Multifactor machine learning models for predicting urinary tract infections: a pilot study

Comprehensive Summary

Grizzi et al. present a study detailing the use of 12 different machine learning models in analyzing a four-predictor panel to predict the occurrence of urinary tract infections (UTIs) in patients. Urinary vitamin D levels were tested for each of the 358 patients selected for the study and their age, gender, and urinary pH was also taken into account. The following 12 machine learning models were used: Random Forest Classifier (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Logistic Regression (LR), Decision Tree (DT), Gaussian Naïve Bayes (GNB), K-Nearest Neighbors (KNN), Gradient Boosting Machine (GBM), Adaptive Boosting (AdaBoost), Light Gradient Boosting Machine (LightGBM), Categorical Boosting (CatBoost), and Deep Neural Network (DNN), and evaluated based on accuracy, specificity and other preset parameters. Statistical analysis was then done of the results of these models and quantified into 95% confidence intervals. The top three machine learning models based on the above parameters were RF, XGBoost, and GBM. The findings show that using machine learning models can reduce the work needed to be done to help predict what patients may be at greater risk of UTIs.

Outcomes and Implications

Urinary tract infections are incredibly prevalent in both men and women and can lead to extreme complications if allowed to progress. The implications for the medical community of this study is that it can be significantly easier to predict and diagnose patients with UTIs. This allows for further personalization of care and keeping patients as healthy as possible. It can also help in figuring out what antibiotics work and help identify early patterns of antibiotic resistance.