How are qualitative methods being used alongside complex health service randomized controlled trials? A systematic review

Article type
Authors
Lewin S, Glenton C, Oxman A
Abstract
Background: Qualitative methods are important in evaluating complex interventions. However, little research has examined how qualitative methods are currently used in randomized controlled trials (RCTs) and how they could be used to improve the usefulness and policy relevance of trial findings.

Objectives: To review systematically the use of qualitative approaches in RCTs of complex health service interventions.

Methods: A random sample was selected of 106 RCTs from the Cochrane Effective Practice and Organisation of Care (EPOC) Group's register and published in English in 2001, 2002 and 2003. All published and unpublished qualitative studies linked to these RCTs were identified through database searches and contact with RCT authors. Data were extracted from each study set by two reviewers using a standard form. These data included descriptions of the RCT and qualitative studies; the quality of the RCT and the qualitative studies; an assessment of the complexity of the intervention evaluated in each RCT; and the approaches (if any) used by the authors to combine the RCT and qualitative study findings. A narrative synthesis of the review findings was performed.

Results: Trialists appeared to see a wide range of approaches as constituting 'qualitative research', including using qualitative data collection and analysis; using specific qualitative data collection methods only; and using interview-based survey methods. Nineteen percent (n=20) of RCTs included 'formal' qualitative studies while 3% (n=3) included 'informal' qualitative work. The aim and timing of the qualitative studies and the methods used were highly variable, with many conducted pre-trial. Little explicit integration of qualitative and RCT findings or the use of mixed method models were identified.

Conclusions: The ways in which qualitative approaches are being used alongside RCTs are variable and these studies are of variable quality. More methodological work is needed on how best to use qualitative methods in this context.