Developing an implementation research logic model : Using a multiple case study design to establish a worked exemplar

Background
Implementation science frameworks explore, interpret, and evaluate different components of the implementation process. By using a program logic approach, implementation frameworks with different purposes can be combined to detail complex interactions. The Implementation Research Logic Model (IRLM) facilitates the development of causal pathways and mechanisms that enable implementation. Critical elements of the IRLM vary across different study designs, and its applicability to synthesizing findings across settings is also under-explored. The dual purpose of this study is to develop an IRLM from an implementation research study that used case study methodology and to demonstrate the utility of the IRLM to synthesize findings across case sites.

Method
The method used in the exemplar project and the alignment of the IRLM to case study methodology are described. Cases were purposely selected using replication logic and represent organizations that have embedded exercise in routine care for people with cancer or mental illness. Four data sources were selected: semi-structured interviews with purposely selected staff, organizational document review, observations, and a survey using the Program Sustainability Assessment Tool (PSAT). Framework analysis was used, and an IRLM was produced at each case site. Similar elements within the individual IRLM were identified, extracted, and re-produced to synthesize findings across sites and represent the generalized, cross-case findings.

Results
The IRLM was embedded within multiple stages of the study, including data collection, analysis, and reporting transparency. Between 33-44 determinants and 36-44 implementation strategies were identified at sites that informed individual IRLMs. An example of generalized findings describing “intervention adaptability” demonstrated similarities in determinant detail and mechanisms of implementation strategies across sites. However, different strategies were applied to address similar determinants. Dependent and bi-directional relationships operated along the causal pathway that influenced implementation outcomes.

Conclusions
Case study methods help address implementation research priorities, including developing causal pathways and mechanisms. Embedding the IRLM within the case study approach provided structure and added to the transparency and replicability of the study. Identifying the similar elements across sites helped synthesize findings and give a general explanation of the implementation process. Detailing the methods provides an example for replication that can build generalizable knowledge in implementation research.