Current rehabilitation definitions do not allow correct classification of Cochrane systematic reviews: an overview of Cochrane reviews

BACKGROUND: During the first three years of its work, Cochrane Rehabilitation was faced with the challenge of defining the inclusion and exclusion criteria of what is rehabilitation on four different occasions: when we worked on classifying all Cochrane systematic reviews (CSRs) for relevance to rehabilitation, when we checked for newly published CSRs, when we started the process to set up the reporting guidelines for the Randomized Controlled Trials Rehabilitation Checklist (RCTRACK) project, and during our collaboration with the World Health Organization for the Package of Rehabilitation Interventions. The aim of this paper was to check how the word "rehabilitation" gets used by researchers in the health field. METHODS: This overview of reviews included all CSRs that used the term "rehabilitation" in the title. They were compared with the authors' judgement (AJ) and with the contents of two main sources: CSRs identified by Cochrane Rehabilitation as relevant to rehabilitation (CRDB), and PubMed MeSH tern "rehabilitation." We also performed a content analysis classifying all CSRs by field and type of intervention and checked the internal coherence of the two databases in order to verify whether all CSRs on interventions in a specific rehabilitation field were included in the databases or not. RESULTS: Out of 14,816 PubMed entries, we analyzed 89 CSRs. We found four reviews that were judged by all classifications as not rehabilitation: they were related to mouth, nutritional, penile and schizophrenia rehabilitation. While CRDB and AJ included 94% and 91% of CSRs respectively, PubMed included only 50%. One CSR about cardiac rehabilitation was excluded only by CRDB and four by AJ. In the 50% CSRs excluded by PubMed, we found that all CSRs on cancer and vestibular rehabilitation, and those on cognitive and neuropsychological interventions, were always omitted, even if all other CSRs on neurological rehabilitation were included. CONCLUSIONS: Our results clearly highlight the need for a comprehensive rehabilitation definition that is able to point out what should be included and excluded from rehabilitation interventions. This will consequently inform all of Cochrane Rehabilitation's work and will serve the wider community of research and rehabilitation.