Comparative Analysis of Machine Learning Models For Predicting Hospital- and Community-associated Urinary Tract Infections Using Demographic, Hospital, and Socioeconomic Predictors.

Details

• Alternative Title:
  The Journal of Hospital Infection, 2025, v. 162: Comparative Analysis of Machine Learning Models For Predicting Hospital- and Community-associated Urinary Tract Infections Using Demographic, Hospital, and Socioeconomic Predictors.
• Journal Article:
  The Journal of Hospital Infection
• Personal Author:
  Arjmand, Arash ; Bani-Yaghoub, Majid ; Sutkin, Gary ; Corkran, Kiel ; Paschal, Susanna
• Description:
  Differentiating between community-associated (CA-) and healthcare-associated (HA-) urinary tract infection (UTI) is crucial for understanding their epidemiology, identifying risk factors, and developing appropriate treatment strategies.
  
  To build, validate, and compare machine learning models: Decision Tree, Neural Network, Logistic Regression, Random Forest, and Extreme Gradient Boosting to differentiate between the incidences of HA-UTI and CA-UTI; additionally, to identify key predictors of UTI using demographic, hospital, and socioeconomic variables.
  
  Patient demographics, hospital, and socioeconomic data from 2019 to 2023 were analysed.
  
  The Decision Tree model demonstrated the highest sensitivity, particularly in handling the highly imbalanced data of HAI, with a sensitivity of 87%. Logistic Regression achieved the best overall accuracy, at 95.9% for distinguishing HA-UTI from UTI-free and 93.2% for HA-UTI vs CA-UTI. Random Forest performed best in cross-validation, reaching 99.1% for HA-UTI vs UTI free and 96.2% for HA-UTI vs CA-UTI. Neural Network showed the highest specificity, at 93.4%, for HA-UTI vs CA-UTI. The area-under-the-curve values further supported these findings, ranging from 71.9% for Neural Network to 94% for Random Forest, reflecting the robustness of these models across different annual datasets. Among patient demographics, hospital, and socioeconomic variables, all models consistently identified the Nurse Units (e.g. Inpatient Units and Mental Health Units) as the most significant predictors of UTI. In addition to Nurse Units, Logistic Regression and Decision Tree identified location (e.g. various clinics and medical centres) as a key predictor.
  
  The machine learning models demonstrated comparable overall accuracy, but differed in sensitivity and specificity across the two classification tasks-HA-UTI vs CA-UTI and HA-UTI vs UTI-free. Nurse Units consistently emerge as the most significant predictors across all years. The importance of all predictors varies from year to year.
• Content Notes:
  Author manuscript
• Subjects:
  • Humans
  • Machine Learning
  • Urinary Tract Infections
  • Community-Acquired Infections
  • Male
  • Female
  • Middle Aged
  • Cross Infection
  • Aged
  • Hospitals
  More +
• Keywords:
  • Urinary tract infection
  • community-associated infection
  • healthcare-associated infection
  • machine learning
  • socioeconomic factors
• Source:
  The Journal of Hospital Infection, 2025, v. 162
• DOI:
  https://doi.org/10.1016/j.jhin.2025.04.024
• ISSN:
  0195-6701 ; 1532-2939
• Pubmed ID:
  40339918
• Pubmed Central ID:
  PMC12875383
• Document Type:
  Text
• Funding:
  U01 CK000671/CK/NCEZID CDC HHSUnited States/ ; U01CK000671/ACL/ACL HHSUnited States/
• Genre:
  Journal Article
• Pages in Document:
  263-271 (19 pdf pages)
• Volume:
  162
• Download URL:
  https://stacks.cdc.gov/view/cdc/261427/cdc_261427_DS1.pdf
• File Type:
  [PDF - 297.03 KB]
• Collection(s):
  CDC Public Access
• Main Document Checksum:
  urn:sha-512:68ecbba84e48330617a7b2eda616f21687fb70ae012bfeddce60541ce77942e42ac7830f92fddcb8001e003aecca3466e25ceab59cf949ae1d957b516441eaa5