Implementation Science – Ensuring the return on our research investment

In their article, Implementation – The Missing Link in the Research Translation Pipeline, Lynch and colleagues report that only a small fraction of published stroke rehabilitation research in leading clinical journals evaluates the implementation of evidence-based interventions into health care practice (1). Their findings are a wake-up call. If we are to achieve the end goals of our research investment and improve population health, then we need to ensure that the evidence we generate is translated into real-world use.

Implementation is the use of strategies to adopt and integrate evidence-based health interventions and change practice patterns. Implementation science studies the process of disseminating health information and integrating evidence into practice at individual-, organizational-, and community-levels. The concept embodied in “implementation” goes by different names depending upon the context – for example, Dissemination & Implementation (D&I) research in the U.S., Knowledge Translation (KT) in Canada, Knowledge to Action in public health, and Continuous Quality Improvement in learning healthcare systems. Implementation science is multi-disciplinary drawing upon expertise in clinical care and public health, health services research, social and behavioral sciences, systems engineering, and marketing and communication. Recognizing that implementation is a human process, social science frameworks and theories inform how we design for dissemination and how we evaluate and affect the social mechanisms of action involved in the decision to de-implement outdated practices, to translate new evidence and sustain behaviors into routine care.

Implementation encompasses a full spectrum of scientific methodologies in its research arsenal. Qualitative methods from medical anthropology and sociology help us understand knowledge, attitudes and beliefs and identify the mutable factors affecting health behaviors and clinical practice. These insights inform where to target interventions for improving evidence translation. Formative research informs dissemination messaging and communication strategies. Large simple studies, cluster-randomized pragmatic trials and effectiveneness-implementation hybrid trial designs compare implementation interventions in real-world settings (2). Quasi-experimental studies using electronic health data are another means for studying the effectiveness of intervention implementation at scale while balancing internal validity goals with external validity considerations (such as diverse subpopulations, cost, and sustainability) (3). The use of mixed methods (e.g., pre-post interviews-surveys coupled with clinical trials) is increasingly being used to understand pragmatic trial results and the success or failure of implementation efforts (4). Mixed methods increase our understanding of the context in which interventions are implemented and provide critical information necessary for evaluating an intervention’s implementation-readiness, i.e., generalizability and potential for scalability. As the field of implementation science in health research has matured, standards for reporting implementation studies have been reported (5, 6) and a seminal textbook published, now in its second edition (7).

So why might implementation research be under-represented in the stroke rehabilitation clinical literature? The authors discuss some possibilities. One relates to the fact that in a multi-disciplinary research field, the body of work can become scattered in different scientific journals and thereby disseminate implementation scholarship amongst different research communities. The authors found evidence of this possibility in their exploration of where rehabilitation implementation findings were published. While this may not be an issue for researchers, accustomed to searching the scientific literature via PubMed and other tools, it is a barrier for clinicians who may follow just one or two clinical journals. The net effect is that implementation research appears invisible. Journals, like Neurorehabilitation and Neural Repair, could advance cross-pollination of the disciplines by encouraging submission of implementation research and/or spotlighting implementation research reported in other journals.

Another possibility is that implementation research does not have the same funding priority resulting in less implementation research to report. Historically, dissemination and implementation research has been funded “pennies on the dollar” when compared with public funding of basic science and intervention efficacy research (8). However, there is growing recognition on the importance of real-world evidence as a condition for behavior change and for promoting adoption of evidence-based strategies. Diffusion of Innovation Theory teaches us that early innovators are willing to take a risk on a good idea or anecdotal evidence when adopting new behaviors; however, the majority of people want solid evidence in similar settings and organizations as their own. Without this type of applied evidence, there is a translational gap. Comparative effectiveness research, pragmatic trials and other forms of real-world evidence seek to bridge this gap and thereby contribute to the field of implementation research, whether or not the work is labeled as an implementation study.

The authors discuss promising signs that the implementation research funding landscape may be changing, providing examples in Canada and Australia. In the United Stated, the Patient Centered Outcomes Research Institute (PCORI) has been critical in funding and influencing research relevant to stakeholders and designed with dissemination and implementation in mind from the beginning (9). The National Institutes of Health National Center for Advancing Translational Sciences funds major translational science infrastructure through Clinical and Translational Science Awards (CTSAs) in more than 50 academic medical research centers across the U.S. CTSA hubs are expected to develop and demonstrate solutions to translational roadblocks and dissemination of successful solutions is an explicit goal and expectation. These hubs are also charged with training and development so that we have the skilled research workforce capability of designing and conducting implementation research.

If we are to achieve the end goal of our research, that is, to improve individual and population health, then implementation science is indeed a missing link. What private sector company invests only in research and development without investment in its marketing and sales organizations (their equivalent implementers)? That is not a sound business model. Likewise, a failure to prioritize, study and celebrate implementation and translation of publicly funded medical research evidence into practice is not a sound taxpayer model. The report by Lynch and colleagues underscores the implementation gap in rehabilitation and stroke research and charges us all with closing that gap.