Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Urinary Tract Infection Detection in Digital Remote Monitoring: Strategies for Managing Participant-Specific Prediction Complexity

Kexin Fan, Alexander Capstick, Ramin Nilforooshan, Payam Barnaghi

TL;DR

The paper tackles early UTI detection in people living with dementia using in-home passive sensors (Minder dataset) and introduces multitask-inspired MLP designs to address participant variability and sex fairness. It evaluates feature clustering, loss-dependent clustering, and participant ID embeddings, finding that loss-dependent approaches—especially Final Layer Separated—substantially improve validation metrics (e.g., precision from 48.92% to 72.60% and sensitivity from 27.44% to 70.52%), with better gender parity. The results highlight the potential for more reliable, equitable screening in real-world home settings, while revealing challenges in generalizing to new participants and leveraging ID embeddings due to limited per-person data. These insights support integrating refined models into the Minder platform to enable earlier, more accurate UTI management in PLWD, guiding clinical decision-making and resource allocation.

Abstract

Urinary tract infections (UTIs) are a significant health concern, particularly for people living with dementia (PLWD), as they can lead to severe complications if not detected and treated early. This study builds on previous work that utilised machine learning (ML) to detect UTIs in PLWD by analysing in-home activity and physiological data collected through low-cost, passive sensors. The current research focuses on improving the performance of previous models, particularly by refining the Multilayer Perceptron (MLP), to better handle variations in home environments and improve sex fairness in predictions by making use of concepts from multitask learning. This study implemented three primary model designs: feature clustering, loss-dependent clustering, and participant ID embedding which were compared against a baseline MLP model. The results demonstrated that the loss-dependent MLP achieved the most significant improvements, increasing validation precision from 48.92% to 72.60% and sensitivity from 27.44% to 70.52%, while also enhancing model fairness across sexes. These findings suggest that the refined models offer a more reliable and equitable approach to early UTI detection in PLWD, addressing participant-specific data variations and enabling clinicians to detect and screen for UTI risks more effectively, thereby facilitating earlier and more accurate treatment decisions.

Urinary Tract Infection Detection in Digital Remote Monitoring: Strategies for Managing Participant-Specific Prediction Complexity

TL;DR

The paper tackles early UTI detection in people living with dementia using in-home passive sensors (Minder dataset) and introduces multitask-inspired MLP designs to address participant variability and sex fairness. It evaluates feature clustering, loss-dependent clustering, and participant ID embeddings, finding that loss-dependent approaches—especially Final Layer Separated—substantially improve validation metrics (e.g., precision from 48.92% to 72.60% and sensitivity from 27.44% to 70.52%), with better gender parity. The results highlight the potential for more reliable, equitable screening in real-world home settings, while revealing challenges in generalizing to new participants and leveraging ID embeddings due to limited per-person data. These insights support integrating refined models into the Minder platform to enable earlier, more accurate UTI management in PLWD, guiding clinical decision-making and resource allocation.

Abstract

Urinary tract infections (UTIs) are a significant health concern, particularly for people living with dementia (PLWD), as they can lead to severe complications if not detected and treated early. This study builds on previous work that utilised machine learning (ML) to detect UTIs in PLWD by analysing in-home activity and physiological data collected through low-cost, passive sensors. The current research focuses on improving the performance of previous models, particularly by refining the Multilayer Perceptron (MLP), to better handle variations in home environments and improve sex fairness in predictions by making use of concepts from multitask learning. This study implemented three primary model designs: feature clustering, loss-dependent clustering, and participant ID embedding which were compared against a baseline MLP model. The results demonstrated that the loss-dependent MLP achieved the most significant improvements, increasing validation precision from 48.92% to 72.60% and sensitivity from 27.44% to 70.52%, while also enhancing model fairness across sexes. These findings suggest that the refined models offer a more reliable and equitable approach to early UTI detection in PLWD, addressing participant-specific data variations and enabling clinicians to detect and screen for UTI risks more effectively, thereby facilitating earlier and more accurate treatment decisions.

Paper Structure

This paper contains 15 sections, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. ML Approaches
  4. The Minder Dataset
  5. Methods
  6. Baseline
  7. Feature Clustering
  8. Loss Dependent
  9. ID Embedding
  10. Results and Interpretation
  11. Baseline
  12. Feature Clustering
  13. Loss Dependent
  14. ID Embedding
  15. Discussion and Conclusion

Figures (2)

  • Figure 1: Loss Analysis and Clustering at Elbow Epoch: \ref{['fig:minder_train_clusters']}: Clusters of average losses for participants from the training dataset at epoch 3, identified as the 'elbow' epoch; \ref{['fig:minder_cluster']}: Clusters of average losses for participants by ID groupings: training on unique train IDs and common IDs, predicting unique test IDs.
  • Figure 2: A visualisation of data points from both the training and test datasets. The colour gradient indicates the corresponding loss values of each data point at the third epoch.