Computational Decision Support For Penicillin Allergy Risk Stratification

Background:

Antibiotic allergy constitutes a major health problem with serious medical and financial consequences when under- but also overdiagnosed.

About one and a half million Belgians report having an allergy to penicillin. However, in 1.35 million (90%) of these "penicillin allergies", the suspicion is wrong. The issue of false penicillin allergy is therefore seen as an important medical problem with significant implications for the individual patient but also for society.

Today, therefore, there is a global consensus to contain the pandemic of false penicillin allergy as much as possible. This should ultimately lead to correct antibiotic policies for the patient and reduce antibiotic resistance and the cost to society.

The problem with this is that there is no diagnostic test that one can perform on a population of one and a half million people. This is practically and financially unfeasible. So there is a need for proper risk stratification based on anamnestic data to better guide our diagnostics.

Specific question:

The investigators currently have a database of about 1000 patients with a possible hypersensitivity to antibiotics who have been fully diagnosed according to current guidelines. The aim of the study is to use this database to see whether a tool can be developed to reliably perform an initial "screening" to determine in whom further testing is necessary and in whom, based purely on the story, it can be decided that the likelihood of penicillin allergy is extremely low and additional diagnostics are unnecessary.