Ten studies were included: seven RCTs, one non-randomised controlled trial and two controlled before and after studies. Where reported, the number of health care professionals ranged from six to 175 and the number of children ranged from 324 to 13,460. The three non-randomised studies were judged at high risk of bias for sequence generation and allocation concealment and low risk of bias for all other items. Three of the RCTs were judged to be at low risk of bias for sequence generation and two were judged at low risk of bias for allocation concealment; the other RCTs were judged as unclear risk of bias for these items. All studies except one were judged at low risk of bias for blinding and avoidance of selective outcome reporting; all were at low risk of bias for incomplete outcome data.
Computer interventions (three RCTs): All three studies reported that antibiotic behaviour was improved following computer interventions. One study reported that antibiotic prescription rates were reduced by 34% (95% CI 14% to 54%) and that the intervention increased compliance with the recommended management plan (RD 41%, 95% CI 29% to 53%). One study reported a significant decrease in duration of prescribed antibiotic courses in the intervention compared to the control arm (RD 34%, 95% CI 29% to 39%) but found no significant difference in terms of the proportion of antibiotics prescribed. The third study found a significant improvement in antibiotic treatment (RD 15%, 95% CI 2% to 13%) in the intervention group, but no significant difference in amoxicillin use or duration of antibiotic prescription.
Educational Sessions (two RCTs, one non-randomised study, one controlled before-and-after study): One study found that an educational intervention increased adherence to recommendations (RD 8%, 95% CI 4% to 11%) and another found that the educational intervention improved knowledge of compliance enhancing strategies (RD 28%, 95% CI 4% to 51% and RD 29%, 95% CI 5% to 53%) but showed little evidence of improvement in self report of compliance. Two studies reported improvement in antibiotic prescribing rates in the educational intervention group compared to control (RD 2%, 95% CI 0 to 5% and RD 17%, 95% CI 10% to 25%).
Other interventions (two RCTs, one controlled before-and-after study): A combination of a training video and risk factor checklist increased appropriate referral by the general practitioner to the otolaryngologist (RD 37%, 95% CI 25% to 49%). A collaborative protocol development combined with educational materials reduced antibiotic prescription rate (OR 0.60, 95% CI 0.43 to 0.83). Collaborative protocol development alone also increased the proportion of prescriptions that followed the protocol (p<0.001). Discussion and providing a protocol decreased prescription rates of courses of multiple antibiotics (RD 22%, 95% CI 10% to 35%).