Changing physician behavior is hard—but comparing doctors' performance to their peers can help, Laura Landro reports for the Wall Street Journal.
Health care providers are testing different ways to use EHRs and other technologies to "nudge" clinicians into making better decisions, but it isn't easy to find an approach that works. "Doctors are ingenious in circumventing anything in a computer," explains Jeffrey Linder, director of the Primary Care Practice Based Research at Brigham and Women's Hospital. "If it makes their job harder they are just going to work around it."
Looking at the data
Consider antibiotic overuse. A recent study in JAMA found that nearly a third of antibiotics prescribed in the United States in outpatient settings are unnecessary. Most doctors know antibiotic overuse is a serious issue, but overprescribing persists anyway, in part because doctors feel pressured by patients and their family members to prescribe antibiotics.
In February, Linder and colleagues published a study in JAMA that examined the effectiveness of three electronic behavioral interventions aimed at reducing unnecessary antibiotic prescriptions for acute respiratory infections. The researchers gave about 250 physicians at 77 primary care practices a primer on appropriate prescribing. The doctors then received one or more electronic interventions over the course of 18 months:
- A "suggested alternatives" prompt which provided alternative treatments to antibiotics;
- An "accountable justification" prompt which required doctors to explain their treatment decision; and
- A "peer comparison" email, sent monthly, which showcased top-performing doctors and provided information on how a clinician's prescribing practices compared with those of his or her peers.
A control group received none of the electronic interventions.
The study found that the "suggested alternatives" prompt had no effect on prescribing practices, while the other two interventions had statistically significant effects. The most effective intervention was the peer comparison email, which was associated with a 16 percentage point drop in unnecessary prescriptions over the course of the study period, from 20 percent to 4 percent.
Linder notes that the peer comparison approach presented challenges. "We expected and got a lot of pushback, but it was the most effective intervention," he says.
Connecting with clinicians
Michael Healy, a physician in Brigham and Women's primary care group, says the peer comparison emails are exactly what some doctors need to stay on track. The emails are a "reality check" that can help doctors rethink their choices, he says.
But he notes that it is hard to design effective, evidence-based interventions that don't make doctors feel like they are being micromanaged. Hardeep Singh, a patient-safety researcher at the Michael E. DeBakey VA Medical Center and Baylor College of Medicine, agrees. The information in such prompts must be "accurate, presented to physicians at the right moment, perceived by them to be of high value, and must make it easier for them to do the next step," he says (Landro, Wall Street Journal, 6/26).
Advisory Board's take
John Kontor, EVP, Clinovations
Despite widespread investment in EHR technology, health systems are still struggling to deliver upon the promise of intuitive, patient-specific, workflow-embedded guidance.
Without clear and effective strategies to manage decision support interventions, attempts to expand clinical decision support (CDS) will only frustrate and disengage providers. One recent study, for instance, found that providers are inundated with an average of nearly 63 alerts per day— 91 percent of which are irrelevant or inaccurate.
Though vast stores of information live within the EHR, most health systems are still operating under a "deploy it and forget it" model, which diminishes the impact of CDS. Few systems have a governance model that assigns clinical and technical owners to monitor the effectiveness of decision support interventions after they go live. Even fewer have processes in place to share this adherence data back with providers. Should it surprise us that the intervention that performed best in the JAMA study is the only intervention that shared any performance data back with the end user?
Ultimately, organizational leaders must maintain a manageable EHR environment with CDS content that providers perceive as valuable and impactful. Until that happens, health systems will struggle to achieve a uniformly high standard of care, and attempted CDS interventions will too often prove to be ineffective, inflexible, and invasive.
To learn more about how we can help you meet these challenges, please contact Mike Murray at firstname.lastname@example.org.