BMJ Publishing Group, BMJ Open, 10(8), p. e019993, 2018
DOI: 10.1136/bmjopen-2017-019993
Full text: Download
ObjectivesThis study compares rapid and traditional analyses of a UK health service evaluation dataset to explore differences in researcher time and consistency of outputs.DesignMixed methods study, quantitatively and qualitatively comparing qualitative methods.SettingData from a home birth service evaluation study in a hospital in the English National Health Service, which took place between October and December 2014. Two research teams independently analysed focus group and interview transcript data: one team used a thematic analysis approach using the framework method, and the second used rapid analysis.ParticipantsHome birth midwives (6), midwifery support workers (4), commissioners (4), managers (6), and community midwives (12) and a patient representative (1) participated in the original study.Primary outcome measuresTime taken to complete analysis in person hours; analysis findings and recommendations matched, partially matched or not matched across the two teams.ResultsRapid analysis data management took less time than thematic analysis (43 hours vs 116.5 hours). Rapid analysis took 100 hours, and thematic analysis took 126.5 hours in total, with interpretation and write up taking much longer in the rapid analysis (52 hours vs 8 hours). Rapid analysis findings overlapped with 79% of thematic analysis findings, and thematic analysis overlapped with 63% of the rapid analysis findings. Rapid analysis recommendations overlapped with 55% of those from the thematic analysis, and thematic analysis overlapped with 59% of the rapid analysis recommendations.ConclusionsRapid analysis delivered a modest time saving. Excessive time to interpret data in rapid analysis in this study may be due to differences between research teams. There was overlap in outputs between approaches, more in findings than recommendations. Rapid analysis may have the potential to deliver valid, timely findings while taking less time. We recommend further comparisons using additional data sets with more similar research teams.