Too deductive too soon? Toward an inductive renewal of implementation science

Abstract Background Over the past two decades, implementation science has developed a strong conceptual foundation through the proliferation and widespread use of theories, models, and frameworks (TMFs). These have provided coherence, shared vocabulary, and methodological discipline across a rapidly expanding field. However, this success has also produced an unintended consequence: increasing reliance on deductive modes of inquiry, in which a limited set of established TMFs are repeatedly applied as analytic templates across diverse empirical contexts. This tendency toward early deductivism risks constraining theoretical development, reducing sensitivity to heterogeneity, complexity, and temporality inherent in implementation, and reinforcing methodological circularity. Toward an inductive renewal of implementation science In this conceptual paper, we argue for an inductive renewal of implementation science that rebalances deduction with stronger inductive and abductive forms of reasoning. Rather than abandoning established TMFs, we propose reframing them as evolving heuristics – resources for structuring inquiry that remain open to refinement, extension, and selective reconfiguration through empirical engagement. We clarify the complementary roles of induction, abduction, and deduction in theory development, emphasizing abductive iteration as a mechanism for translating empirical discovery into cumulative conceptual advancement. We outline strategies for advancing this agenda across three interdependent levels. At the study level, this involves treating TMFs as provisional heuristics, re-embracing qualitative discovery, and using abductive reasoning to refine theory through engagement with unexpected findings. At the field level, shared infrastructures for synthesis and longitudinal learning are needed to support cumulative, context-sensitive theorizing and to account for the temporal dynamics of implementation. Institutionally, journals and funders must recalibrate incentives to value theory development, adaptation, and transparency alongside theory application. Drawing on examples from research on knowledge brokering and implementation scale-up, we show how theoretically informative contributions emerge when empirical surprises, temporal dynamics, and analytic tensions are used to interrogate and refine existing TMFs rather than being absorbed into pre-specified categories. Conclusion A mature implementation science must move beyond asking which TMF best fits a study, toward examining how empirical phenomena challenge, extend, and reshape theory. Sustaining this balance is essential for theoretical coherence and continued conceptual innovation.