A theory-based approach to evidence synthesis: the example of audit and
Background: Behaviour change theories offer frameworks for categorising and
evaluating the active ingredients of interventions and identifying the likely
mechanisms through which effects are achieved. Yet systematic reviews of
behaviour change interventions rarely classify intervention components using
theory, and perhaps consequently, seldom produce clear patterns of results.
Aims: This project aims to illustrate the application of theory to evidence
synthesis by using constructs from Control Theory to recode and reanalyse data
from a Cochrane review of audit and feedback interventions to improve
compliance with clinical practice.
Methods: Data received from Cochrane reviewers was supplemented by data
from an updated literature search. 88 comparisons, reported in 68 studies, were
eligible for analysis. Interventions were recoded according to whether, in addition
to feedback, a specific behavioural target was set, and/or an action plan formed
for improving performance. Data were analysed using univariate and multivariate
Results: 88 comparisons, reported in 67 studies, were retained for analysis.
Only 8 interventions included behavioural targets, and only 3 interventions
featured all three techniques. 24 interventions included targets and/or action
plans. In line with predictions from Control Theory, feedback was more effective
where supplemented by targets and/or action plans.
Gardner, B., Whittington, C., McAteer, J., Eccles, M., & Michie, S. (in prep) A
theory-based approach to evidence synthesis: the example of audit and
Michie, S., McAteer, J., & Gardner, B. (2009) Using theory to specify the details
of complex behavioural interventions. Paper presented to UKSBM conference,
Exeter, UK, 7th January.