Key Point

Clinical decision support systems help physicians to prescribe antimicrobials less frequently and more appropriately, although the improvements appear to be modest. Antibiotics are overprescribed for children with sore throat.

SALT LAKE CITY AND BOSTON—The frequency of antimicrobial prescriptions has fallen as a result of efforts to educate physicians and the public about the dangers of unnecessary antibiotic use. "But there is still room for improvement," asserted J. Todd Weber, MD, in an editorial.¹

Dr. Weber, Director of the CDC’s Office of Antimicrobial Resistance, was commenting on two recent studies of antimicrobial prescribing habits. One of the studies, by Jeffrey A. Linder, MD, MPH, and colleagues, found that antibiotics are still inappropriately prescribed for children with sore throat.² Antibiotic use in these cases far exceeded reported rates of pediatric group A ß-hemolytic streptococci (GABHS) infection, the only common cause of sore throat in children that warrants antibiotics, reported Dr. Linder, Instructor in Medicine at Harvard University Medical School in Boston.

In addition, while an overall reduction in antibiotic prescribing was observed over time, it was because of the decreased use of agents recommended for GABHS. "The use of nonrecommended antimicrobials (eg, amoxicillin/clavulanate, clarithromycin, and azithromycin) remained stable over the nine-year study period but at an unacceptably high level of 27% among those who received antimicrobials," pointed out Dr. Weber.

The other study, by Samore et al, had a positive aspect. It showed that more appropriate antimicrobial use is possible with either of two kinds of clinical decision support systems—a paper document or a handheld personal digital assistant—designed to help primary care physicians manage respiratory tract infections.³

However, the clinical decision support systems were only modestly effective for a narrow range of diagnoses within small communities, according to Dr. Weber. Furthermore, the educational effort surrounding their use required a significant amount of time and resources.

ANTIBIOTICS FOR CHILDREN WITH SORE THROAT

Dr. Linder’s group analyzed data from office, emergency department, and hospital outpatient department visits for sore throat between 1995 and 2003. The study included 4,158 children ages 3 to 17 years. Data were obtained from two CDC surveys—the National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey.

"Slightly more than half of the children in the study received antibiotics even though GABHS is only found in 15% to 36% of children with sore throat," Dr. Linder told Pulmonary Reviews. An antibiotic that is not recommended for GABHS was prescribed for 27% of children who received an antibiotic, he noted.

Overall, the antibiotic prescription rate fell from 66% in 1995 to 54% in 2003. However, the apparent improvement was attributed to a decline in the prescribing of antibiotics recommended for GABHS, from 49% at the beginning of the study period to 38% at the end. Rates of nonrecommended antibiotic prescribing were essentially the same at those times (18% in 1995 and 16% in 2003).

The overall rate of GABHS testing with the rapid test or a culture was 53%. Physicians prescribed antibiotic therapy in 48% of visits that included a GABHS test and in 51% of visits in which a GABHS test was not performed.

"We were disappointed with our findings," Dr. Linder remarked. "Physicians prescribe antibiotics too often for children with sore throat, even though this is one place where we should be getting it right nearly 100% of the time."

The study also clearly illustrated that physicians do not test for GABHS as frequently as they should, and too often, they choose inappropriate GABHS treatments, said Dr. Linder. "Unless a child is allergic, there is no reason to give anything other than penicillin or amoxicillin for GABHS infection," he explained. "GABHS is never resistant to penicillin, but it is to many other antimicrobials."

TWO YEARS OF CLINICAL DECISION SUPPORT SYSTEMS

In the study by Samore and colleagues, 12 small rural communities were randomized to one of two strategies to facilitate appropriate antimicrobial prescribing in cases of respiratory tract infection—a community intervention alone or with clinical decision support systems. The community intervention included the promotion and distribution of a printed guide discussing patient self-management of common respiratory tract infections and better communication with clinicians.

Of the primary care physicians in areas randomized to the community intervention plus clinical decision support systems, 18% chose the paper version of the clinical decision support systems: a flowchart that guided the user to the correct diagnosis and treatment of respiratory tract infections. Fifty-four percent selected the personal digital assistant, which generated diagnostic and therapeutic recommendations based on patient data entered into the device. Twenty-four percent used both types of clinical decision support systems. Notably, only 5% relied on a paper form that patients used to initiate care for suspected respiratory tract infection.

Antimicrobial prescribing rates were analyzed the year before the study (2001) and during the two subsequent years (2002 to 2003). By the end of the study, the prescribing rate decreased from 84.1 to 75.3 per 100 person-years in the clinical decision support systems arm and rose from 84.3 to 85.2 per 100 person-years in areas that received only the community intervention. The rate was unchanged in a reference group of six rural communities.

The decline in antimicrobial prescriptions in the clinical decision support systems arm occurred entirely in the second year of the study; no significant decreases were observed in the first year. In year 2, "prescribing rates in clinical decision support systems communities decreased 10% from baseline whereas in the community intervention–alone communities and [reference] communities, prescribing rates in 2003 increased by 1% and 6%, respectively," related the authors.

The clinical decision support systems–related change in antimicrobial use appeared to be due mainly to reductions in macrolide prescriptions of 12% and 28% in study years 1 and 2, respectively. Clinical decision support systems were also associated with a 32% reduction in the likelihood of antimicrobials being prescribed when they were never indicated.

—Timothy Begany

References
1. Weber JT. Appropriate use of antimicrobial drugs: a better prescription is needed. JAMA. 2005;294:2354-2356.
2. Linder JA, Bates DW, Lee GM, Finkelstein JA. Antibiotic treatment of children with sore throat. JAMA. 2005; 294:2315-2322.
3. Samore MH, Bateman K, Alder SC, et al. Clinical decision support and appropriateness of antimicrobial prescribing: a randomized trial. JAMA. 2005;294:2305-2314.

Return to table of contents