Our search strategy, selection of studies, assessment of bias risk, and reporting of results for the review will be conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses statement [17].
2.1. Review Inclusion Criteria
2.1.1. Participants
This review will include all patients with schizophrenia who received/were subject to a community-based intervention program. Further, the results will be presented separately, based on the duration of the patients’ illness (divided into the first episode or chronic status).
2.1.2. Intervention
This review will consider studies that evaluate any type of intervention programs that originated from the Community-based Intervention Program for patients with schizophrenia. Those interventions may include but are not limited to, case management, cognitive behavioral therapy, occupational rehabilitation, and physical intervention programs. In addition, the review will include the patients’ symptomatic severity and quality of life as outcomes of the community-based intervention.
2.1.3. Comparators
The comparison groups will include patients with schizophrenia receiving standardized treatment that includes only medication. Additionally, cohort study designs will be included in this review, so studies without a comparison group will be included.
2.2. Outcomes
Schizophrenia causes frequent relapses and poor quality of life when treatment is ineffective or not sustained. In view of this, exploratory analysis will be conducted to identify relapse, recovery, and/or remission rates of psychotic symptoms. The patients’ conditions will be verified by the number and duration of hospitalizations after the community-based intervention, by their scores on the Positive and Negative Syndrome Scale (PANSS), Brief Psychiatric Rating Scale (BPRS), and Global Assessment of Functioning (GAF). Quality of life will be defined by the patients’ and their respective caregivers’ Quality of Life Scale scores. Thus, any studies that report any of the above outcomes will be included.
2.3. Study Design
This review will consider randomized controlled trials (RCT), non-randomized controlled trials (non-RCT), and cohort studies, but will exclude single case studies and reviews. It will report the specific characteristics of all included studies, and the inclusion criteria are: studies must be written in either English or Korean, must have been published in the last 10 years (from 2009 to 2019), and must be a full paper (i.e., conference abstracts will not be considered).
2.4. Electronic Bibliographic Databases
Electronic searches will be conducted on the following databases: MEDLINE (PubMed), Cochrane Central Register of Controlled Trials (CENTRAL), PsycINFO, EMBASE and the Cumulative Index to Nursing and Allied Health Literature, Research Information Sharing Service (Korean database). Databases will be searched for studies from September 2009 to September 2019.
2.5. Search Strategy
The search strategy aims to find published studies that are in accordance with the Population Intervention Comparison and Outcome Process. An initial search of PubMed will utilize text words related to the systematic review research question: “schizophrenia” and “community based” or “community mental health services.” Then, we will identify relevant keywords by an analysis of the text words contained in the title and abstract, and of the index terms used to describe the relevant articles to refine our search. Specific search strategies are provided as attachments.
2.6. Study Screening and Selection
Search results will be downloaded using Endnote software, X9 version, and duplicate studies will be eliminated. In the first review, we will review the title and abstract of the selected studies to identify populations, intervention and outcome variables, and study designs to eliminate non-relevant literature; in the second review, a full-text review will identify the final literature of the selected studies. Each selected study will be independently reviewed by two researchers and will be cross-reviewed by both researchers. During the process, if opinions do not agree among researchers, the text will be reviewed together until the researchers reach an agreement.
2.7. Data Extraction
Data extraction will include specificities about populations, types of interventions, study designs, and outcome variables. Researchers will select five articles to create a pilot-format data extraction tool, and this tool will use the Excel software program (Microsoft).
In addition to the outcome data, descriptive details such as study designs (e.g., RCT or non-RCT), participants' characteristics (e.g., age, gender, disease status), methods used in the analysis, and methods of intervention (handling) will also be recorded and reviewed. The amount, duration, frequency, and intensity of each reported intervention will also be included in the record. The demographic characteristics and types of interventions will also be specified to enhance the study analysis and synthesis.
If there are any missing or unclear data, we will contact the author of the original research to clarify.
2.8. Assessing Risk of Bias
For RCT studies, we will use the Cochrane RoB 2.0 tool [18]. RoB is a tool that combines both the checklist method and area evaluation method, and it is an important tool because its area evaluation randomizes sequence generation, which blinds parts of the study and study personnel, blinds the outcome assessments, and does not include incomplete outcome data, which helps us avoid selective reporting and other possible types of biased selection. To avoid the risk of biases in each question, they will be judged as “high,” “low,” and “uncertified” bias, in accordance with the specific presented guidelines.
Non-RCTs, including cohort studies, will use a Risk of Bias in Non-randomized Studies-of Interventions (ROBINS-I). ROBINS-I evaluation areas are related to the selection of target populations, confounding variables, measurement of exposure, blinding of outcome assessments, incomplete result data, and selective reporting [19]. Each area can be divided according to study type: cohort, comparative clinical trials, and before-after studies. This will be performed to determine the risk of warping in that specific area. To avoid the risk of biases in each question, they will be judged as “high,” “low,” and “uncertified” bias. Two reviewers will independently assess the methodological quality of selected studies, and any disagreement will be resolved by discussion.