Analysis of difficult decisions in categorization rehabilitation reviews in the Cochrane library

In 2017, Cochrane Rehabilitation created an online relational database to crowd-source the “tagging” of Cochrane publications, both protocols and completed reviews, for relevance to rehabilitation. Rehabilitation professionals were invited to contribute to this tagging work through a public call for interest. Twenty-five people from 13 countries responded to this call. We set up the tagging system to allow one rehabilitation physician and one allied health professional to independently categorize all Cochrane reviews for: a) relevance to rehabilitation, b) relevance to specific professional groups, and c) broad areas of clinical practice by patient pathology. We resolved all disagreements between contributors by discussing these reviews during Review Committee meetings, which involved two physiotherapists and one physician, to ensure consistency of categorization decisions. This dataset contains a list of all Cochrane reviews and protocol, published between 1996 and August 31, 2019 (excluding withdrawn publications, and including only the most recent version of each included protocol or completed review) where there was some disagreement about whether they were about rehabilitation or not, and so had to be discuss by Cochrane Rehabilitation's Review Committee. These reviews and protocols are separated into two groups (those that were considered to be about rehabilitation interventions and those that were not). This dataset also includes data on the rationale for these inclusion/exclusion decisions, plus information on where there were identified conflicts or errors in the initial categorization of reviews/protocols as being about or not about rehabilitation.