Deimplementation of ineffective and harmful medical practices: a data-driven commentary.

Deimplementation is the discontinuation or abandonment of medical practices that are ineffective or of unclear effectiveness, ranging from simply unhelpful to harmful. With epidemiology expanding to include more translational sciences, epidemiologists can contribute to deimplementation by defining evidence, establishing causality, and advising on study design. An estimated 10%-30% of health care practices have minimal to no benefit to patients and should be targeted for deimplementation. The steps in deimplementation are (1) identify low-value clinical practices, (2) facilitate the deimplementation process, (3) evaluate deimplementation outcomes, and (4) sustain deimplementation, each of which is a complex project. Deimplementation science involves researchers, health care and clinical stakeholders, and patient and community partners affected by the medical practice. Increasing collaboration between epidemiologists and implementation scientists is important to optimizing health care delivery.