Assessing the link between implementation fidelity and health outcomes for a trial of intensive case management by community health workers: a mixed methods study protocol

Chronic disease is a major contributor to the health differentials in Indigenous populations. This is evident in Australia where the health status of Aboriginal and Torres Strait Islander people sits at the bottom of the league table of first world nations, with chronic disease the biggest contributor to burden of illness [1]. In rural and remote areas of Australia the hospitalisation rates and prevalence of risk factors of chronic disease for Aboriginal and Torres Strait Islander populations is higher than non-indigenous people [2]. It is widely recognized that better systems of care are required to address chronic disease and ensure that Aboriginal and Torres Strait Islander people have access to the care they need [3, 4].

Implementation of effective health service interventions to improve chronic disease outcomes in Aboriginal and Torres Strait Islander populations is complex. Factors that contribute to chronic illness relate to the characteristics of Indigenous populations, infrastructure in Indigenous communities and the capacity of health services to respond to Indigenous health issues. Characteristics of Indigenous populations that contribute to poor health outcomes include the social determinants of health, environmental factors and intergenerational grief and trauma caused by colonisation and racism [5]. Compared with urban areas, there is less access to comprehensive primary health care services in rural and remote locations and greater challenges with accessing specialist care [6, 7]. Workforce supply, skills and capacity to respond to health needs in a culturally competent manner, poor orientation, inappropriate service delivery models, lack of knowledge of chronic disease care guidelines, poor access to information technology and lack of resources are all barriers to effective chronic disease interventions in Indigenous Aboriginal and Torres Strait Islander populations [8–10].

Indigenous health research has produced evidence about effective models of care and chronic disease strategies that address barriers to health improvement [11–13], however the transfer of research findings from trials into routine clinical practice has proven challenging [14, 15]. Campbell et al. suggested that a solution to this problem is to strengthen research design and execution by improved theory development, modelling using qualitative and quantitative methods and evaluating long term implementation [16]. It has been suggested that poor implementation may be due to the characteristics of the intervention itself, the target setting, the research design and the interaction between these elements [15]. Another reason for poor transferability of research into practice is that research reports do not describe the intervention well enough to enable application or authentic replication in the real world setting [17]. A review of 80 studies published in the journal Evidence Based Medicine found that only 39% of non-drug trials adequately described interventions to enable replication [18]. More robust theory development and better description of interventions contribute to improved understanding of intervention components, but they do not help measure if the intervention is being implemented as intended. Therefore additional strategies are needed to monitor the implementation of interventions in a primary health care setting.

Knowing if an intervention was implemented as planned is fundamental to knowing what has contributed to the success of an intervention [19]. Complex interventions, such as those related to chronic disease, have many components that are often poorly implemented and hence rarely achieve implementation fidelity. Implementation fidelity is “the degree to which programs are implemented as intended by the program developer” [20, 21]. Breitenstein et al. state that fidelity is critical to the systematic implementation of evidence based practice [21]. Therefore health service evaluation models need to include measures that assess implementation fidelity.

The RE-AIM framework provides a model for evaluating public health initiatives by scoring the impact of an initiative in a real world setting. The dimensions measured to obtain the score include: reach, efficacy, adoption, implementation and maintenance. These dimensions are evaluated at an individual, organisational and/or community level [14]. While this model is useful for evaluating the effectiveness of chronic disease interventions in real world settings [22, 23], implementation fidelity is not being measured.

In comparison, Carroll et al. developed a conceptual model to measure implementation fidelity by allocating a score for implementation based on adherence to the intervention. The theory underpinning the model is that “implementation fidelity is the moderator between interventions and their intended outcomes” [20]. A score for implementation fidelity is arrived at by measuring adherence to an intervention which can then be correlated to health outcomes. The conceptual model provides a mechanism for assessing the effectiveness of individual components to help identify what is essential to implementing an intervention. Dimensions evaluated include: coverage, frequency and duration.

Carroll et al’s model was adapted and applied by Keith and colleagues to measure implementation fidelity of a nurse lead case management intervention for patients with cardiovascular disease [24]. Using this approach researchers and service managers were able to understand the level of implementation fidelity achieved and the impact of different components of the model on health outcomes. A key addition made by Keith et al. was the inclusion of information about context to help explain the score awarded. Implementation of any new intervention requires active change management and context plays an important role in how an intervention is received and implemented. A predictor of successful change is that the intervention proposed is consistent with the values of the organisation and there is capacity in the organisation to manage change [25]. If the innovation being proposed does not fit with the values of the organisation or is not communicated to those expected to implement it, no matter how effective the intervention may be in research conditions, it is unlikely to be implemented as planned. Measuring implementation fidelity and analysing results using a theoretical lens of change management would assist managers to improve service planning for complex chronic disease interventions and better manage the risks with implementation of chronic disease interventions in Aboriginal and Torres Strait Islander communities.