Predicting Multiple Clinical Outcomes Related to Functional Recovery and Social Isolation Among Older Adults After Lower-Limb Fracture or Hip Replacement
Authors: Santosh Ray, Pratik K. Mishra, Ali Abedi, Charlene H. Chu, Amir Ahmad, Shehroz S. Khan
Organizations: Faculty of Information Technology, Liwa University, Abu Dhabi, United Arab Emirates · Institute of Biomedical Engineering, University of Toronto, Toronto, ON, Canada · KITE Research Institute - Toronto Rehabilitation Institute, University Health Network, Toronto, Canada · Peter Munk Cardiac Centre, University Health Network, Toronto, Canada · Lawrence Bloomberg Faculty of Nursing, University of Toronto, Toronto, Canada · College of Information Technology, United Arab Emirates University, Al Ain, United Arab Emirates · College of Engineering and Technology, American University of the Middle East, Egaila, 54200, Kuwait
Older adults recovering after lower-limb fracture or hip replacement may experience complex recovery trajectories. Most of the time, these clinical aspects are studied in isolation, masking their joint impact on recovery. This study used the MAISON-LLF dataset, which contains multimodal sensor and clinical assessment data from 18 older adults recovering in the community after lower-limb fracture or hip replacement. Participants were monitored for up to eight weeks, corresponding to a maximum of 1,008 participant-days of sensor monitoring. Forty-six daily features were extracted from indoor motion, acceleration, step count, heart rate, out-of-home mobility, and sleep data. Five clinical outcomes were assessed every two weeks: the Social Isolation Scale, Oxford Hip Score, Oxford Knee Score, Timed Up and Go test, and 30-second Chair Stand test. We utilize an inherent relationship between multi-modal sensor data and different clinical scores and formulate it as a multi-output regression problem. We tested various machine learning and deep learning single- and multi-output regression algorithms to predict these scores simultaneously. The results showed that predicting clinical scores jointly was better than separately. The tabular DL multi-output regressor, NODE, gave a remarkable performance of MSE=3.96 and MAE=1.02 in comparison to other multi- and single-output regressors. The SHAP feature analysis further showed the importance of including multimodal sensors to provide a good estimate of patients' recovery trajectory. This work may support the simultaneous assessment of functional recovery and social engagement among community-dwelling older adults and ultimately help improve their care and quality of life.
Figures & tables
Data modality 1
Feature 1
Description
acceleration-
count, entropy, kurtosis, mean, skew, std, and sum
The total count, entropy, kurtosis, mean, skewness, standard deviation, and sum of acceleration data in a day.
coefficient-of-variation
Measures the relative variability of daily physical activity, indicating how consistently an individual moves throughout the day in relation to their average activity level.
minutes-with-data
Represents the total number of minutes within a 24-hour period (from 12:00 AM to the following 12:00 AM) during which acceleration data was recorded, providing insight into the smartwatch’s active recording time, even if it was not necessarily worn on the wrist.
hours-with-data
Represents the total number of hours within a 24-hour period (from 12:00 AM to the following 12:00 AM) during which acceleration data was recorded, providing insight into the smartwatch’s active recording time, even if it was not necessarily worn on the wrist.
Represents the number of detected movement events between 12:00 AM and 6:00 AM, 6:00 AM and 12:00 PM, 12:00 PM and 18:00 PM, 18:00 PM and 12:00 AM, based on acceleration values exceeding a dynamic threshold, and reflects physical activity or restlessness.
movement-events-24h
Represents the total number of detected movement events over a full 24-hour period (from 12:00 AM to the following 12:00 AM), based on acceleration values exceeding a dynamic threshold, and provides an overall measure of daily physical activity.
TABLE I: The data modalities, feature names, and their description. 1 Feature naming convention in the dataset: ‘data modality’ +‘-’+‘feature’.
Fig. 1: Correlation matrix of the clinical outcomes for the overall dataset.
Fig. 2: Correlation matrix of the outcomes for each individual fold as part of the leave-one-participant-out cross validation.
TABLE II: Results of multi-output and single-output regression for leave-one-person-out cross-validation in terms of (a) MSE and (b) MAE. Unless the errors are equivalent, the best values for each outcome are highlighted in bold. The blue shaded cells signify the better values between single and multi-output regression.
TABLE III: Percentage improvement of multi-output over single-output regressors for leave-one-person-out cross-validation. The best values for each outcome across all the regressors are highlighted in bold.
TABLE IV: Percentage improvement of multi-output regressors with respect to LR baseline for leave-one-person-out cross-validation. The best values for each outcome across all the regressors are highlighted in bold.
TABLE V: Results of multi-output and single-output regression for leave-one-week-out cross-validation in terms of (a) MSE and (b) MAE. Unless the errors are equivalent, the best values for each outcome are highlighted in bold. The blue shaded cells signify the better values between single and multi-output regression.
TABLE VI: Percentage improvement of multi-output over single-output regressors for leave-one-week-out cross-validation. The best values for each outcome across all the regressors are highlighted in bold.
TABLE VII: Percentage improvement of multi-output regressors with respect to LR baseline for leave-one-week-out cross-validation. The best values for each outcome across all the regressors are highlighted in bold.
TABLE VIII: SHAP-based ranking of features for NODE multi-output regressor for leave-one-person-out cross-validation.
Wearable and mobile sensing technologies enable continuous monitoring of human behavior and health in real-world settings. However, predictive modeling in longitudinal multimodal data remains challenging, particularly when targeting complex or clinically derived outcomes. In this work, we present a longitudinal multimodal study of 66 older adults conducted in real-world conditions and combining wearable sensing, behavioral monitoring, and clinical assessments. This setting provides a rare opportunity to study an underrepresented population in long-term, into-the-wild conditions. Building on this dataset, we investigate how the alignment between sensed signals and target variables affects predictive performance across health-related tasks. We design a unified evaluation framework spanning tasks with increasing levels of observability, including Activity Levels prediction, Sleep Duration estimation, and Sleep Apnea Severity classification. Our results reveal a clear gradient of predictability: highly observable behavioral targets achieve robust performance (macro-F1 65%), while more abstract outcomes remain challenging despite consistent improvements over baseline models. Moreover, through explainability analysis, we show that historical features consistently emerge as the most informative predictors, highlighting the central role of longitudinal information.
Flavio Di Martino, Mattia G. Campana, Marcello Magno +2
Objective sleep assessment relies on polysomnography (PSG), yet clinical impact is often better reflected in patient-reported outcomes (PROs) such as sleepiness and fatigue. Existing summary indices, including the Apnea-Hypopnea Index (AHI), provide limited insight into the multidomain physiology underlying functional recovery. We propose an interpretable, causal-discovery-guided framework for deriving a hierarchical Sleep Recovery Score (SRS) from multimodal PSG. Using two large population cohorts (MESA: n=1,540; MrOS: n=825), we apply directed acyclic graph (DAG) learning to identify candidate physiological drivers spanning respiratory burden, hypoxic burden, sleep fragmentation, sleep architecture, and autonomic regulation. Although derived from clinical PSG, these domains map naturally to sensing streams increasingly available in connected health technologies, including wearable ECG, oximetry, and sleep-stage estimation devices. To preserve mechanistic plausibility, we introduce a two-stage screening process that combines physiology-based constraints with constrained LLM-assisted auditing to identify and remove structural confounders and construct-overlapping variables. Across cohorts, these five domains emerge as recurrent physiological domains associated with recovery, and the resulting SRS shows up to 3.4× stronger alignment with perceived recovery than AHI. By linking multimodal sleep physiology to patient-centered outcomes through an interpretable, bias-aware, and domain-structured framework, this work provides a practical foundation for recovery modeling across both clinical sleep studies and emerging smart and connected health settings.
Saba A. Farahani, Elahe Khatibi, Manoj Vishwanath +2
Functional decline in older adults is typically recognized only after falls or observable gait impairment, closing the window for prevention. We investigated whether temporal trajectories of routine biomarkers, already recorded but rarely analyzed longitudinally, can identify patients in the pre-clinical phase of mobility decline. Using the All of Us Research Program (N = 297,861; 11.1% cases), we derived trajectory features (slope, variability, delta, mean) for twelve biomarkers over a three-year pre-index window. LightGBM models incorporating trajectories significantly outperformed static laboratory summaries (AUROC 0.797 vs. 0.755; DeLong p < 0.001; AUPRC 0.380 vs. 0.304). A 1:1 age- and sex-matched analysis confirmed an independent trajectory signal (AUROC 0.727 vs. demographics-only 0.680). A horizon analysis demonstrated sustained prediction 3-12 months before decline onset (AUROC 0.768-0.740). Because the model uses only measurements already ordered in routine care, it supports passive, zero-burden EHR integration for early detection of pre-clinical functional decline.
Rashmita Kudamala, Aravind V. Kuruvikkattil, Lalitha Pranathi Pulavarthy +1
Dept. of Biomedical Engineering and Informatics, Luddy School of Informatics, Computing, and Engineering, Indiana University Indianapolis, IN, USA