To achieve this objective, we will conduct an overview of systematic reviews according to the principles of the Cochrane Handbook’s chapter on methods for overviews of reviews (25) and other recent methodological papers (27–29). Overview of reviews (OvR) is designed to compile, synthesize, and integrate evidence from multiple systematic reviews into one accessible and usable document (25, 27). Despite their increasing popularity in healthcare research over the past years, there are currently no systematically developed reporting guidelines for OvR (24, 30). However, this protocol has been prepared in consultation with the PRISMA-P statement (31) (checklist provided in the supplemental documents to this protocol, in the additional file). This review has been registered in the PROSPERO database (CRD42020176676). Any amendments to the protocol will be described in the final review article.
Literature searches
A professional librarian and members of the research team will help to develop the search strategy based on the objective of the review. We used the PICO model to develop our search strategy: 1) MTBI/concussion (patient); 2) Chronic/persistent symptoms (intervention); 3) Prognosis (outcome); and 4) Systematic review (types of study). The “comparison” concept is not relevant for our review, and we chose to add a “types of study” concept considering that we will focus our search on systematic reviews only. We will search in five relevant databases (Cochrane (Wiley), Medline (Ovid), CINAHL (EBSCO), Embase (Elsevier), PsycINFO (Ovid)) for systematic reviews published in peer-reviewed journals without date restrictions. An example of a search strategy in Medline (Ovid) is displayed in Table 1.
Table 1
Example of search strategy in Medline (Ovid) database
1) Concept : MTBI concussion |
#1 | [ti.] | (concuss* OR commotio* cerebr* OR cerebral* commotio* OR mtbi) or ((mild OR minor OR minimal) adj3 (traumatic brain OR tbi)) |
#2 | [ab.] | (concuss* OR commotio* cerebr* OR cerebral* commotio* OR mtbi) OR ((mild OR minor OR minimal) adj3 (traumatic brain OR tbi)) |
#3 | [Mesh] | Brain injuries, traumatic/ |
#4 | [Mesh] | Brain concussion/ |
#5 | #1 OR #2 OR #3 OR #4 |
2) Concept : Chronic/persistent symptoms |
#6 | [ti.] | (prolong* OR persist* OR unresolved OR delay* OR chronic* OR post-concuss* OR postconcuss*) |
#7 | [ab.] | (prolong* OR persist* OR unresolved OR delay* OR chronic* OR post-concuss* OR postconcuss*) |
#8 | [Mesh] | Post-Concussion Syndrome/ |
#9 | #6 OR #7 OR #8 |
3) Concept: Prognosis |
#10 | [ti.] | (prognos* OR predict* OR course* OR outcome*) |
#11 | [ab.] | (prognos* OR predict* OR course* OR outcome*) |
#12 | [Mesh] | Prognosis/ |
#13 | #10 OR #11 OR #12 |
#14 | #5 AND #9 AND #13 |
4) Concept: Systematic reviews |
#15 | [pt.] | Systematic review |
#16 | #14 AND #15 |
Eligibility criteria
We will only include studies that have employed a systematic review design, peer-reviewed systematic reviews, or meta-analyses. We will consider publications to be systematic reviews if they were clearly described in the report as being based on a systematic search, by precisely presenting the search strategy and the selection process of the articles (such as a PRISMA flow diagram). French or English publications will be included. We will consider reviews targeting only adult populations with mTBI or concussion. In order to have a common definition of the condition, we will refer to the WHO Collaborating Centre Task Force on Mild Traumatic Brain Injury definition (32). We will include manuscripts that reviewed the course of at least one persistent symptom in mTBI patients. Narrative, non-systematic reviews, editorials/ commentaries, or grey literature, such as thesis work, will not be eligible for this OvR. We will also exclude reviews about moderate, severe, or non-traumatic brain injuries, and reviews targeting a child population.
Data management
All references will be imported from the databases in a reference management software (Endnote). Duplicates will be removed using the software command “Find duplicates” based on the title of the references. Then, once duplicates have been removed, the screening and selection process will be performed by using Covidence online software (33).
Screening and selection process
Two reviewers will independently screen all references identified from the literature search. All references judged potentially eligible after screening for title and abstract will be reviewed in a full text. Any disagreement will be discussed by the two reviewers, and if consensus is not reached, we will consult a third member of the team.
Data extraction
Relevant information from all selected reviews will be grouped in an extraction grid. We will create the grid and extract data from the reviews, such as study details (authors, year of publication, dates of search, design, number of studies included, types of analysis, purpose of the reviews, etc.), population studied, prognostic factors, outcomes measured (mainly post-concussion symptoms), post-injury timeframe, and other relevant results found in the reviews. Extraction will be limited to information presented in the included systematic reviews and not in the primary research studies (26). Two independent reviewers will extract data and then will compare their respective grid and reach a consensus.
Risk of bias assessment
Risk of bias of the included reviews will be rated using the Risk of Bias in Systematic Reviews (ROBIS) tool (34). ROBIS is useful and reliable for systematic reviewers to identify areas where bias may be introduced into systematic review methods: study eligibility criteria; identification and selection of studies; data collection and study appraisal; and synthesis and findings (34). We will summarize the number of systematic reviews that had a low, high, or unclear concern for each domain and the number of reviews at high or low risk of bias. Risk of bias ratings will be used in the data synthesis process to inform the conclusions of this review. No study will be excluded based on its risk of bias evaluation.
Data synthesis
Based on the data extracted, we will organize information from the reviews about prognostic factors into a comprehensive map. This synthesis will be inspired by the proposed model of Hou et al. (21) which organized the predictors in larger categories of prognostic factors, i.e. predisposing, precipitating, and perpetuating factors. We will also classify the factors regarding International Classification of Functioning, Disability and Health (ICF) (35). We will present data in tabular form and graphically if possible, in order to visually demonstrate the diversity of data in terms of quality of evidence and quality of reviews. We will produce a synthesized list of prognostic factors ranked with consideration of the quality of the evidence, risk of bias, and magnitude of long-term gravity of symptoms. Narrative descriptions from each review will complement the data and provide a comprehensive understanding of the synthesis.