Written by Charles Goss, PhD, assistant professor of biostatistics and medicine, Washington University School of Medicine in St. Louis
On January 11, the Center for Dissemination and Implementation hosted Daniel Almirall, PhD, associate professor at the University of Michigan, for his talk, “Multi-level Adaptive Implementation Strategies: Design Principles, Optimization Questions and Choosing the Right Experimental Design”. This event brought together 90 global participants. Professor Almirall, both a statistician and clinical trialist, gave a thought-provoking presentation on research that has important implications for implementation science.
By way of a brief introduction to the field, implementation science focuses on methods (aka: implementation strategies) to facilitate the use of evidence-based practices (EBPs) in routine care. Examples of implementation strategies include providing educational resources, financial incentives, and ongoing feedback to healthcare providers. Because implementation science is an emerging scientific field, there is still ongoing research regarding the most appropriate implementation strategies to maximize delivery of EBPs within and across different contexts.
This is where Professor Almirall’s research comes in. He proposes a new approach to both conceptualizing and guiding implementation at multiple levels (e.g., system, clinic, clinician). His proposed approach integrates adaptive treatments (i.e., flexible approaches that can be tailored to patients based on history of treatment responses) and implementation strategies to yield a new approach to implementation science termed “MAISYs” (Multilevel Adaptive Implementation Strategies).
At the most basic level, MAISYs can be used as tools to make decisions regarding the best strategies for improving outcomes. Implementation strategies that are not effective in a given context are further adjusted such that they better meet the needs of the community to which they are applied. For example, if holding educational meetings on implementing a new treatment at the healthcare system level does not improve the health of patients, then there may be a decision that would lead training that is more intensive at the clinic level. This approach provides an “instruction manual” for healthcare providers that enables both simple and tailored solutions for improving health outcomes. Importantly, MAISYs allow for flexibility to meet the changing needs both within a particular clinic and across clinics and providers. Not only do MAISYs have practical implications, they also have important implications for conducting implementation research. Rigorous study designs such as randomized clinical trials can be used to evaluate whether a particular MAISY outperforms usual care and MAISYs can be further optimized using more sophisticated clinical-trial design approaches (e.g., Sequential Multiple Assignment Trials [SMARTs]).
Overall, the work of Professor Almirall and colleagues at the University of Michigan represents a promising development in the field of implementation science that can ultimately help to improve healthcare systems (and therefore the lives of people) around the world.
This seminar was part of the Dissemination and Implementation Science seminar series hosted by the Center for Dissemination and Implementation at the Institute for Public Health. The center advances the growing body of Dissemination & Implementation (D&I) research methods by building training opportunities and catalyzing newly applied D&I research across health specialties. The center ensures that the most effective services are delivered in clinical and public health settings.