notesum.ai

Published at November 26

Explainable AI for Classifying UTI Risk Groups Using a Real-World Linked EHR and Pathology Lab Dataset

cs.LG

cs.AI

Released Date: November 26, 2024

Authors: Yujie Dai¹, Brian Sullivan¹, Axel Montout¹, Amy Dillon¹, Chris Waller², Peter Acs², Rachel Denholm¹, Philip Williams³, Alastair D Hay¹, Raul Santos-Rodriguez¹, Andrew Dowsey¹

Aff.: ¹University of Bristol; ²NHS England; ³University Hospitals Bristol and Weston NHS Foundation Trust

Arxiv: http://arxiv.org/abs/2411.17645v1

Demographic Information	Group
	UTI Group¹¹1Individuals with a UTI likelihood $>$ 0	Control Group²²2Individuals with a UTI likelihood $=$ 0
	N=147518	N=610602
Age (years)
18-24	9445 (6.4%)	73137 (12.0%)
25-44	49822 (33.8%)	259189 (42.4%)
45-64	37688 (25.5%)	179170 (29.3%)
65-84	39151 (26.5%)	87452 (14.3%)
85+	11412 (7.7%)	11654 (1.9%)
Gender
Male	45082 (30.6%)	340534 (55.8%)
Female	102436 (69.4%)	270068 (44.2%)
Comorbidities
Incontinent Urinary	538 (0.4%)	405 (0.1%)
Dementia	4693 (3.2%)	3951 (0.6%)
Covid High Risk	22433 (15.2%)	23130 (3.8%)
Covid Increased Risk	84644 (57.4%)	185056 (30.3%)
Organ Transplant	172 (0.1%)	222 (0.0%)
Living Conditions
Housebound	6750 (4.6%)	4031 (0.7%)
Nursing/Caring Home	3304 (2.2%)	2651 (0.4%)
Homeless	159 (0.1%)	502 (0.1%)