Section 11: De-Implementation in Pragmatic Research

11.5 De-implementation Outcomes and Equity Considerations

[CONTENT NOTE: THIS IS FROM HELFRICH (2022), PERMISSION NEEDED TO USE]

There are several considerations for assessing outcomes when studying de-implementation or developing de-implementation programs.

Unintended consequences. Psychological reactance (anger & mistrust): both patients and providers could potentially experience de-implementation efforts as an infringement, for patients on their right to receive services, and for providers on their professional autonomy.

Intervention-outcome asymmetry. In implementation efforts, the intervention outcome is the benefit from implementing the evidence-based practice; at least in principle, the clinician or practitioner who is implementing the evidence-based practice is delivering some benefit to their patient. However, for de-implementation efforts, the expected benefit is typically an absence of bad outcomes – from the provider’s or practitioner’s perspective, the best expected outcome is often literally nothing: a low-value inhaler is eliminated and the patient doesn’t experience a breathing exacerbation; a low-value cancer screening is forgone and the patient never develops cancers; a patient who has an upper respiratory infection doesn’t receive an antibiotic and it resolves on its own in a couple of weeks. The problem this creates for the provider or practitioner is that they may experience a real risk of a bad outcome, e.g., an angry patient or a random bad event, and conversely fail to perceive any real benefit. More so than implementation interventions, de-implementation may require implementation researchers to engineer feedback that helps reveal the benefits to stakeholders and create positive reinforcement for de-implementation.

Outcome measurement. Proctor and colleagues established a set of outcomes specific to implementation, meaning factors that implementation strategies could or need to influence in order to achieve high levels of implementation. These include acceptability, adoption, appropriateness, feasibility, fidelity, implementation cost, penetration, and sustainability. These can be equally applied to much de-implementation work, but in some cases, e.g., certain tests and imaging, patient and provider (or other stakeholder) perceptions of the low-value practice are difficult to assess and even potentially introduce confusion.

Equity challenges. Within insured populations (e.g., a given managed care organization, or within Medicare), we can end up in a situation
where some patients subsidize low-value care delivered to other patients. This is because all patients pay into insurance, but receipt of low-value care can vary substantially among patients. There is some research that finds more socially or economically advantaged patients are more likely to receive low-value care, while other research suggests that socioeconomically-disadvantaged patients are subject to both more low-value care & less high-value care, or that these populations may receive less care overall – less high-value care but also less low-value care. There are also documented differences among patients by race and gender in their relative concerns about overuse and under-use, which have profound implications for how patients respond to de-implementation efforts.