (AOR = 3.37; 95% CI: 1.85–4.89) more likely among patients who had the surgical procedure as compared to no surgical procedure. Similarly, based on the pooled effect of three studies [9, 16, 27], patients who had underlying non-communicable disease were 2.81 times more likely to have HCAI as compared to those without the underlying disease ().

Discussion

In this systematic review and meta-analysis, the pooled prevalence of HCAI was 16.96% in Ethiopia. The authors also found that surgical procedures and underlying non-communicable diseases were identified as determinants of HCAI.

From the study, the national pooled prevalence of HCAI in Ethiopia was 16.96% (95% CI: 14.10%-19.82%). The result was higher than studies conducted in China (3.12%) [35], Morocco (10.3%) [36], Botswana (13.54%) [6], and South Africa (7.67%) [7]. The possible reasons for high prevalence in this study might be very low hand hygiene practice by physicians and resource constraints [37], low adherence to infection prevention practice [38], low level of job satisfaction [39], morally distressed nurses [40], and low implementation of the nursing process [41] in our settings, and also less attention given to HCAI.

Regarding hand hygiene, only 7% of the physicians working at two University hospitals in the capital of Ethiopia, Addis Ababa, performed hand hygiene before patient contact [37]. As a result, the acquisition of HCAI from these healthcare professionals might be high. Evidence also showed that 35% of the nurses in southwest Ethiopia were non-adherent to infection prevention practice [38], thereby contributing to high HCAI in Ethiopia. Besides, nearly 68% of the health professionals were less satisfied with their work in one of the regions in the country [39]. Hence, the nosocomial infection becomes inevitably high because these less satisfied health professionals are less likely to deliver quality healthcare. Additionally, about 84% of the nurses in the northwestern part of the country [40] were morally distressed, thus causing HCAI as morally less prepared nurses were unable to deliver quality nursing care.

Resource constraints could also increase HCAI in the country because lack of hand hygiene agents and sinks were reported as hindering factors of infection prevention practice in Addis Ababa, Ethiopia [37]. Implementation of the nursing process was below half (49%) in the northwest part of the country [41], so nursing intervention would not be planed for patients at risk of nosocomial infection. Furthermore, healthcare providers, patients, and/or families are more curious about the primary reason for admission or healthcare visits, so less attention is given to HCAI.

From the subgroup analysis, HCAI was found the highest in ICU (26%). This finding is consistent with studies conducted in China [42], India [43], and Morocco [36]. The reasons for high HCAI in ICU may be due to the highest incidence of HCAI, the severity of the disease, and highly invasive procedures. The incidence of HCAI at a referral medical center in Jimma University Hospital, Ethiopia was 207.6/1000 patient-days [9]. This highest incidence may be augmented by the severity of the disease [36] among ICU patients. Added, highly invasive procedures like intubation, peripheral, and central venous catheters are highly likely among ICU patients. Consequently, the risk of HCAI is higher among intubated patients and those on vascular catheterization [44].

In the time trend analysis, we found that HCAI was slightly increasing in Ethiopia from 1983 to 2017. The possible reasons might be more emphasis given on healthcare coverage than quality, increase in technological advancement, and overutilization of invasive procedures. Evidence revealed that advances in life-saving medical practices increase exposure to invasive procedures which increase the occurrences of nosocomial infections [11]. On top of this, nurses’ burnout might contribute to the increasing trend. Evidence in the United States (US) revealed that nurses’ burnout was found as a single most important associated factor for increased nosocomial urinary tract infection (UTI) and surgical site infection (SSI) [45].

The current systematic review and meta-analysis revealed surgical procedure and underlying non-communicable disease as determinants of HCAI. Accordingly, patients who had the surgical procedure were 3.37 times more likely to acquire HCAI as compared to patients who have no surgical procedure. The finding is in line with previous studies done in South Africa [7] and Poland [44]. The reason for the observed association could be explained by less compliance to hand hygiene practice and high prevalence of surgical site infection (25.22%) in Ethiopia [46]. Compliance with hand hygiene practice is pivotal for the prevention and control of nosocomial infection, but only 3.6% keep hand hygiene before performing aseptic procedures at Debre Birhan referral hospital, central Ethiopia [47]. In 2017, the overall compliance to hand hygiene practice was 18.7% [48] and 22% [47] at Hiwot Fana Specialized Hospital and Debre Birhan Referral Hospital in Ethiopia, respectively. Consequently, unable to keep and maintain hand hygiene practice increased the acquisition of HCAI.

In this systematic review and meta-analysis, the odds of having HCAI among patients who have underlying non-communicable disease were nearly 3 times higher than their counter-part. This finding is supported by studies that reported positive association of HCAI with diabetes mellitus [49] and underlying renal disease [50]. The possible explanations for the observed association in the current study might be due to high prevalence of underlying diseases and the immune-suppressive effects of these diseases. In Ethiopia, a meta-analysis of studies showed high burden (6.5%) of diabetes mellitus [51]. Besides, another meta-analysis in other setting (North America, Europe, Latin America, and the Caribbean) reported immune-suppression as risk factor for HCAI [49]. Thus, higher odds of HCAI may be ascertained to the aforementioned antecedents.

Strengths and limitations of the study

This systematic review and meta-analysis was the first national report on the prevalence of HCAI and its determinants in Ethiopia. However, it may lack national representativeness because no data were found from Benishangul Gumuz, Afar, Gambella, Somalia, Dire Dawa, and Harari regions of the country. Besides, the use of only English language, the absence of grey pieces of literature, and the unlimited time-period for the inclusion of studies may limit the conclusiveness of the finding. On top of this, the time-trend analysis might not reflect the exact trend because all the considered years didn’t have reported data.

Conclusions

The prevalence of healthcare-associated infection has remained a problem of public health importance in Ethiopia. Based on the subgroup analysis, the highest prevalence of HCAI was found in ICU followed by pediatrics, surgical, and obstetrics wards in descending order. Surgical procedures and underlying non-communicable disease were found as determinants of HCAI. Therefore, policy-makers and program officers should give due emphasis to the prevention of healthcare-associated infection with more attention for patients admitted to ICU. Furthermore, the existing infection prevention and control practices for patients who had surgical procedures and underlying non-communicable disease should be strengthened in Ethiopia.

PRISMA checklist.

(DOC)

Click here for additional data file.

24 Sep 2020

PONE-D-20-09587

Prevalence and risk factors of healthcare-associated infection in Ethiopia: a systematic review and meta-analysis

PLOS ONE

Dear Abebaw Alemu Yeshambel,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by 30th November 2020. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Professor Kwasi Torpey, MD PhD MPH

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. At this time we ask that you include in your manuscript an explanation for the pooling of prevalence data from case-control studies and cross-sectional studies.

3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Additional Editor Comments (if provided):

The manuscript titled " Prevalence of risk factors of health associated infections in Ethiopia: a systematic review and meta-analysis" is an important addition in understanding the factors associated with health associated infections within the Ethiopian context. Though the manuscript makes an important contribution it is not publishable in its current form as a result numerous language errors. The manuscript should be reviewed by a native speaker and copyedited thoroughly.

1. Below are some examples (not exhaustive)

a. Abstract Background: emerging of multidrug resistance microbial infection............. emerging should read emergence

b. Following sentence on survey or surveillance should be revised for clarity

c sentence with varied instead of varying

d. Methods: Addis Ababa university: University must start with a capital letter

e. 14,240 were participated for prevalence estimates - Correct language error. participants?

f. Last paragraph on background : Varieties of studies can read as a number of studies

Study selection

g. Screening of title and abstract: This should be plural

h. Quality Assessment: Disagreement was resolved by interference by third reviewer. Appropriate word could be involvement, assessed etc

plus many more

2. a. Reference #9 : remove CAPS

b. #20 - Check

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: No

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Title : Prevalence and risk factors of healthcare-associated infection in Ethiopia: a systematic review and meta-analysis

General Comments

o It will be good express why the research focused only “patients because “Healthcare-associated infection” include acquired infection in the health care set up for patients, healthcare workers, and visitors.

o It would be good to see subgroup analysis and observe the prevalence of HCAI in each ward by observing the I2 , rather than “region, study design, diagnostic method and sample size” this may give good insight for the reader as well as specific decision actions?

Specific comments

� Title:

� Abstract:

1. in the abstract section “Addis Ababa university repository.” It is not a research data base but it is a student’s thesis data base, better to use as unpublished grey literature source rather than consider as a scientific repository site?

2. Is it possible to say “The national prevalence of healthcare-associated infection” for this pooled result because, the exposure of Healthcare workers and patients vary from type of facility and vary ward to ward? Did you get similar population to pool the prevalence of HCAI in Ethiopia, for example it may vary , ICU to ophthalmology and Teaching hospital and district hospital in Ethiopian condition?

� Method:

o Is there any reason “There was no restriction of the study period” for this dynamic scientific world, don’t you consider the implementation of IP programs procedural and guidelines change have impact the prevalence of Health care associated infections in Ethiopia? For example, Infection guideline development since 2004 GC has an impact in Ethiopia to manage IP practices? How do you see this?

o “Not applicable because no primary data were collected. Did the protocol published? , why not ?

“

� Result:

o Subgroup analysis section: it is good to conduct and observe subgroup analysis?

o For risk factors pooled value what was the measure of effect? Is that OR or RR, , don’t you think papers must be similar?

� Discussion:

o Start the discussion with one summary paragraph.

o In the third paragraph of the discussion section “ The possible reasons for high prevalence in this study might be very low hand hygiene practice by physicians,” it needs a reference? May be should be supported by evidence?

� Strength and limitation

Some of the limitation that I observed in this manuscript

- 1. Only one single language English for the inclusion criteria

- 2. Absence of Grey literature in the pooled result

- 3. Absence of time period for the inclusion criteria?

Reviewer #2: Prevalence and risk factors of healthcare-associated infection in Ethiopia: a systematic review and meta-analysis

This study is supposed to be an important work considering the escalation of healthcare-associated infections globally, however, the major drawback noticed is the lack of the systematic review of the identified studies. There is an obvious lack of information on healthcare-associated infections in the manuscript vis-a viz the different types and how the integration of these was done to culminate into meta-analysis.

Please, see comments in the reviewed submitted attachment.

All the best.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: walelegn worku yallew

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: PONE-D-20-09587_reviewer submission.pdf

Click here for additional data file.

6 Oct 2020

Point by a point response letter

Dear Academic Editor (Professor Kwasi Torpey, MD Ph.D. MPH), Reviewer #1, and Reviewer #2

After going through the entire manuscript, you forwarded your constructive comments which we missed to touch. Therefore, we are glad enough to express our sincerest thanks for your constructive editorial comments that could help to improve the novelty of our effort.

Editors general comments

Comment: 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Author response: Yes indeed, accessing the PLOS ONE style templates from the given links, our manuscript has been made to meet PLOS ONE's style requirements, including those for file naming. These changes were made to meet PLOS ONE's style requirements and found throughout the revised version of the manuscript.

Comment: 2. At this time we ask that you include in your manuscript an explanation for the pooling of prevalence data from case-control studies and cross-sectional studies.

Author response: Yes Sure! We use case-control and cross-sectional studies together to estimate the pooled effect of the determinants of healthcare-associated infection. But, we did not pool the prevalence from case-control and cross-sectional studies together unless specified. Nonetheless, the authors hadn’t encountered a prevalence report from a case-control study, as depicted in Table 1 the prevalence estimate of a pocket study was not presented under the prevalence column for a case-control study included. Now we include a statement that specifies the purpose of including case-control studies in the methods section under the sub-heading inclusion and exclusion criteria.

Comment: 3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information

Author response: Thank you. We include captions for supporting information files and update the in-text citations accordingly. You can find the caption at the end of the manuscript file following the figures caption.

Additional Editor Comments (if provided):

Comment: The manuscript titled " Prevalence of risk factors of health-associated infections in Ethiopia: a systematic review and meta-analysis" is an important addition in understanding the factors associated with healthcare-associated infections within the Ethiopian context. The manuscript should be reviewed by a native speaker and copyedited thoroughly.

1. Below are some examples (not exhaustive)

a. Abstract Background: emerging of multidrug resistance microbial infection............. emerging should read emergence

b. Following sentence on survey or surveillance should be revised for clarity

c sentence with varied instead of varying

d. Methods: Addis Ababa university: University must start with a capital letter

e. 14,240 were participated for prevalence estimates - Correct language error. participants?

f. Last paragraph on background : Varieties of studies can read as a number of studies

Study selection

g. Screening of title and abstract: This should be plural

h. Quality Assessment: Disagreement was resolved by interference by third reviewer.

Appropriate word could be involvement, assessed etc

plus many more

2. a. Reference #9 : remove CAPS

b. #20 – Check

Authors’ response: Sure! We have tried our best to address the comments from the reviewers. Moreover, from repeated proof-reading of the manuscript, we found several grammatical errors, interlinings, police titles, punctuation errors, wordings, and spelling errors. Therefore, finding our colleague who has a Master of Arts in English, we have done our best to thoroughly copyedit the manuscript for English language usage. These editorial changes are found throughout the revised version manuscript.

Reviewers' general comments:

Reviewer's Responses to Questions

Comments to the Author

Comment: 1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: No

Authors’ response: The design of our research is systematic review and meta-analysis. Based on the quality and strength level of information, evidence from systematic review and meta-analysis is the strongest. The finding is reported based on PRISMA guidelines and the sample size for the current study is high. ________________________________________

Comment: 2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Authors’ response: Thank you for your appreciation! ________________________________________

Comment: 3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Author response: Thank you indeed!

________________________________________

Comment: 4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: No

Authors’ response: Sure! We have tried our best to address the comments from both reviewers. Moreover, from repeated proof-reading of the manuscript, we found several grammatical errors, interlinings, police titles, punctuation errors, wordings, and spelling errors. Therefore, finding our colleague who has a Master of Arts in English, we have done our best to thoroughly copyedit the manuscript for English language usage. These editorial changes are found throughout the revised version manuscript.________________________________________

Review Comments to the Author

Reviewer #1: Comments

General Comments

Comment: It will be good express why the research focused only “patients because “Healthcare-associated infection” include acquired infection in the health care set up for patients, healthcare workers, and visitors.

Authors’ response: Firstly, the meta-analysis is depending on the original studies report. Secondly, whoever he/she is, once they acquired infection in healthcare settings, he/she is considered and treated as a patient.

Comment: It would be good to see subgroup analysis and observe the prevalence of HCAI in each ward by observing the I-square, rather than “region, study design, diagnostic method and sample size” this may give good insight for the reader as well as specific decision actions?

Authors’ response: Yes sure! As per your comment, we do a subgroup analysis based on ward type and include in the revised version of the manuscript, and we found important findings indicating the highest (25.8%) prevalence of HCAI in the intensive care unit (ICU) and pediatrics ward (24.16%). The revision is included in Figure 4 in the results section. You can appreciate the result in the track change version of the manuscript.

Specific comments

Comment: 1. in the abstract section “Addis Ababa university repository.” It is not a research data base but it is a student’s thesis data base, better to use as unpublished grey literature source rather than consider as a scientific repository site?

Authors’ response: Thank you. In fact, Addis Ababa University’s online research repository is not considered as a database as PubMed and other databases. Unpublished works of its own students’ research are deposited to the University online repository. So, we correct the statement as per your comment you can find it as track change on the revised document in the abstract and methods section.

Comment: 2. Is it possible to say “The national prevalence of healthcare-associated infection” for this pooled result because, the exposure of Healthcare workers and patients vary from type of facility and vary ward to ward? Did you get similar population to pool the prevalence of HCAI in Ethiopia, for example it may vary, ICU to ophthalmology and Teaching hospital and district hospital in Ethiopian condition?

Authors’ response: One of the reasons to do systematic review and meta-analysis is the absence of pooled evidence for a country. Different fragmented studies found in different areas could not represent the national level. However, pooling these primary studies could represent the national level. By considering its strength and limitation, the current meta-analysis could be reported as being national prevalence because no data and/or other meta-analysis studies at the national level were published before ours. Besides, as per your comment subgroup analysis is done by ward type and the result is incorporated in the results section under subheading subgroup analysis as Figure 4.

Regarding variations from hospital to hospital, we have found the studies were done at referral and above level hospitals so we were unable to compare the varaiations across district and referral hospitals.

� Method:

Comment: Is there any reason “There was no restriction of the study period” for this dynamic scientific world, don’t you consider the implementation of IP programs procedural and guidelines change have impact the prevalence of Health care associated infections in Ethiopia? For example, Infection guideline development since 2004 GC has an impact in Ethiopia to manage IP practices? How do you see this?

Authors’ response: Usually, the year of publication could not be limited in case of systematic review and meta-analysis. It is known that, the variation of reports between studies conducted at different time periods and across geographical locations. There may be different updates throughout a different time period. There may be an impact due to the involvement of guidelines, policy change, human resources increment, or else. With the presence of such kinds, it is recommended to pool all the available studies to have pooled data. And, though it is not statistically significant the trend analysis showed an increase in the trend of HCAI in Ethiopia as incorporated in the original manuscript in Figure 6.

Comment: “Not applicable because no primary data were collected. Did the protocol published? , why not ?

Author response: Thank you indeed! We include the above statement under ethics approval and consent to participate section of the manuscript in the methods section because no primary data is collected from patients. However, currently, the protocol is registered in the PROSPERO database with a registration number CRD4202016676, and available online:https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020166761

Result:

Comment: Subgroup analysis section: it is good to conduct and observe subgroup analysis?

Author response: Thank you. Yes, Sure! As per your comment, we add the subgroup analysis based on the ward type. You can see Figure 4 in the revised version of the manuscript at the results section under the subheading subgroup analysis.

Comment: For risk factors pooled value what was the measure of effect? Is that OR or RR, don’t you think papers must be similar?

Authors’ response: For this review, we found only 3 studies with cohort study design, and the other is done through a cross-sectional study design. Of these, only one study’s RR pooled with OR. We may combine anything that's trying to estimate the same thing if we have the estimate and standard error, using the inverse-variance method. The question is whether it makes sense to pool them. We may have RR and OR, then assuming the OR is a good approximation to the RR in our case. Thus, we theoretically combine them. Moreover, when the outcomes are rare, or in conducting a nested case-control study then these are approximately equal and can readily be combined. Some researchers decide these are similar enough to combine; others do not. The judgment of the meta-analyst in the context of the aims of the meta-analysis will be required to make such decisions on a case by case basis. OR, RR and HR are all measures of relative risks. Thus, combining OR & RR is sometimes acceptable

� Discussion:

Comment: Start the discussion with one summary paragraph.

Authors’ response: Thank you and we do it now. The changes made could be appreciated in the first paragraph of the discussion section.

Comment: In the third paragraph of the discussion section “The possible reasons for high prevalence in this study might be very low hand hygiene practice by physicians,” it needs a reference? May be should be supported by evidence?

Authors’ response: It is now supported by evidence and updates as per your comment.

Strength and limitation

Comment: Some of the limitation that I observed in this manuscript

- 1. Only one single language English for the inclusion criteria

- 2. Absence of Grey literature in the pooled result

- 3. Absence of time period for the inclusion criteria?

Authors’ response: Thank you! With due respect, we consider these as limitations too. It can be depicted in the strength and limitations section of the revised manuscript in track changes.

Reviewer #2: Comments

Comment: This study is supposed to be an important work considering the escalation of healthcare-associated infections globally, however, the major drawback noticed is the lack of the systematic review of the identified studies. There is an obvious lack of information on healthcare-associated infections in the manuscript vis-a viz the different types and how the integration of these was done to culminate into meta-analysis.

Author response: Thank you. Now all the comments are revised as per the comment.

Specific comments by reviewer #2 taken from the attachment

Comment: Editorial comments throughout the document

Author response: Yes Sure! From repeated proof-reading of the manuscript, we found several grammatical errors, interlinings, police titles, punctuation errors, wordings, and spelling errors. Therefore, finding our colleague who has a Master of Arts in English, we have done our best to thoroughly copyedit the manuscript for English language usage. These editorial changes are found throughout the revised version manuscript.

Background section

Comment: In the background section “Please endeavor to select references that are archived such as that from the World Health Organisation, US Centres for Disease Control and Prevention, etc. Not internet materials for HCAI”

Author response: Yes indeed! We have revised as per your comment, and we use the definitions used by CDC. The changes made could be seen in the introduction section.

Comment: “There is a published document by WHO. Please cite that in the stead. See number 10 citation”

Author response: Thank you indeed. The citation is revised as per the comment given. The revision made is available in the introduction section as track change.

Results section

Comment: The number of pocket studies included were 19 but written as 18

Author response: Thank you for raising this issue. Yes, the total number of studies included for meta-analysis was 19, but the total number of studies used to estimate the prevalence was 18. Nevertheless, the 19th study was case-control which we use it for determinant estimation. So, dear reviewer #2, the discrepancy in the number of studies was due to this scenario. Now, the number of articles included is revised accordingly. You can find the revisions made in the abstract section and the results section as track changes.

Comment. In the results section at subheading risk factors of HCAI, you recommend the risk factors to be presented in the table. “This is better presented as a table where all denominators are known before statistical analysis done on it”

Author response: Yes sure! The revision is made according to your comment and included in Table 3 in the revised version of the manuscript. The changes can be depicted in the results section under sub-heading Determents of healthcare-associated infection.

8 Oct 2020

Healthcare -associated infection and its determinants in Ethiopia: A systematic review and meta-analysis

PONE-D-20-09587R1

Dear Mr Abebew Alemu,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Professor Kwasi Torpey, MD PhD MPH

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Comments addressed. There are still a few language corrections needed. Final copyediting will be helpful

Reviewers' comments:

15 Oct 2020

PONE-D-20-09587R1

Healthcare-associated infection and its determinants in Ethiopia: A systematic review and meta-analysis

Dear Dr. Alemu:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Kwasi Torpey

Academic Editor

PLOS ONE

Table 1

Characteristics and quality status of the studies included.

First author year	Study region	Study design	Sample size	Prevalence	Quality status
Gedebu M. et al./1987 [19]	Addis Ababa	Cross-sectional	2506	13.40	Low risk
Gedebu M. et al./1988 [20]	Addis Ababa	Cross-sectional	700	17.00	Low risk
Habte-Gaber E. et al./1988 [21]	Addis Ababa	Cohort	1006	16.40	Low risk
Berhe N. et al./2001 [22]	Addis Ababa and SNNPR	Cohort	247	5.90	Low risk
Endalfer N. et al./2008 [23]	Addis Ababa	Cross-sectional	854	9.00	Low risk
Tesfahun Z. et al./2009 [24]	Tigray region	Cross-sectional	246	27.60	Low risk
Endalfer N. et al./2011 [25]	Addis Ababa	Cross-sectional	215	35.80	Low risk
Melaku S. et al./2012 [26]	Amhara region	Cross-sectional	1383	17.80	Low risk
Melaku S. eta al/2012 [27]	Amhara region	Cross-sectional	1254	9.40	Low risk
Mulu W. et al./2013 [16]	Amhara region	Cross-sectional	294	10.90	Low risk
Sahile T. eta al/2016 [28]	Oromia region	Cross-sectional	500	35	Low risk
Yallew WW. et al./2016 [8]	Amhara region	Cross-sectional	908	14.90	Low risk
Tolera M. et al./2018 [18]	Oromia region	Cross-sectional	394	6.90	Low risk
Gashaw M. et al./2018 [17]	Oromia region	Cross-sectional	1015	11.60	Low risk
Ali S. et al./2018 [9]	Oromia region	Cohort	1069	19.40	Low risk
Alemayehu T. et al./2019 [10]	SNNPR	Cross-sectional	939	21.40	Low risk
Gebremeskel S. et al./2018 [31]	Addis Ababa	Cross-sectional	410	19.80	Low risk
Yallew WW. et al./2017 [29]	Amhara region	Case-control	545		Low risk
Zewdu et al./2017 [30]	Oromia region	Cohort	300	14.00	Low risk

Note: SNNPR: Southern Nations Nationalities and Peoples Region; Low risk: a study scored > 50% in the JBI quality assessment scale.

Table 2

The pooled prevalence of HCAI, 95% CI, and heterogeneity estimate with a p-value for the subgroup analysis, by region, study design, sample size, and diagnostic method.

Variables	Characteristics	Pooled prevalence (95% CI)	I2
Region	Addis Ababa	18.44% (14.02–22.86)	99%
	Oromia	17.37% (9.2–25.56)	99.5%
	Amhara	13.27% (9.00–17.52)	98.5%
	Tigray	27.6% (25.37–29.83)	-
	SNNPR	21.4% (20.3–22.5)	-
	Addis Ababa & SNNPR	5.9% (4.25–7.55)	-
Study design	Cross-sectional	17.83% (14.39–21.27)	99.3%
	Cohort	13.96% (8.78–19.14)	98.4%
Diagnostic method	Clinical and laboratory	18.2% (14.85–21.51)	99.2%
	Culture-confirmed	12.71% (6.4–19.02)	99%
Sample size	<1000	18.15% (13.28–23.03)	99.3%
	≥1000	14.66% (11.72–17.59)	98.6%

Note: SNNPR: Southern Nations Nationalities and Peoples Region; I2: reported for the pooled effect of two or more studies.

Table 3

Determinants of healthcare-associated infection in Ethiopia.

Determinants	Author/year	HCAI		Effect size (95% CI)	Pooled effect size (95% CI)	I2
		Yes	No
Surgical procedure	Endalfer N. eta al/2011 [25]	71	66	3.96 (2.82–5.09)	3.37 (1.85–4.89)	46.9%
	Yallew WW. et al./2017 [29]	64	164	2.35 (0.35–4.34)
Take prophylaxis	Melaku S. et al./2012 [27]	54	237	1.76 (1.21–2.3)	1.76 (1.21–2.3)	-
underlying non-communicable disease	Melaku S. et al./2012 [27]	16	71	4.3 (2.32–6.28)	2.81 (1.39–4.22)	54.5%
	Mulu W. et al./2013 [16]	10	36	2.72 (0.42–5.01)
	Ali S. et al./2018 [9]	44	135	2.01 (1.15–2.87)
Age ≥ 51 years	Mulu W. et al./2013 [16]	6	16	6.38 (-10.61–23.37)	6.38 (-10.61–23.37)	-
Duration of operation 90–150 minutes	Mulu W. et al./2013 [16]	3	6	11 (-18.41–40.41)	11 (-18.41–40.41)	-
Hospital stay >5 days	Mulu W. et al./2013 [16]	2	3	8.2 (5.2–11.2)	5.32 (0.01–10.65)	89.8%
	Alemayehu T. et al./2019 [10]	58	183	2.76 (1.13–4.37)
Age 1–14 years	Yallew WW. et al./2016 [8]	14	148	0.25 (-0.06–0.56)	0.25 (-0.06–0.56)	-
Admission to the surgery ward	Yallew WW. et al./2016 [8]	75	240	2.86 (1.33–4.38)	2.86 (1.33–4.38)	-
Patients admitted at Felege Hiwot Hospital	Yallew WW. et al./2016 [8]	74	261	1.99 (1.2–2.77)	1.99 (1.2–2.77)	-
Immuno-deficiency	Yallew WW. et al./2017 [29]	31	92	2.34 (0.57–4.1)	2.34 (0.57–4.1)	-
Central vascular catheter	Yallew WW. et al./2017 [29]	5	4	6.92 (-11.17–25.01)	6.92 (-11.17–25.01)	-
Patient received antimicrobial	Yallew WW. et al./2017 [29]	104	294	8.63 (-1.79–19.05)	8.63 (-1.79–19.05)	-
Medical waste container at room	Yallew WW. et al./2017 [29]	102	431	0.18 (-0.290–0.65)	0.18 (-0.290–0.65)	-
Previous hospitalization	Ali S. et al./2018 [29]	20	25	1.65 (0.91–2.39)	2.13 (0.71–3.55)	49.4%
	Gebremeskel S. et al./2018 [31]	27	43	3.22 (1.16–5.28)
Age 18–30 years	Ali S. et al./2018 [9]	28	179	0.54 (0.22–0.85)	0.54 (0.22–0.85)	-
Chest tube insertion	Ali S. et al./2018 [9]	3	1	4.14 (1.57–6.71)	4.14 (1.57–6.71)	-
Mechanical ventilation	Ali S. et al./2018 [9]	12	22	1.99 (0.65–3.32)	1.99 (0.65–3.32)	-
Malnutrition	Alemayehu T. et al./2019 [10]	39	135	2.1 (0.78–3.41)	2.1 (0.78–3.41)	-
Male sex	Gebremeskel S. et al./2018 [31]	38	95	2.1 (0.45–3.67)	2.1 (0.45–3.67)	-
Hospital stay < 5 days	Gebremeskel S. et al./2018 [31]	5	119	0.03 (-0.01–0.07)	0.03 (-0.01–0.07)	-

Note: I2: reported for the pooled effect of two or more studies.