Article

Electronic Tool Developed to Help Clinicians Detect Antibiotic Allergic-Type Reactions

Author(s):

New research has led to the development of an algorithm designed to help improve detection of allergic reactions to antibiotics, given that antibiotics are among the most common medication adverse event causes.

Westyn Branch-Elliman, MD, MMSc

Credit: LinkedIn

Westyn Branch-Elliman, MD, MMSc

Credit: LinkedIn

A high-sensitivity algorithm has been developed that provides clinicians’ feedback on the harms of unnecessarily prolonged exposures to antibiotics, according to new research on patients with periprocedural antibiotic prophylaxis.1

Antibiotics are some of the most frequently-used drugs associated with patients in hospitals reporting hypersensitivity reactions, which may lead to higher health costs and even death.2

The treatments which are prescribed by clinicians for infection prophylaxis at surgical sites are most commonly the antibiotics which can cause hypersensitivity reactions.

This study, led by Westyn Branch-Elliman, MD, MMSc—from the Center for Organization and Implementation Research at VA Boston Healthcare System in Massachusetts—was conducted with the aim of designing an informatics tool for improving antibiotic allergic-type event detection.

Branch-Elliman and colleagues wrote that their goal was to “develop an electronic tool using structured and unstructured data elements available within the Veterans Affairs (VA) electronic health record (EHR) to identify incident antibiotic allergic-type reactions and (2) to characterize the nature and severity of these reactions.”

Background and Findings

The investigators conducted—from October of 2015, to September of 2019—a retrospective cohort study that they carried out in VA hospitals which focused on patients who underwent cardiovascular electronic device implant procedures and also received periprocedural antibiotic prophylaxis.

The research team’s analysis of the data took place between July of 2021 and January of 2022. The team divided the cohort into training and test cohorts and manually reviewed the cases included to determine the presence and severity of reactions considered allergic-type.

They identified possible variables that may indicate allergic-type reactions in advance. These variables they assessed included the following:
Allergies recorded in the VA's Allergy Reaction Tracking system
Medications given for treating allergic reactions
Allergy diagnosis codes
Text searches of clinical notes with specific keywords or phrases associated with potential allergic-type reactions

The investigators developed a model on their training cohort to then detect reaction events that were allergic-type and then applied it to the test cohort. The performance of the algorithm was later evaluated by the team’s assessment of its test characteristics.

In total, the investigators used data from a cohort of 36,344 patients, primarily males with a mean age of 72 years, who had cardiovascular implantable electronic device procedures and also had antibiotic prophylaxis.

The research team added that the median duration of postprocedural antibiotic use was found to be 4 days. The team was able to develope an algorithm using 7 variables, which were as follows:

  • Historic and observed entries in the Veteran Affair's hospitals ART
  • PheCodes related to skin symptoms
  • Urticaria
  • Antibiotic allergies
  • Adverse events
  • Keyword detection in clinical notes
  • Antihistamine administration

The final algorithm was found to have successfully identified antibiotic allergic-type reactions with a probability of 30% or more. The investigators noted that the positive predictive value of the algorithm was shown to be 61%, indicating that 61% of the identified cases were truly antibiotic allergic reactions.

The team added that the sensitivity of their algorithm, which assesses its ability to correctly detect positive cases, was found to be 87%. This suggests that the algorithm performed well in identifying true positive cases of antibiotic allergic-type reactions.

“This model had a good PPV for detecting antibiotic allergic-type reactions and can be operationalized to provide clinicians with real-time audit and feedback about when allergic-type reactions occur to support efforts to reduce inappropriate and unnecessary antibiotic exposures that increase patient-level harm without improving outcomes,” they wrote.

References

  1. Reyes Dassum S, Mull HJ, Golenbock S, et al. A Novel Informatics Tool to Detect Periprocedural Antibiotic Allergy Adverse Events for Near Real-time Surveillance to Support Audit and Feedback. JAMA Netw Open. 2023;6(5):e2313964. doi:10.1001/jamanetworkopen.2023.13964.
  2. Thong BYH, Tan TC. Epidemiology and risk factors for drug allergy. Br J Clin Pharmacol. 2011;71(5):684-700. doi:10.1111/j.1365-2125.2010.03774.x.
Related Videos
Jonathan Meyer, MD: Cognitive Gains, Dopamine-Free Schizophrenia Treatment with Xanomeline Trospium Chloride
Allysa Saggese, NP | Credit: Weill Cornell Medicine
Zobair Younossi, MD, MPH | Credit: American College of Gastroenterology
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
Thumbnail for schizophrenia special report around approval of Cobenfy.
© 2024 MJH Life Sciences

All rights reserved.