AI reduces diagnostic workload of detecting UTIs

Urinary tract infections make up a significant portion of microbiological screening in diagnostic laboratories, yet nearly two-thirds of samples come up negative. But AI has the potential to improve the process by reducing the number of query samples and enabling diagnostic services to concentrate on those that many have actual infections, according to findings published in BMC Medical Informatics and Decision Making.

Researchers from Cardiff University in Bristol, U.K., conducted a retrospective analysis of all urine microscopy, culture and sensitivity reports over one year against two methods of classification––a heuristic model, which used a combination of white blood cell count and bacterial count, and a machine learning approach with three algorithms. Overall, more than 212,000 urine reports were analyzed.

“Taking advantage of recent developments in ‘big data’ technologies, our observational study [analyzed] data representing an entire year of urine analysis at a large pathology service that covers sample processing for multiple hospitals as well as the community in the Bristol/Bath region in the Southwest of the U.K.,” wrote first author Ross Burton, a PhD student at the School of Medicine at Cardiff University, et al.

The machine learning algorithms outperformed the heuristic model, with a classification sensitivity of 95% while achieving a relative workload reduction.

However, further analysis found that samples from pregnant patients and children required further independent evaluation. The three algorithms had to be trained independently to classify pregnant patients, children, then other patients. With this combined system, the relative workload reduction was 41% and sensitivity was 95% for each patient group.

“The work presented here shows that supervised machine learning models can be of significant utility in predicting whether urine samples are likely to require bacterial culture,” Burton et al. concluded. “This could potentially improve service efficiency at a time when demand is surpassing the resources of public healthcare providers."