Objectives To study if Randomized Controlled Trials (RCTs) on complex interventions published in top journals include all the practical details needed to replicate the intervention in everyday clinical practice (clinical replicability). We chose rehabilitation as a case-study because the World Health Organization calls for its development within health services, and due to its intrinsic complexities.
Method Online survey of a pre-defined sample of clinical expert teams from different world regions with diverse rehabilitation competences. Forty-seven clinicians from 7 Physical and Rehabilitation Medicine (PRM) teams (Belgium, Italy, Malaysia, Pakistan, Poland, Puerto Rico, USA), including 20 physicians, 12 physiotherapists, 6 occupational therapists, 6 psychologists and 3 others. The team leaders were active researchers. All RCTs published between January and July 2017 in the top PRM journals (76 RCTs) were reviewed by each team leader. 14 questions developed using CONSORT and TIDIeR checklists through consensus and piloting.
Results The response rate was 99%. Inter-rater agreement was moderate/good. All participants considered unanimously 12 (16%) RCTs clinically replicable and none not replicable. Of the other, 56 (74%) RCTs have been considered replicable and 45 (59%) not replicable by at least one complete team. At least one ‘absent’ information was found by all participants in 60 RCTs (79%), and by a minimum of 85% in the remaining 16 (21%). Information considered to be less well described (8-19% ‘perfect’ information) included two providers (skills, experience) and two delivery (cautions, relationships) items. The best described (50-79% ‘perfect’) were the classic methodological items included in CONSORT (descending order: participants, materials, procedures, setting and intervention).
Conclusions This rehabilitation case-study shows problems of clinical replicability of RCTs on complex interventions and suggests the need to better define some clinical items not described by classical methodological checklists like CONSORT.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.