Strengthening scaling up through learning from implementation: comparing experiences from Afghanistan, Bangladesh and Uganda

Background: Many effective innovations and interventions are never effectively scaled up. Implementation research (IR) has the promise of supporting scale-up through enabling rapid learning about the intervention and its fit with the context in which it is implemented. We integrate conceptual frameworks addressing different dimensions of scaling up (specifically, the attributes of the service or innovation being scaled, the actors involved, the context, and the scale-up strategy) and questions commonly addressed by IR (concerning acceptability, appropriateness, adoption, feasibility, fidelity to original design, implementation costs, coverage and sustainability) to explore how IR can support scale-up.

Methods: We draw upon three IR studies conducted by Future Health Systems (FHS) in Afghanistan, Bangladesh and Uganda. We reviewed project documents from the period 2011-2016 to identify information related to the dimensions of scaling up. Further, for each country, we developed rich descriptions of how the research teams approached scaling up, and how IR contributed to scale-up. The rich descriptions were checked by FHS research teams. We identified common patterns and differences across the three cases.

Results: The three cases planned quite different innovations/interventions and had very different types of scale-up strategies. In all three cases, the research teams had extensive prior experience within the study communities, and little explicit attention was paid to contextual factors. All three cases involved complex interactions between the research teams and other stakeholders, among stakeholders, and between stakeholders and the intervention. The IR planned by the research teams focussed primarily on feasibility and effectiveness, but in practice, the research teams also had critical insights into other factors such as sustainability, acceptability, cost-effectiveness and appropriateness. Stakeholder analyses and other project management tools further complemented IR.

Conclusions: IR can provide significant insights into how best to scale-up a particular intervention. To take advantage of insights from IR, scale-up strategies require flexibility and IR must also be sufficiently flexible to respond to new emerging questions. While commonly used conceptual frameworks for scale-up clearly delineate actors, such as implementers, target communities and the support team, in our experience, IR blurred the links between these groups.