In my last blog post, I wrote about how some of what we do in healthcare is outdated or just plain wrong. Yet, physicians often persist in continuing these practices. For example, the Society of Hospital Medicine recommends that doctors do not routinely order daily chest x-rays for hospitalized patients unless there are specific clinical indications. Still, a recent study found that only about half of surveyed critical care clinicians reported any attempt to operationalize such recommendations. Why is this? Why do we continue to practice medicine as we did four or five decades ago, even when we know better?
In a 2017 editorial, researchers succinctly summarized the issue, “While there are likely multiple causes to the problem of overtreatment, an under-recognized cause may be the inability to effectively and in an evidence-based manner stop habitual behaviors and care practices that consume time but have low-value.” They shared that, “Simply urging or informing health care providers to abandon these kinds of outdated and low-value clinical practices often fails to extinguish unwanted habits.” It is not enough to ask (or even mandate) that clinicians stop performing certain procedures or ordering particular medications; we need to study how successful implementations work and apply those learnings to the opposite process: de-implementation.
Scholars have studied the terms researchers use in addition to de-implementation. In a BMC Medicine article, the authors identified 43 terms used to describe the same principle. The most frequent is disinvestment, followed by decrease use, discontinuation, and abandonment. In fact, de-implementation was closer to the bottom with respect to frequency. No matter; I am going to use that word because I think it quickly and effectively communicates that all the work and thought that goes into starting a new project or instituting a complex therapy must also be applied in reversing long-held beliefs, training, and even muscle memory.
If you are a frequent reader of this blog, you will likely know that I often comment that one of our most important jobs as clinical informaticians or operational leaders is to make it easy to do the right thing and slightly difficult to do the wrong thing (keeping in mind that sometimes the “wrong” thing is the right thing given certain circumstances). De-implementation experience shows that “making it slightly difficult to do the wrong thing” may not be enough. Case in point: years of research have demonstrated little value in treating children with uncomplicated viral bronchiolitis with a medicine called albuterol. When I was in training back in the Stone Age, nebulized albuterol was the standard of care for kids with bronchiolitis. We know better now, yet this medication is still often ordered anyway, especially in emergency departments.
Clinicians at the Children’s Hospital of Philadelphia have aimed to reduce albuterol use in their emergency department and hospital. In a research article published in Pediatrics, they write that initially, they simply removed the albuterol order from the electronic health record (EHR) bronchiolitis order set. This maneuver follows the recommendation to make it slightly more difficult to do the “wrong” thing. Since the medicine was no longer in the order set, physicians had to scroll to a box near the bottom of the page, type in, and then configure the albuterol order. This requires significantly more mental energy than clicking on a box associated with a perfectly configured order, so one (ok, I) might predict that most doctors would not do it. Guess what: one (ok, I again) would be wrong!
Many physicians still put forth the energy to manually order the albuterol despite its absence from the order set. Researchers learned that simply removing the ingrained and commonly used order annoyed clinicians and may have caused some to avoid the order set altogether. To achieve the desired effect and decrease the routine use of albuterol, they re-inserted the medication in the order set but attached a prominent statement just above the order recommending that it not be used. In their paper, they wrote that “[t]he inclusion of an undesired order in an order set with the explicit statement that albuterol is not recommended for routine use (the ‘do not order set’ approach, as opposed to leaving the option out) was a novel and effective strategy. This approach provides reinforcement to clinicians about an improvement treatment goal along with needed orders, such as an order for nasal suctioning that was encouraged as part of the pathway.”
Clearly, de-implementation calls for more than just informing physicians of updated data or recommendations. Research published earlier this year in the journal Implementation Science notes that patients themselves might need to be in the mix. The authors report that patient decision-making is a key determinant of the de-implementation of low-value surgery. Patients’ opinions and desires can sway their caregivers; hence ignoring their input often leads to sub-optimal results.
A well-thought-out approach with a rigorous methodology and ample research is required to convince clinicians to stop delivering outdated care. This is a team sport and requires clinical, operational, and technical leaders to row in the same direction to achieve the desired outcome: better patient care.