A Holistic Approach

Implementation research? Yes – but what kind of research do we need?

Von Joan Muela Ribera and Koen Peeters Grietens

Nowadays few people would argue against the need of accompanying project implementation with some form of research. There is a general thirst for scientific evidence in all implementing agencies, from global multilateral organisations to local NGOs, and yet researchers and implementers still move on stony paths when engaging in collaboration, often resulting in deep frustrations and limited fruitful outcomes.

Implementation research? Yes – but what kind of research do we need?

Improving insecticide treated hammocks’ design with the community, as part of the implementation research process (Vietnam)


Implementers tend to argue that researchers do not meet their expectations or needs while researchers complain about a lack of interest in scientific solidity on the implementers’ side. No doubt, this dissonance can generally be traced back to misunderstandings and divergent priorities. The conundrum for effective collaboration is therefore how to develop a research process that accounts for the intrinsically dynamic and complex nature of the projects without violating the principles of scientific rigor.

Marrying implementation and research

In our field implementation research is applied research accompanying and informing projects targeted at improving the health of populations. In other words, it is both research for and research of implementation that aims at improving the implementation and also at documenting it and at building related theory. It’s mainly centred on:

  1. elaborating project or program strategies and putting them into action; often involving communities or assessing their response to the implementation (why, under what circumstances and how does an implementation succeed or fail, and how can it be adapted in order to thrive);
  2. evaluating the results of the implementation and its planned and unplanned consequences;
  3. developing conceptual frameworks and models allowing both the comparison and adaptation of implementations to different socio-cultural contexts and the further development of the general model.

What kind of research do we need?

Implementations, whether targeting behavioural change or introducing/scaling up medical resources and strategies (preventive measures, diagnostic or treatment tools, health system strengthening) are overtly complex and dynamic processes. A rigid, predetermined, and strictly quantitative research design is unlikely to provide in-depth insights in project-related change and is not the right tool to meet programmatic challenges during the course of implementation as this requires adapting the design when these challenges appear instead of at predetermined intervals.  But then what kind of science and research methodology is appropriate for project implementation in health? What kind of research is required when we want to start from local realities, optimize interventions during their course and offer measures that include relevant stakeholders from the beginning?

In response to these questions, we argue that implementation research, if it is to be capable of meeting the expectations of both researchers and implementers, needs to be holistic, transdisciplinary and flexible.

A holistic approach

Great emphasis has been placed on the need to consider the socio-cultural context for understanding enabling factors and bottlenecks to an implementation. However, current practices tend to focus on a reduced number of factors, typically identified using KABP (knowledge, attitudes, behavior and practices) surveys; or, on analyzing ‘key factors’ in isolation from their socio-economic and political contexts. Such approaches have been widely criticized for being reductionist. In order not to omit relevant, but often less obvious explanations, we advocate the use of large scope theoretical models, e.g. socio-ecological models or access models. These models offer a holistic perspective, allowing researchers to categorize, discard/select, interconnect and evaluate in a contextualized manner all possible factors at play, leading to a systematic and systemic research process. In addition, in relation to the implementation, a holistic perspective accounts for the socio-political issues and power relations at stake, and includes all relevant actors and institutions during the process.

A systemic analysis, e.g. studying the socio-political structure, economic strategies, social values, beliefs and social practices, allows an understanding of the problems affecting the course of an intervention and the promotion of feasible and sustainable solutions. In this sense, via its analysis of the context throughout the whole process, implementation research aims to continuously orient implementers and provide the basis for decision-making.

Villagers contributed to improve the design of an insecticide-treated hammock in Central Vietnam © Joan Muela

A transdisciplinary and multi-sectorial approach

Due to the complexity of the problems, of the solutions to be implemented, and of the very context of implementation a transdisciplinary and multisectorial approach is required. A transdisciplinary approach means starting from a concrete problem to find solutions that are from the outset jointly defined and developed by the various relevant disciplines (e.g. social and economic sciences, epidemiology and public health) and involve the key players for the development of the implementation (e.g. policy makers, NGOs, donors, communities).

Likewise, a multisectorial approach implies working together with the different sectors involved, both at the institutional level (public, private, civil society) and of the administration (e.g. finance, agriculture, housing, health, education).

Flexible and iterative research

The course of a project is neither static nor predictable. In addition, an intervention aims at producing the best possible outcome instead of a predefined outcome measure. Likewise, implementation research cannot be deadlocked due to a misunderstanding of methodological rigor that does neither permit changes nor adjustments to the implementation research strategy. Conversely, it is noxious to call a mere documentation of activities ‘research’, or to undertake monitoring and evaluation without a scientifically constructed methodology, including e.g. valid qualitative or quantitative sampling strategies and analytic frames.

One of the problems when researching an implementation is often the overemphasis on quantification, where reliance on numbers is the sole valid way to generate evidence on and determine the magnitude of a problem. However, there is common awareness that quantitative designs per se do not allow us to comprehend the systems that we study and its interactions, nor to adjust the implementation process. In response, there is an increasing trend of systematically introducing qualitative methods in the monitoring and evaluation field. Complementing qualitative, quantitative and participatory methods (with increasing community participation) in so-called transdisciplinary mixed methods approaches are a way forward for conducting implementation research as this allows both flexibility and quantification.

The role of theory

Theory is sometimes perceived as produced by academics in their ivory towers, without apparent practical utility or relation to the interests of health projects. Nothing is less true. Theory helps us understand what is going on and allows us to formulate research questions, to open perspectives and detect problems that are not self-evident. Theory is necessary to organize action, and the results of an implementation are the best way to test the hypotheses of the project. Theory also permits us to generalize and to establish the bases for the scaling-up of interventions, adapting them to different contexts, as well as to compare the results of these interventions across contexts.

No success without dialog 

Implementation research is not an exercise in paying lip service to trendy words and nor is it a do-it-yourself activity of a public health implementer in the field. Implementation research requires disciplinary training, skills and capabilities that go beyond mere applications of basic methods. It requires a mutual understanding of people with different theoretical and practical backgrounds, and above all, the willingness to engage in an open, unbiased dialog with different stakeholders. In this sense, we welcome an NGO-driven platform that involves universities (in the North and the South), and health implementers (from the North and the South) to sit around the table and start discussing and acting together to address concrete problems and seek solutions together based on rigorous but flexible methodologies, social commitment, and strong theoretical underpinnings.

Joan Muela Ribera and Koen Peeters Grietens

Joan Muela Ribera
PASS-Suisse, Neuchâtel, Switzerland,

Koen Peeters Grietens
Medical Anthropology Unit, Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium,