Article Text
Abstract
Background Evidence-based healthcare innovations require complementary evidence-based implementation strategies to support their translation into practice. Efforts to test, refine and replicate implementation strategies are frustrated by insufficient description. Our aim was to examine the extent to which implementation strategies could be specified using the Behaviour Change Technique (BCT) taxonomy, a behavioural science tool for describing the active ingredients of interventions.
Methods The data source was a compilation of 73 implementation strategies, developed through evidence synthesis and expert consensus. The definition of each strategy (n=73) was deductively coded using the BCT Taxonomy, containing 93 discrete techniques. A typology was developed iteratively to categorise the extent of overlap between strategies and BCTs. The number of BCTs per strategy and extent of overlap was estimated. In the next stage, 3 experts will independently rate 1) their level of agreement with the categorisation and 2) level of agreement with the BCT(S) identified within each strategy.
Results During preliminary analysis, 87 BCTs were coded across 73 strategies (average 1.2 per strategy). Five types of overlap were identified. For 8% of strategies (n=6), there was direct overlap between the strategy description and BCT (e.g. strategy: remind clinicians/BCT: prompts and cues). For 36% of strategies (n=26), there was at least 1 BCT clearly subsumed under the strategy description which could be used to guide initial operationalisation (e.g. strategy: provide clinical supervision/BCT: restructure social environment). For 26% of strategies (n=19), a BCT(s) was probably subsumed under the strategy given its definition and/or title but other BCTs were possible depending on how the strategy is operationalised (e.g. strategy: visit other implementation sites/BCT: social comparison). For 11% (n=8), there were no BCTs clearly indicated in the strategy definition or title (e.g. strategy: make training dynamic). Finally, 19% of strategies (n=14) did not focus on behaviour change to support implementation (e.g. strategy: access new funding).
Conclusion Many implementation strategies require further specification in order to apply them in a setting, relying on assumptions and inference on the part of the intervention developer, be it researcher or practitioner. This creates an opportunity for inconsistent application and limits the potential for replication and synthesis of evidence of effectiveness.
This study is the first step towards moving from general descriptions of implementation strategies to full descriptions of their active ingredients. This is essential to understand how strategies at an organisational and professional level can lead to observable changes in individual behaviour.