3.1. Search Strategy
In this systematic review and meta-analysis study, All studies that reported in-hospital death in admitted patients to hospital due to COVID-19 has been used. The literature in Iran was acquired through Iranian scientific data base including SID, IranDoc, Health.barakatkns, MagIran and Civilica and international searching databases including PubMed, Scopus, Elsevier and Web of Science. Additional search was also done by Google Scholar search engine.
The last search took place on 11 April 2021. To search and include related studies as many as possible, we used the following terms: “Deceased”, “Mortality”, “Death”, “COVID 19”,“ Cross sectional”,“Cohort”,“ Observational”,“Iran”(or the names of its provinces), as keywords for titles and/or abstracts in MeSH word search database with such combination. (“Cohort”[Title/Abstract] OR “Observational”[Title/Abstract] OR “Cross Sectional”[Title/Abstract])) AND (“Mortality”[Title/Abstract] OR “Death”[Title/Abstract] OR “Deceased”[Title/Abstract]) AND (“COVID 19”[Title/Abstract] AND ( “Iran”[Title/Abstract] OR “Name of provinces,like Tehran, Isfahan, …”[Title/Abstract].
3.2. Selection of Studies and Data Extraction
Published studies were regarded as qualified for the analysis if they met the following criteria: (1) Observational, cross-sectional or cohort studies with the full text of the paper available in the Persian or English languages, (2) studies with a sample size of more than 30, and (3) studies reporting the prevalence of in-hospital death in admitted patients to hospital due to COVID-19, (4) studies with mean Age ranges from 16 to 100, (5) studies with approved diagnosis of COVID-19 by means positive RT-PCR of throat-swab specimens or the chest CT scans according to the WHO interim guidance including ground glass opacity in addition to ill-defined margins, smooth or irregular interlobular septal thickening, air bronchogram, crazy-paving pattern, and thickening the adjacent pleura [22–24] conversely, the following studies were excluded: (1) non-English or non-Persian full-text reports, (2) studies not providing enough data to estimate the death prevalence, (3) studies designed as letters to the editor, expert opinions, editorials, commentaries, case-reports, case-series, and reviews, and (4) Studies reporting overlapping data.
3.3. Data Extraction
All articles categorized as potentially relevant were reviewed separately by both of the authors (Seyed Hassan Saadat and Alireza Najimi-Varzaneh), In case of inconsistency in the results of the two authors, the other two authors ((Mostafa Ghanei and Mohammad Gholami Fesharaki ) review the articles and finally the results of the articles summarized the following data using Excel datasheets: First author’s name, duration of study, number of COVID-19 hospitalation patients, study sample size, name of the province, mean age and gender of responders. The analysis was conducted according to the preferred reporting items for systematic reviews and meta-analysis (PRISMA)[25]. In this study, The “Newcastle Ottawa Scale (NOS)” was used to evaluate the accuracy of the data in this analysis.
3.3. Statistical Analysis
The prevalence of in-hospitalized death among COVID-19 patient from each province of Iran was computed by metan command. Statistical tests of heterogeneity among the studies were carried out using the Q test (P < 0.10 indicates statistically significant heterogeneity) and I-squared statistics. We also used a funnel plot to investigate publication bias. In this study, "metafor" package in R software version 3.6. In this study, the publication bias was assessed graphically and statistically by funnel plot, and based on visual inspection of the funnel plot, Begg's Test, and Egger’s test.