Systematic Review on Concussion Recovery Time Lacks Comprehensiveness.

Abstract

I read with interest the article by Williams et al. [1] titled ‘‘Concussion recovery time among high school and collegiate athletes: a systematic review and meta-analysis’’. Though the authors followed the reporting standards set forth in the PRISMA guidelines [2], they (like many other authors of systematic reviews) fell short in conducting a comprehensive review of the literature and reporting a reproducible search strategy. Expertly devised search methods for identifying relevant published and unpublished studies are the bedrock upon which the validity of the entire review is dependent; as in order for a systematic review to perform its function as a reliable resource for clinical decision making, it must include all relevant clinical studies. The shortcomings within this systematic review are commonplace. Though an electronic search strategy for one database was presented, the search as reported is not reproducible. The authors state that their search was run in PubMed from database inception to December 2013, yet a large discrepancy exists between the 141 citations the authors state were found and the 1913 citations that result when the following strategy is run through December 31, 2013: (athletes or sports) and (concussion or mild traumatic brain injury) and (recovery or symptoms or cognition). Further, the search fails to qualify as a comprehensive literature search; these searches are innately sensitive and complex. The strategy used in this study does not use controlled vocabulary or proper syntax, nor does it provide an inclusive concatenation of those terms published on sports, athletes, concussions, recovery, symptoms, or cognition. Exhaustive searches that form the foundation of systematic reviews reflect an intimate understanding of the variability of keywords and database-specific indexing. They also take advantage of database-specific limiters/filters. For example, many studies written about concussions are sport specific. A search that only includes general terms for athletes or sports will miss some studies conducted on single sport athletes, such as football or soccer players. A comprehensive PubMed search for the concept of athletes or sports follows: (‘‘Sports’’[Mesh] OR ‘‘Athletes’’[Mesh] OR sport OR sports OR athlete* OR rugby OR football OR soccer OR lacrosse OR wrestling OR swimming OR hockey OR skiing OR skiers). Additional sports could be added, as appropriate. By improving syntax use and expanding upon the comprehensiveness of terms, the authors would have found several hundred additional articles in PubMed alone. Including other databases such as PsycINFO, Web of Science, Scopus, the Cochrane Library, and clinical trial registers would yield dozens more. Databases are designed so that all searches harvest results; however, only trained and experienced searchers have the capability to conduct the exhaustive searches required for systematic reviews. This is recognized in the Institute of Medicine’s ‘‘Recommended Standards for Finding and Assessing Individual Studies’’, which stipulate that authors of systematic reviews ‘‘Work with a librarian or other information specialist trained in performing systematic reviews (SRs) to plan the search strategy’’ and ‘‘Use an independent librarian or other information specialist to peer review the search strategy’’ [3]. The expertise of a medical librarian will ensure the comprehensive collection and examination of relevant studies. Additionally, owing to their intimate knowledge of literature in the & Jill Livingston jill.livingston@uconn.edu