Quality of inpatient antimicrobial use in hematology and oncology patients

To compare antimicrobial prescribing practices in Australian hematology and oncology patients to noncancer acute inpatients and to identify targets for stewardship interventions.

Retrospective comparative analysis of a national prospectively collected database.

Using data from the 2014-2018 annual Australian point-prevalence surveys of antimicrobial prescribing in hospitalized patients (ie, Hospital National Antimicrobial Prescribing Survey called Hospital NAPS), the most frequently used antimicrobials, their appropriateness, and guideline concordance were compared among hematology/bone marrow transplant (hemBMT), oncology, and noncancer inpatients in the setting of treatment of neutropenic fever and antibacterial and antifungal prophylaxis.

In 454 facilities, 94,226 antibiotic prescriptions for 62,607 adult inpatients (2,230 hemBMT, 1,824 oncology, and 58,553 noncancer) were analyzed. Appropriateness was high for neutropenic fever management across groups (83.4%-90.4%); however, hemBMT patients had high rates of carbapenem use (111 of 746 prescriptions, 14.9%), and 20.2% of these prescriptions were deemed inappropriate. Logistic regression demonstrated that hemBMT patients were more likely to receive appropriate antifungal prophylaxis compared to oncology and noncancer patients (adjusted OR, 5.3; P < .001 for hemBMT compared to noncancer patients). Oncology had a low rate of antifungal prophylaxis guideline compliance (67.2%), and incorrect dosage and frequency were key factors. Compared to oncology patients, hemBMT patients were more likely to receive appropriate nonsurgical antibacterial prophylaxis (aOR, 8.4; 95% CI, 5.3-13.3; P < .001). HemBMT patients were also more likely to receive appropriate nonsurgical antibacterial prophylaxis compared to noncancer patients (OR, 3.1; 95% CI, 1.9-5.0; P < .001). However, in the Australian context, the hemBMT group had higher than expected use of fluoroquinolone prophylaxis (66 of 831 prescriptions, 8%).

This study demonstrates why separate analysis of hemBMT and oncology populations is necessary to identify specific opportunities for quality improvement in each patient group.