An audit and feedback intervention study increased adherence to antibiotic prescribing guidelines at a Norwegian hospital

June Utnes Høgli, Beate Hennie Garcia, Frode Skjold, Vegard Skogen, Lars Småbrekke

Appropriate antibiotic prescribing is associated with favourable levels of antimicrobial resistance (AMR) and clinical outcomes. Most intervention studies on antibiotic prescribing originate from settings with high level of AMR. In a Norwegian hospital setting with low level of AMR, the literature on interventions for promoting guideline-recommended antibiotic prescribing in hospital is scarce and requested. Preliminary studies have shown improvement potentials regarding antibiotic prescribing according to guidelines. We aimed to promote appropriate antibiotic prescribing in patients with community-acquired pneumonia (CAP) and acute exacerbations of chronic obstructive pulmonary disease (AECOPD) at a respiratory medicine department in a Norwegian University hospital. Our specific objectives were to increase prescribing of appropriate empirical antibiotics, reduce high-dose benzylpenicillin and reduce total treatment duration. We performed an audit and feedback intervention study, combined with distribution of a recently published pocket version of the national clinical practice guideline. We included patients discharged with CAP or AECOPD and prescribed antibiotics during hospital stay, and excluded those presenting with aspiration, nosocomial infection and co-infections. The pre- and post-intervention period was 9 and 6 months, respectively. Feedback was provided orally to the department physicians at an internal-educational meeting. To explore the effect of the intervention on appropriate empirical antibiotics and mean total treatment duration we applied before-after analysis (Student's t-test) and interrupted time series (ITS). We used Pearson's χ2 to compare dose changes. In the pre-and post-intervention period we included 253 and 155 patients, respectively. Following the intervention, overall mean prescribing of appropriate empirical antibiotics increased from 61.7 to 83.8 % (P < 0.001), overall mean total treatment duration decreased from 11.2 to 10.4 days (P = 0.015), and prescribing of high-dose benzylpenicillin decreased from 48.8 to 38.6 % (P = 0.125). With ITS we found that six months post-intervention, the effect on appropriate empirical antibiotic prescribing had increased and sustained, while the effect on treatment duration was at pre-intervention level. The combination of audit and feedback plus distribution of a pocket version of guideline recommendations led to a substantial increase in prescribing of appropriate empirical antibiotics, which is important due to favourable effect on AMR and clinical outcomes.