Personalized antibiograms for machine learning driven antibiotic selection

doi:10.1038/s43856-022-00094-8

Personalized antibiograms for machine learning driven antibiotic selection

2022-04-08

34

OA

AI

DOI：10.1038/s43856-022-00094-8

原文链接

原文求助

摘要

En 中文

BackgroundThe Centers for Disease Control and Prevention identify antibiotic prescribing stewardship as the most important action to combat increasing antibiotic resistance. Clinicians balance broad empiric antibiotic coverage vs. precision coverage targeting only the most likely pathogens. We investigate the utility of machine learning-based clinical decision support for antibiotic prescribing stewardship.MethodsIn this retrospective multi-site study, we developed machine learning models that predict antibiotic susceptibility patterns (personalized antibiograms) using electronic health record data of 8342 infections from Stanford emergency departments and 15,806 uncomplicated urinary tract infections from Massachusetts General Hospital and Brigham & Women's Hospital in Boston. We assessed the trade-off between broad-spectrum and precise antibiotic prescribing using linear programming.ResultsWe find in Stanford data that personalized antibiograms reallocate clinician antibiotic selections with a coverage rate (fraction of infections covered by treatment) of 85.9%; similar to clinician performance (84.3% p = 0.11). In the Boston dataset, the personalized antibiograms coverage rate is 90.4%; a significant improvement over clinicians (88.1% p < 0.0001). Personalized antibiograms achieve similar coverage to the clinician benchmark with narrower antibiotics. With Stanford data, personalized antibiograms maintain clinician coverage rates while narrowing 69% of empiric vancomycin+piperacillin/tazobactam prescriptions to piperacillin/tazobactam. In the Boston dataset, personalized antibiograms maintain clinician coverage rates while narrowing 48% of ciprofloxacin to trimethoprim/sulfamethoxazole.ConclusionsPrecision empiric antibiotic prescribing with personalized antibiograms could improve patient safety and antibiotic stewardship by reducing unnecessary use of broad-spectrum antibiotics that breed a growing tide of resistant organisms. Plain language summaryAntibiotic resistance is an increasing threat to public health. The World Health Organization estimates that 700,000 people die annually due to antibiotic resistant infection. By 2050 the annual death toll is expected to reach 10 million. The Centers for Disease Control and Prevention list the importance of appropriate prescribing of antibiotics as the number one action advised to reduce the spread of resistant bacteria. When selecting appropriate antibiotics, clinicians aim to maximize the likelihood that individual patients will respond whilst limiting the use of options that have action against a large number of different bacteria. Overuse of valuable wide acting antibiotics can increase the rate at which bacteria develop resistance to them. Here we show that machine learning models that predict antibiotic susceptibility have the potential to guide clinicians when choosing antibiotics in a way that maintains or improves patient safety while reducing the overall use of wide acting antibiotics. Corbin et al. train machine learning models on electronic health record data to predict susceptibility of infections to particular antibiotics (personalized antibiograms). Antibiotic selection driven by personalized antibiograms achieves similar coverage rates to those seen in actual clinical practice using fewer broad spectrum antibiotics.

Keyword:

BROAD-SPECTRUM ANTIBIOTICS

PROPHYLAXIS

STEWARDSHIP

INFECTIONS

RISK

VANCOMYCIN

GUIDELINES

RESISTANCE

OUTCOMES

Personalized antibiograms for machine learning driven antibiotic selection

摘要

期刊

机构

引用论文