Study design
This systematic review has been designed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) statement and has been registered on PROSPERO (CRD42020196610) [15].
Eligibility criteria
Population
Included are COVID-19 infection patients among adults (≥18 years of age) who are diagnosed with four types of psychiatric and neuropsychiatric syndromes (anxiety, depression, delirium, and post-traumatic stress disorder). We will use the author's definitions of psychiatric and neuropsychiatric syndromes, as well as diagnostic criteria for SARS-CoV-2 infection, with no age, gender or setting, location, or ethnicity restrictions.
Excluded are studies explored the indirect effects of SARS-CoV-2 on the mental health of family members, care providers, or isolated people who did not infect will be excluded. We also excluded studies including populations with other coronavirus diseases (severe acute respiratory syndrome (SARS) or Middle East respiratory syndrome (MERS), unless the trial authors provided subgroup data for people with COVID-19.
Type of outcomes
The primary outcomes are the prevalence of signs or symptoms (depression, anxiety, delirium, and PTSD). Secondary outcomes are the symptom severity (depression, anxiety, delirium, and PTSD), the incidence of mortality in COVID-19 patients with depression, anxiety, delirium or PTSD, and the measurement of health-related quality of life using a validated scale, such as the Short Form 36 Health Survey questionnaire. The outcome will be classified as examining the acute or post-illness psychiatric consequences of infection on the basis of whether the information is collected during the patient’s illness or the period after the illness.
If a study met our criteria reports on COVID-19 infection that were not specified a priori as outcomes of interest for this review, the results will be noted in a narrative synthesis, but not necessarily pooled for meta-analyses nor reported in the summary of findings table.
Studies design
We will include only peer-reviewed RCT, cohort/case-control/cross-sectional studies, case reports, case series, and qualitative studies. Conference abstracts, commentaries, or opinion pieces will be excluded because they lack adequate information for meta-analysis. Only studies published in English and Chinese will be included in this review.
Search strategy
A senior investigator (Y.G.) would examine the published and gray literature sources to extract the studies reporting the prevalence of depression, PTSD, anxiety, or delirium in COVID-19 patients. An experienced medical information specialist (J.H.T.) would further check and approve the search methodology. We will conduct a comprehensive search of the Cochrane Library, PubMed, Web of Science, Embase, and Chinese Biomedicine Literature to extract articles/abstracts published between the inception of this disease (December 2019) until the completion of this review will be included. There will be no restrictions on language or year of publication. An additional file, which would describe the complete search strategy for PubMed as well as other electronic databases will be provided. We will also thoroughly search the reference lists of the relevant reviews and research trials. We have presented the search strategy using PubMed as an example in Table 1. The search strategy will be adapted to fit other online databases as well. Studies published before 12 December 2019 will be excluded.
Update plan
We will perform identical search operations at regular pre-defined intervals to identify newly published data. There are no robust standards for the update frequency based on current research; however, due to the unprecedented number of publications on COVID-19, we will update the literature searches every month, and perform meta-analysis if any new eligible studies or data are obtained. We will submit an updated systematic review if we observe any changes in the outcomes and heterogeneity after the addition of new studies or provide data on additional outcomes [11, 12]. We chose this updating frequency to allow quick updates and to highlight the most recent information to the researchers, clinicians, nurses, and policymakers [11, 14, 16].
Study selection
Original literature search records will be imported into Endnote X9 software tool (Thomson Reuters, New York, NY, USA) management software. Two authors (JYS and YG) will independently retrieve full-text of potential studies after deduplication to assess their eligibility according to the abovementioned inclusion criteria. Any disagreement will be resolved by the third reviewer (JHT).
Data extraction
Two independent reviewers (JYS and MMN) will be involved in data extraction; we will extract country of patients, population type (e.g., old people and children), age, study design (such as RCT/cohort/case-control), diagnostic criteria for the viral infection (such as WHO criteria), stage and severity of the disease, length of follow-up, sample size (such as number of cohort, number of cases), and gender.
Risk of bias (quality) assessment.
Two independent reviewers (JYS and YG) will use the following tools to examine the risk of bias in the included studies: Cochrane Collaboration RoB 2.0 tool for RCT [17], the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I Scale) for non-randomized controlled trials [18], the Newcastle-Ottawa Scale (NOS) for cohort and case-control studies [19]. The 11-item checklist recommended by the Agency for Healthcare Research and Quality (AHRQ) will be used to assess the the quality of the cross-sectional studies that were included [20]. We will classify the methodological quality of each individual study as having a low, high, or unclear risk of bias as describing in Table 2. Any disagreement regarding inclusion of some studies will be resolved by discussion and consensus between the two reviewers. If this failed, it shall be resolved by the third reviewer (JHT).
Processing missing data
We will contact the corresponding or other primary authors to obtain missing data or insufficiently reported data after selecting the studies. Randomized controlled trials will be treated as cohort studies; all data from the control and experimental group will be extracted if they met our criteria. In addition, we will estimate missing data if they can be extracted from tables or figures. Trials with missing data that cannot be obtained will be excluded for reasons. Any significant deviations between the protocol and the final review will be reported clearly.
Differences between the protocol and the final review
Any significant deviations between the protocol and final review will be reported clearly.
Data analysis
The Stata (v13.0; StataCorp) and Revman 5 were used for statistical analysis. The statistical heterogeneity will be examined using the Cochran's Q and the I² statistic. An I² > 50%, and a p-value < 0.05 will correspond to significant heterogeneity, and a random-effects model will be used for the subgroup analyses and pooled estimates. On the contrary, an I² < 50% and a p-value > 0.05 will correspond to insignificant heterogeneity, and the fixed-effect model shall be used for the subsequent meta-analysis. The effect size measures were mean difference with 95% CI (for the severity of the symptoms and degree of diagnoses) and prevalence with 95% CI (number of psychiatric diagnoses (depression, anxiety, delirium, and PTSD); severity of depression, anxiety, delirium, and PTSD). The heterogeneity/publication bias will be examined using the Egger's test or the symmetry of the funnel plot. In the Egger's test, bias will be significant when p-value < 0.05.
Subgroup analysis
The following subgroup analyses will be planned for main outcomes if data are sufficient: Age (< 60 vs. ≥ 60 years), symptom severity (mild vs. severe vs. ICU patients), high and middle-high vs. middle-low and low-income countries, databases (data from Chinese databases vs. data from English databases), study design (RCT, cross-sectional, non-randomized control trials, cohort-study case-control trials, and qualitative studies), and follow-up time (1, 3, 6 months of acute and post-illness for COVID-19 patients).
Sensitivity analyses
We will perform sensitivity analyses by repeating meta-analysis with studies with an unclear or low risk for bias. Additional issues suitable for sensitivity analysis will be identified during the review process. We will report the sensitivity analysis by generating summary tables
Quality of the evidence assessment
The Grades of Recommendations, Assessment, Development, and Evaluation (GRADE) Working Group method will be used to examine the quality of the evidence for each outcome. We will assess each outcome based on each of the following five aspects: imprecision, inconsistency, limitations, indirectness, and publication bias. They will be rated as very-low, low, moderate, or high level [21].