Variability in antimicrobial prescribing among hospital-based physicians is not associated with patient characteristics or clinical outcomes, data suggest. The lowest level of such prescribing within each hospital could be considered a target for antimicrobial stewardship, according to the researchers.
In a multicenter study of 124 physicians responsible for more than 124,000 hospitalized patients, the difference in mean prescribing between the highest and lowest quartiles of prescription volume was 15.8 days of treatment per 100 patient-days.
Baseline patient characteristics were similar across the quartiles, and there were no differences in patient outcomes, including in-hospital deaths, hospital length of stay, intensive care unit transfer, and hospital readmission.
Although the investigators expected variation in prescribing, “what surprised us most was the limited association with any differences in clinical outcomes, particularly when it came to the amount of antimicrobials used,” study author Mark T. McIntyre, PharmD, pharmacotherapy specialist at the Sinai Health System in Toronto, told this news organization.
“Importantly, this is not a study that defines quality of care,” he said. “We looked at natural variation in practice and association with outcomes. So, I don’t want clinicians to think, ‘Well, I’m high, therefore I’m bad,’ or, ‘I’m low, therefore I’m good.’
“This is an early explanatory analysis that asks whether this is an opportunity to optimize prescribing in ways we hadn’t thought of before,” he said. “Now that we don’t have an association with higher or lower prescribing and outcomes, we can look at what else is driving that antimicrobial prescribing and what we can do about it. Comfort level, risk tolerance, and social, cultural, and contextual factors all likely play a role.”
The study was published online in the Canadian Medical Association Journal.
Antimicrobial reductions possible
The investigators conducted a retrospective cohort study using the General Medicine Inpatient Initiative database to assess physician-level volume and spectrum of antimicrobial prescribing in adult general medical wards. Four academic hospitals in Toronto were evaluated for the period 2010 to 2019.
The investigators stratified physicians into quartiles by hospital site on the basis of volume of antimicrobial prescribing (specifically, days of therapy per 100 patient-days and antimicrobial-free days) and antibacterial spectrum (modified spectrum score, which assigns a value to each antibacterial agent on the basis of its breadth of coverage).
They also examined potential differences between physician quartiles in patient characteristics, such as age, sex, the Laboratory-Based Acute Physiology Score, discharge diagnosis, and the Charlson Comorbidity Index.
Multilevel modeling allowed the investigators to evaluate the association between clinical outcomes and antimicrobial volume and spectrum.
The primary measure was days of therapy per 100 patient-days.
As noted, the cohort included 124 physicians who were responsible for 124,158 hospital admissions. The median physician-level volume of antimicrobial prescribing was 56.1 days of therapy per 100 patient-days. Patient characteristics were balanced across the quartiles of physician prescribing.
The difference in mean prescribing between physician quartile 4 and quartile 1 was 15.8 days of therapy per 100 patient-days, meaning the median physician in quartile 4 prescribed antimicrobials at a volume that was 30% higher than that of the median physician in quartile 1.
No significant differences were noted for any clinical outcome with regard to quartile of days of therapy, antimicrobial-free days, or modified spectrum score after adjustment for patient-level characteristics.
In addition, no significant differences in the case mix between quartile 4 and quartile 1 were found when the cohort was restricted to patients admitted and discharged by the same most responsible person, nor were differences found in an analysis that was restricted to those without a discharge diagnosis code of palliative care.
In-hospital mortality was higher among patients cared for by prescribers with higher modified spectrum scores (odds ratio, 1.13). “We still can’t fully explain this finding,” Dr. McIntyre acknowledged. “We only saw that in our primary analysis. When we did several sensitivity analyses, that finding didn’t appear.”
The authors concluded, “Ultimately, without discernible benefit in outcomes of patients of physicians who prescribe more frequently, less antimicrobial exposure may be possible, leading to lower risk of antimicrobial resistance.”