Bayesian network modeling of patterns of antibiotic cross-resistance by bacterial sample source.

Antibiotics are drugs that kill some bacteria. Antibiotic resistant bacteria are bacteria that continue to grow despite the presence of an antibiotic drug. These bacteria are a major problem in healthcare, particularly if the bacteria are resistant to multiple drugs. Here, we study bacteria that are resistant to several antibiotics that are present in patients in hospital. We find that patterns of cross-resistance differ between the location bacteria were sampled from, such as blood or urine. Our results highlight the importance of considering sample sources when assessing the likelihood that bacteria is resistant to multiple antibiotics. The information and methods described in our study should enable further analysis and prediction of the presence of cross-resistant bacteria, enabling appropriate antibiotic treatments to be used.