Influence of a guideline or an additional rapid strep test on antibiotic prescriptions for sore throat: the cluster randomized controlled trial of HALS (Hals und Antibiotika Leitlinien Strategien)

Abstract

Background: Pharyngitis due to Group A beta-hemolytic streptococci (GAS) is seen as the main indication for antibiotics for sore throat. In primary care settings prescription rates are much higher than the prevalence of GAS. Recommendations in international guidelines differ considerably. A German guideline suggested to consider antibiotics for patients with Centor or McIsaac scores ≥ 3, first choice being penicillin V for 7 days, and recommended analgesics for all. We investigated, if the implementation of this guideline lowers the antibiotic prescription rate, and if a rapid antigen detection strep-test (RADT) in patients with scores ≥ 3 lowers the rate further. Methods: HALS was an open pragmatic parallel group three-arm cluster-randomized controlled trial. Primary care practices in Northern Germany were randomized into three groups: Guideline (GL-group), modified guideline with a RADT for scores ≥ 3 (GL-RADT-group) or usual care (UC-group). All practices were visited and instructed by the study team (outreach visits) and supplied with material according to their group. The practices were asked to recruit 11 consecutive patients ≥ 2 years with an acute sore throat and being at least moderately impaired. A study throat swab for GAS was taken in every patient. The antibiotic prescription rate at the first consultation was the primary outcome. Results: From October 2010 to March 2012, 68 general practitioners in 61 practices recruited 520 patients, 516 could be analyzed for the primary endpoint. Antibiotic prescription rates did not differ between groups (p = 0.162) and were about three times higher than the GAS rate: GL-group 97/187 patients (52%; GAS = 16%), GL-RADT-group 74/172 (43%; GAS = 16%) and UC-group 68/157 (43%; GAS = 14%). In the GL-RADT-group 55% of patients had scores ≥ 3 compared to 35% in GL-group (p < 0.001). After adjustment, in the GL-RADT-group the OR was 0.23 for getting an antibiotic compared to the GL-group (p = 0.010), even though 35 of 90 patients with a negative Strep-test got an antibiotic in the GL-RADT-group. The prescription rates per practice covered the full range from 0 to 100% in all groups. Conclusion: The scores proposed in the implemented guideline seem inappropriate to lower antibiotic prescriptions for sore throat, but better adherence of practitioners to negative RADTs should lead to fewer prescriptions. Trial registration: DRKS00013018, retrospectively registered 28.11.2017.