Inappropriate antibiotic use for acute bronchitis decreased with decision support

Decision support resulted in less inappropriate use of antibiotics for acute bronchitis in primary care, a recent study showed.

Researchers performed a three-arm cluster randomized trial of 33 primary care practices in an integrated health care system to determine whether provision of decision support would lead to more appropriate use of antibiotics for acute bronchitis. Eleven practices received printed brochures about decision support for acute cough illness, 11 practices received decision support via the electronic medical record, and 11 practices served as controls. At practices in the printed and electronic decision support groups, clinicians also received education and feedback about prescribing and patients received education brochures at check-in. Clinicians at the control sites were not aware of the study objectives.

The intervention was aimed at care of patients who were at least 13 years old and had an office visit for acute bronchitis during the baseline and intervention periods. The researchers used an intention-to-treat analysis to compare antibiotic prescription rates for uncomplicated acute bronchitis after the intervention (2009-2010) with rates in the same time period over the previous three years (2006-2007, 2007-2008 and 2008-2009). Only visits from Oct. 1 through March 31 were considered in each time period in order to capture the highest volume of visits for acute bronchitis. The study results were published online Jan. 14 by JAMA Internal Medicine.

Overall, 9,808 visits took place during the baseline periods and 6,242 took place during the intervention periods. When compared with the baseline periods, antibiotic use rates among adolescents and adults in the intervention period decreased in practices in the printed and electronic decision support groups (from 80.0% to 68.3% and from 74.0% to 60.7%, respectively) but increased in the control group (from 72.5% to 74.3%). The researchers controlled for patient and clinician characteristics and for clustering of observations by clinician and by practice site and found that the differences between each type of decision support group and the control group were statistically significant (P=0.003 for control vs. printed decision support and P=0.01 for control vs. electronic decision support). However, the effectiveness of the two types of decision support did not differ significantly (P=0.67).

The authors noted that their study took place in a health care delivery system that had been using a comprehensive electronic medical record for several years, and that all of the study sites were in rural or semi-rural areas. They also pointed out that they could not tell how each component of the multipart interventions (patient education, physician education, etc.) affected antibiotic prescribing rates. However, they concluded that a decision support strategy for acute bronchitis can lead to reduced rates of antibiotic use in primary care and that written support is as effective as computer-based support.

“Studies of computer-assisted decision support tools that do not include a comparison with more traditional implementation strategies may significantly overestimate the value of this type of decision support,” the authors concluded.

The author of an invited commentary pointed out that although the interventions were deemed effective, more than 60% of patients still received antibiotics inappropriately, indicating that more work needs to be done. He recommended that clinicians communicate the risks and benefits of antibiotic prescribing more clearly to patients; that continuous quality improvement techniques and measurement of results replace randomized, controlled trials; and that new interventions involving organizational change be investigated to try to change physician behavior. Last, he suggested that successful reduction of antibiotic prescribing in this context be redefined.

“Success is not reducing the antibiotic prescribing rate by 10%,” he wrote. “Success is reducing the antibiotic prescribing rate to 10%.”