One of the major debates in implementation research turns around fidelity and adaptation. Fidelity is the degree to which an intervention is implemented as intended by its developers. It is meant to ensure that the intervention maintains its intended effects. Adaptation is the process of implementers or users bringing changes to the original design of an intervention. Depending on the nature of the modifications brought, adaptation could either be potentially positive or could carry the risk of threatening the theoretical basis of the intervention, resulting in a negative effect on expected outcomes. Adaptive interventions are those for which adaptation is allowed or even encouraged. Classical fidelity dimensions and conceptual frameworks do not address the issue of how to adapt an intervention while still maintaining its effectiveness. We support the idea that fidelity and adaptation co-exist and that adaptations can impact either positively or negatively on the intervention’s effectiveness. For adaptive interventions, research should answer the question how an adequate fidelity-adaptation balance can be reached. One way to address this issue is by looking systematically at the aspects of an intervention that are being adapted. We conducted fidelity research on the implementation of an empowerment strategy for dengue prevention in Cuba. In view of the adaptive nature of the strategy, we anticipated that the classical fidelity dimensions would be of limited use for assessing adaptations. The typology we used in the assessment—implemented, not-implemented, modified, or added components of the strategy—also had limitations. It did not allow us to answer the question which of the modifications introduced in the strategy contributed to or distracted from outcomes. We confronted our empirical research with existing literature on fidelity, and as a result, considered that the framework for implementation fidelity proposed by Carroll et al. in 2007 could potentially meet our concerns. We propose modifications to the framework to assess both fidelity and adaptation. The modified Carroll et al.’s framework we propose may permit a comprehensive assessment of the implementation fidelity-adaptation balance required when implementing adaptive interventions, but more empirical research is needed to validate it.

In global health, researchers and decision makers, many of whom have medical, epidemiology or biostatistics background, are increasingly interested in evaluating the implementation of health interventions. Implementation science, particularly for the study of public policies, has existed since at least the 1930s. This science makes compelling use of explicit theories and analytic frameworks that ensure research quality and rigour. Our objective is to inform researchers and decision makers who are not familiar with this research branch about these theories and analytic frameworks. We define four models of causation used in implementation science: intervention theory, frameworks, middle-range theory and grand theory. We then explain how scientists apply these models for three main implementation studies: fidelity assessment, process evaluation and complex evaluation. For each study, we provide concrete examples from research in Cuba and Africa to better understand the implementation of health interventions in global health context. Global health researchers and decision makers with a quantitative background will not become implementation scientists after reading this article. However, we believe they will be more aware of the need for rigorous implementation evaluations of global health interventions, alongside impact evaluations, and in collaboration with social scientists.

Dengue virus infections are a serious public health problem worldwide. Aedes aegypti is the primary vector of dengue in Cuba. As there is no vaccine or specific treatment, the control efforts are directed to the reduction of mosquito populations. The indiscriminate use of insecticides can lead to adverse effects on ecosystems, including human health. The sterile insect technique is a species-specific and environment-friendly method of insect population control based on the release of large numbers of sterile insects, ideally males only. The success of this technique for the sustainable management of agricultural pests has encouraged its evaluation for the population suppression of mosquito vector species. Here, we describe an open field trial to evaluate the effect of the release of irradiated male Ae. aegypti on a wild population. The pilot trial was carried out in a suburb of Havana and compared the mosquito population density before and after the intervention, in both untreated control and release areas. The wild population was monitored by an ovitrap network, recording frequency and density of eggs as well as their hatch rate. A significant amount of sterility was induced in the field population of the release area, as compared with the untreated control area. The ovitrap index and the mean number of eggs/trap declined dramatically after 12 and 5 weeks of releases, respectively. For the last 3 weeks, no eggs were collected in the treatment area, clearly indicating a significant suppression of the wild target population. We conclude that the sterile males released competed successfully and induced enough sterility to suppress the local Ae. aegypti population.

The reemergence of dengue as an important public health problem reflects the difficulties in sustaining vertically organized, effective, control programs and the need for community-based strategies for Aedes aegypti control that result in behavioral change. We aimed to disentangle the relationships between underlying determinants of dengue related practices. We conducted a cross-sectional study in 780 households in La Lisa, Havana, Cuba. A questionnaire and an observation guide were administrated to collect information on variables related to economic status, knowledge on dengue, risk perception and practices associated with Aedes aegypti breading sites. To test a conceptual model that hypothesized direct relationships among all these constructs, we first used Exploratory Factor Analysis with Principal Component Analysis to establish the relationship between observed variables and the underlying latent variables. Subsequently, we tested whether the observed data supported the conceptual model through Confirmatory Factor Analysis. Exploratory Factor Analysis indicated that the items measured could be reduced into five factors with an eigenvalue >1.0: Knowledge on dengue, Intradomiciliar risk practices, Peridomiciliar risk practices, Risk perception and Economic status. The proportion of the total variance in the data explained by these five factors was 74.3%. The Confirmatory Factor Analysis model differed from our hypothesized conceptual model. Only Knowledge on dengue had a significant, direct, positive, effect on Practices. There was also a direct association of Economic status with Knowledge on dengue, but not with Risk perception and Practices. Clarifying the relationship between direct and indirect determinants of dengue related practices contributes to a better understanding of the potential effect of Information Education and Communication on practices and on the reduction of Aedes aegypti breeding sites and provides inputs for designing a community based strategy for dengue control.

The non-sustainability of vertically organised dengue vector control programmes led to pleas for changing the emphasis towards community-based strategies. We conducted a cluster randomised controlled trial with 16 intervention and 16 control clusters to test the effectiveness of a community empowerment strategy intertwined with the routine dengue vector control programme in La Lisa, Havana City, Cuba. The intervention included four components on top of routine control: organisation and management; entomological risk surveillance; capacity building; and community work for vector control. In the control clusters, routine activities continued without interference. The community participation score increased from 1.4 to 3.4. Good knowledge of breeding sites increased by 52.8% and 27.5% in the intervention and control clusters, respectively. There were no changes in adequate Aedes aegypti control practices at household level in the control clusters, but in the intervention clusters adequacy increased by 36.2%. At baseline, the Breteau indices (BI) were approximately 0.1 and were comparable; they fluctuated over time but became different with the launch of the community-based dengue control activities in the intervention clusters. Over the intervention period, the BI remained 53% (95% CI 22-92%) lower in these clusters than in the control clusters. The empowerment strategy increased community involvement and added effectiveness to routine A. aegypti control.