Keith W. Hamilton
Researchers have developed an antibiogram that uses hospital data as well as patient-specific data to predict the likelihood that a treatment will be successful.
During a presentation at The Society for Healthcare Epidemiology of America spring conference, Keith W. Hamilton, MD, associate health care epidemiologist and director of antimicrobial stewardship at the Hospital of the University of Pennsylvania, and colleagues showed the patient-specific antibiogram (PS-ABG) was more accurate in predicting antibiotic susceptibility than a traditional antibiogram (ABG) that relies solely on hospital statistics.
“The goal of antibiotic stewardship is to make sure the right patient gets the right antibiotic at the right time,” Hamilton told Infectious Disease News. “We are very good at directing empiric antibiotics to patients who fall into the bell curve. There are other patients who may be at greater risk for resistant infections. Using risk factors that we know are associated with antibiotic resistance is important to guide what antibiotics go to which patients.”
Known risk factors of antibiotic resistance include prior antibiotic exposure, a history of infection or colonization with drug-resistant organisms, location in a hospital (ICU vs. non-ICU setting), and certain diagnoses, according to Hamilton.
Through a partnership with the health IT company ILÚM Health Solutions, Hamilton and colleagues developed a software that uses historic patient data to inform antibiotic prescribing practices in real time. With this software, the research team created a PS-ABG that incorporates clinical characteristics into an institutional ABG.
The researchers compared the accuracy of PS-ABG and ABG using data from more than 7,500 patients at the Hospital of the University of Pennsylvania. Looking primarily at organisms that cause urinary tract infections, Hamilton and colleagues found that the PS-ABG was as precise or, in many cases, statistically superior to the traditional ABG in establishing antibiotic susceptibility.
Moving forward, the researchers said they will explore whether other clinical risk factors of antibiotic resistance can be applied to the model. Future research will also assess the clinical implications of PS-ABG.
“Whether using a sophisticated model like this or risk scores to identify patients who are at higher risk for resistant infections, using more precision in prescribing practices will help to improve patient outcomes,” Hamilton said. “It is important to take patient risk factors into consideration to help clinicians determine what antibiotics they should prescribe.” – by Stephanie Viguers
Hamilton KW, et al. Abstract 10157. Presented at: Society for Healthcare Epidemiology of America (SHEA) spring conference; April 18-20, 2018; Portland, Ore.
Disclosure: Hamilton reports no relevant financial disclosures.