A previous post in this series highlighted the critical importance of the patient perspective in de-implementation efforts. At the provider level, there is a rapidly growing evidence base on interventions that can work to discourage providers from overusing services. Yet, there remain multiple areas of uncertainty regarding de-implementation efforts at the provider level that could be addressed by de-implementation research in learning health systems. Three such areas include the impact and design of educational efforts, the role of psychological factors on decision-making and the effectiveness of interventions that target these factors, and the impact on providers of sometimes conflicting incentives and interventions at the provider and organizational level.
The impact of provider education on de-implementation efforts has not been well studied. Education regarding the ineffectiveness of longstanding widely used services appears to work better in some instances (e.g., hormone replacement therapy for prevention of chronic disease) than in others (e.g., antibiotics for acute upper respiratory infections). Further, the design of educational efforts that may work best for different types of clinical problems or health care services are unknown. For example, for some clinical areas general education about the latest evidence may be sufficient, but in others more intensive academic detailing may be needed. Learning health systems could help identify which types of education most effectively reduce overuse in different situations, as well as when education needs to be paired with other strategies in order to be effective. Earlier educational efforts during medical training are likely to be critically important in helping trainees develop the skill set needed to flexibly abandon conventional approaches when future evidence suggests that those approaches should be de-implemented.
Another area in which de-implementation research could fill knowledge gaps is the role of psychologic factors in clinical decision-making, and the effectiveness of leveraging these factors to prompt appropriate decisions. Such approaches are becoming increasingly important as the medical evidence base continues to push the boundaries of providers’ cognitive capacities, making strategies that rely solely on provider education less likely to yield meaningful and sustained reductions in overuse. For example, in some cases providers’ decision-making may be driven by their motivation to avoid errors of omission, with less concern about making errors of commission. Another important barrier may be provider concerns about maintaining patients’ trust in light of new clinical recommendations to avoid services that had previously been recommended to them. De-implementation research in learning health systems could help identify areas in which such psychologic factors and cognitive biases contribute most to overuse and open the door to new strategies to address these factors.
Finally, providers are increasingly inundated with a variety of government, health system, and payer-led initiatives that seek to influence their clinical decisions. These initiatives often take the form of public reporting, pay-for-performance, and quality improvement programs. The impact of the interaction between these sometimes-conflicting incentives that compete for providers’ limited attention is unknown. If interventions to reduce overuse are to successfully compete with other programs, they will need to consider the broader context in which they are aiming to shift clinical decisions and account for other initiatives vying for providers’ efforts.
This four-part blog series is a product of a thematic working group on de-implementation convened by AcademyHealth and the ABIM Foundation as part of the Research Community on Low-Value Care. Access the first two posts in the series focused on viewing de-implementation as part of a learning health system and addressing patient perspectives.