Loading field map
Why do proven medical interventions take years, even decades, to reach patients? This question has driven a distinct subfield within health services research: implementation and dissemination science. The central pressure is not discovering what works—that is the job of clinical trials and comparative effectiveness research—but understanding how to make what works actually happen in real-world settings. Over the past sixty years, researchers have moved from describing how innovations spread naturally to designing deliberate strategies for embedding evidence into routine care. The story of this subfield is one of increasing sophistication: from passive diffusion models to active translation frameworks, and finally to multi-level theories that treat implementation itself as a complex, dynamic process.
The subfield's first major framework, Diffusion of Innovations (1962–1990), emerged from rural sociology. Everett Rogers showed that new ideas spread through populations in an S-shaped curve, driven by characteristics like relative advantage, compatibility, and observability. This framework treated adoption as a largely passive, social process. Researchers could predict which innovations would spread, but they had little guidance for accelerating adoption. The framework's strength was its descriptive power; its weakness was its limited prescriptive reach. It described how change happened, not how to make it happen.
By the 1990s, the gap between research and practice had become a recognized crisis. The Knowledge Translation framework (1990–2005) responded by reframing the problem as an active communication challenge. Knowledge translation emphasized that researchers must tailor messages to specific audiences—clinicians, managers, policymakers—and remove barriers to uptake. This was a narrowing and transformation of diffusion theory: instead of waiting for innovations to spread, knowledge translation called for deliberate push-and-pull strategies. It also introduced the concept of the "knowledge-to-action cycle," distinguishing between knowledge creation and the action steps needed to apply it. However, knowledge translation remained largely a conceptual map; it did not provide detailed guidance on how to design implementation strategies in complex healthcare organizations.
The late 1990s and early 2000s saw the rise of frameworks that treated implementation as a phenomenon requiring its own theory, not just an extension of diffusion or translation. Promoting Action on Research Implementation in Health Services (PARIHS, 1998–Present) argued that successful implementation depends on the interplay of three core elements: evidence (its nature and strength), context (the setting's culture, leadership, and resources), and facilitation (the process of enabling change). PARIHS was a breakthrough because it moved beyond describing adoption to specifying conditions for success. It coexisted with knowledge translation but shifted attention from message design to organizational readiness. Today, PARIHS remains active, often used as a diagnostic framework to assess whether a setting is ready for change before an implementation effort begins.
Almost simultaneously, RE-AIM (1999–Present) offered a practical evaluation lens. RE-AIM stands for Reach, Effectiveness, Adoption, Implementation, and Maintenance. Unlike PARIHS, which focused on preconditions, RE-AIM provided a framework for measuring the public health impact of an intervention across multiple dimensions. It narrowed the field's attention to external validity: an intervention that works in a controlled trial may fail if it cannot reach diverse populations, be adopted by varied settings, or be maintained over time. RE-AIM and PARIHS complemented each other—one assessing readiness, the other evaluating outcomes—and both remain widely used today, often in combination.
By the mid-2000s, the field had accumulated many concepts but lacked a unified vocabulary. The Interactive Systems Framework for Dissemination and Implementation (ISF, 2005–Present) addressed this by distinguishing three interacting systems: the prevention delivery system (where implementation happens), the prevention support system (training and technical assistance), and the synthesis and translation system (distilling evidence). ISF absorbed ideas from both knowledge translation and PARIHS but organized them into a systems-level architecture. It was particularly influential in public health, where implementation often involves multiple organizations working together.
Also in 2005, the Theoretical Domains Framework (TDF) emerged from a collaboration of behavioral scientists. TDF synthesized 33 behavior change theories into 14 domains (e.g., knowledge, skills, social influences, beliefs about consequences). Its contribution was to provide a comprehensive, theory-based checklist for diagnosing barriers to clinician behavior change. TDF narrowed the field's focus to individual and team behavior, complementing PARIHS's organizational lens. Researchers often use TDF to design implementation strategies by identifying which domains are most relevant in a given context.
Practical, Robust Implementation and Sustainability Model (PRISM, 2008–Present) integrated elements from RE-AIM and PARIHS into a single framework that explicitly addressed sustainability—a dimension earlier models had neglected. PRISM emphasized that implementation is not a one-time event but a process requiring ongoing adaptation. It broadened the field's temporal horizon, asking not just "Did it work?" but "Did it last?"
In 2009, the field produced three frameworks that remain central today. The Consolidated Framework for Implementation Research (CFIR) was a landmark synthesis. Its developers reviewed 19 published implementation theories and extracted five major domains: intervention characteristics, outer setting, inner setting, individual characteristics, and process. CFIR did not introduce new concepts; rather, it organized existing ones into a common language. This made it an infrastructure framework—a shared reference point that researchers could use to describe their contexts and compare findings across studies. CFIR quickly became the most widely cited implementation framework, not because it was the most theoretically novel, but because it provided a practical, comprehensive checklist for implementation planning and evaluation.
Also in 2009, Exploration, Preparation, Implementation, Sustainment (EPIS) offered a phase-based model. Unlike CFIR, which was a static list of factors, EPIS described implementation as a temporal process with distinct stages. Each stage—exploration (deciding whether to adopt), preparation (planning), implementation (active rollout), and sustainment (maintenance)—has its own key determinants and strategies. EPIS complemented CFIR by adding a dynamic, developmental perspective. Researchers often use them together: CFIR to identify what matters, EPIS to know when it matters.
Normalization Process Theory (NPT, 2009–Present) took a different approach. Instead of listing factors or stages, NPT asked: what work do people have to do to make a new practice routine? It identified four generative mechanisms—coherence (sense-making), cognitive participation (engagement), collective action (work), and reflexive monitoring (appraisal)—that explain how practices become embedded or "normalized." NPT was a revival of sociological process theory in a field that had become dominated by factor lists. It remains active today, especially in qualitative studies that seek to understand why some implementations succeed while others stall.
Since 2010, Complexity Science Approaches have challenged the assumption that implementation can be fully controlled or predicted. Drawing on concepts from systems theory—feedback loops, emergence, path dependence, adaptation—complexity science treats healthcare settings as dynamic systems where interventions interact unpredictably with local contexts. This perspective does not replace earlier frameworks but transforms how they are used. For example, a complexity-informed researcher might use CFIR not as a fixed checklist but as a starting point for adaptive, iterative implementation. Complexity science also revived interest in formative evaluation and rapid-cycle testing, methods that allow strategies to evolve in real time.
Today, the leading frameworks coexist in a division of labor. PARIHS and TDF are often used for pre-implementation diagnosis: PARIHS assesses organizational context, TDF assesses individual barriers. RE-AIM remains the standard for evaluating public health impact. CFIR serves as the common language for describing implementation contexts across studies. EPIS provides a temporal roadmap for multi-year implementation projects. NPT offers a process-oriented lens for understanding how practices become routine. Complexity science acts as a meta-theoretical corrective, reminding researchers that no framework can fully capture the messiness of real-world change.
Despite their diversity, today's leading frameworks agree on several points. First, implementation is multi-level: it involves individuals, teams, organizations, and broader policy systems. Second, context matters enormously—what works in one setting may fail in another. Third, implementation is not a one-time event but an ongoing process that requires adaptation and sustainment. Fourth, no single framework is sufficient; researchers often combine frameworks to address different aspects of a problem.
The main disagreements center on what to prioritize. Some frameworks (TDF, NPT) emphasize individual and team cognition and behavior; others (PARIHS, CFIR) give equal weight to organizational and structural factors. Some (EPIS, RE-AIM) focus on stages and outcomes; others (complexity science) argue that linear stage models oversimplify the process. A deeper tension concerns generalizability: CFIR and EPIS aim for frameworks that apply across settings, while complexity science suggests that each implementation is so context-dependent that general frameworks have limited predictive power. This disagreement is productive—it keeps the field from settling into a single orthodoxy and drives ongoing refinement of methods and theories.