Canada

The CatchIt Early research project examined how artificial intelligence (AI) and machine learning (ML) could support early detection of neurodegenerative disorders such as Alzheimer’s and Parkinson’s disease in Canada. Data from the Canadian Community Health Survey (CCHS) survey was used to model early indicators of cognitive decline in diagnosis based on five (5) variables: perceived mental health, diagnosed mood disorder, sleep issues, mobility limitations, and memory issues. Following preprocessing and dimensionality reduction, random forest and logistic regression learning models were trained to classify individuals at risk. Both models performed poorly, with area under the curve (AUC) values of 0.507 and 0.505, low specificity, and negligible Cohen’s Kappa scores. While sensitivity was moderate, both models failed to distinguish true positives from false positives reliably. These findings indicate that self- reported data alone is insufficient for early detection of neurodegeneration using current AI/ML approaches. To contextualize the prevalence, the choropleth maps and a line graph were created to illustrate a consistent rise in Alzheimer’s cases from 2002 to 2022 and Parkinson’s from 2006 to 2022, with notable provincial variation. While predictive modeling using existing public health data was unsuccessful, the project emphasizes the urgent need for richer, longitudinal, and clinically validated datasets to improve early detection capabilities. Ultimately, the study highlights the growing burden of neurodegenerative disorders in Canada and the importance of advancing data integration to support proactive public health planning and early intervention.

Aging Powerfully: AI Blueprint for Chronic Disease Management and Social Determinants of Longevity

Predicting Alzheimer’s and Parkinson’s Onset Through Machine Learning and Public Health Data

artificial intelligence (ai)

predictive machine learning models

alzheimer’s

early detection

parkinson’s disease

technical paper

2025 National Inter-University Health Data & AI Inquiry Program - this program is a unique project-based inquiry and science communication program that brings together students from diverse academic backgrounds and research interests. Presented research projects explore how generative AI is being applied to medication management, mental health screening, and chronic disease monitoring to align with the future of pharmacy practice.

2025 National Inter-University Health Data & AI Inquiry Program is a unique project-based inquiry and science communication program that brings together students from diverse academic backgrounds and research interests. 

As populations age, early detection of chronic illness is increasingly vital to prevent declines in independence and quality of life. This study investigates whether wearable-derived physiological signals and clinical health records can be used to predict chronic illness in older adults using machine learning models. We used data from the All of Us (AOU) Research Program, including Fitbit wearable summaries and EHR-based clinical measures, to predict three high-burden chronic conditions: osteoporosis, dementia, and heart failure. After filtering for data completeness, we created participant-level summaries using 21 features spanning physical activity, sleep behavior, cardiovascular metrics, and biometric indicators. A final dataset of approximately 2,000 participants aged 60+ was used for modeling. Logistic regression, random forest, and XGBoost models were trained and evaluated using 5-fold cross-validation with stratified train/test splits. We addressed class imbalance with up-sampling during training. Performance was measured using accuracy, F1-score, and area under the ROC curve (AUC). Osteoporosis prediction achieved the highest performance across models (XGBoost AUC = 0.802, F1 = 0.742), supported by features such as basal metabolic rate (BMR) and daily energy expenditure. Dementia and heart failure results were limited by severe class imbalance, with undefined F1-scores and reduced AUCs. These results demonstrate the feasibility of predicting musculoskeletal chronic conditions using wearable and clinical data, and highlight the challenges of modeling rare disease outcomes in imbalanced datasets. Our findings suggest a scalable framework for early risk stratification using passive health monitoring in aging populations.

Predicting Chronic Illness in Aging Populations Using Physiological Markers

Cognitive decline and dementia are increasing public health concerns among Canada’s aging population. Musical interaction has been a potential non-pharmacologic intervention to enhance cognitive and affective well-being. Its benefits may not be widely distributed to all groups, though. This research investigates the association between musical engagement and cognitive function, and social determinants among older Canadians using analysis of two national surveys: the Canadian Community Health Survey (CCHS – Healthy Aging) and the General Social Survey (GSS – Time Use, Cycle 29). Affordable artificial intelligence (AI) techniques, including multiple linear regression and k-means clustering, were employed to explore predictors of depression and musical engagement trends. Depression scores on the Patient Health Questionnaire-9 (PHQ-9) were approximated based on loneliness, income, education, race, and immigration status. Clustering pinpointed high depression groups with limited music access. GSS data illustrated differences in gender and income participation in music despite regression findings indicating poor predictive capability (R² ≈ 0.00). Loneliness was the strongest predictor of depression (coef = 0.36), and the model explained 34% of the variance (R² = 0.34). Findings indicate that racialized, low-income older adults are least likely to have music-based care. These findings identify that AI can uncover disparities and guide culturally responsive, community-based programming to promote equity in healthy aging.

Predicting Loneliness-Linked Health Risks by Analyzing Digital Behaviours Using AI

As public health and demographic shifts coexist, aging populations present challenges for disease prediction and prevention in developing regions. Focusing on the urban and rural areas of Punjab, Pakistan, and Kandahar and Kabul, Afghanistan, the presented research approaches the task of developing a predictive AI framework to quantify polio and hypertension risk. Bayesian networks were employed to model the probability of causal relationships between the parameters of age, vaccination status, sanitation, urbanity or rurality, and water access. Synthetic, yet realistic, data was integrated to reflect true conditions necessary for training the model to understand real data that was later provided. Risk calculations produced capture the multiplicative effects of low vaccination statuses, poor sanitation, and aging on the probability of being at risk. Results show regional differences due to urbanity and rurality showing higher risk of hypertension in the latter, but a main focus on poor sanitation correlating to higher risk of polio. Hypertension is shown to have a positive correlational relationship with aging, while polio is shown to have a negative correlational relationship. The AI model thus shows how localized interventions like, sanitation campaigns or an increase in elderly care, could mitigate risks. These findings underline a need for data-driven public health strategies in resource-scarce environments. The framework’s ability to adapt to parameters suggests applications in other regions, age groups, diseases, and more, provided that real data is used. This contributes to a replicable method for health interventions where inequities reside.

Predictive Modeling of Polio and Hypertension: A Hybrid Model for Disease Risk Forecasting


Sleep is a crucial determinant of health, yet many older Canadians experience inadequate sleep, particularly those facing socioeconomic disadvantages. This study investigates the relationship between sleep deprivation and socioeconomic status (SES) in predicting cognitive decline and cardiometabolic illness among older adults in Canada. We conducted a secondary analysis of seven Canadian datasets, the Canadian Health Measures Survey (CHMS), the Canadian Community Health Survey (CCHS), the Canadian Longitudinal Study on Aging (CLSA), the Canadian Dementia Imaging Protocol (CDIP), Sleep Rest Inequality (SRI) Dataset, and the Wearable Sleep Inequity Dataset. Adults aged 60 and above were evaluated using self-reported and objective sleep indicators (duration, fragmentation, waking after sleep onset), SES measures (income quintile, education), and health outcomes (cardiometabolic illness, BMI, cognition). T-tests, chi-square tests, linear and logistic regression, and unsupervised K-means clustering were used to examine associations and predict health outcomes. Older adults in the lowest income quintile reported significantly shorter sleep (6.0 hrs vs. 6.6 hrs), higher cardiometabolic disease rates (33% vs. 19%), and greater cognitive impairment (28% vs. 14%). Both the risk of mood disorders (RR = 1.45) and cardiometabolic disease (RR = 1.33) were strongly predicted by SES and sleep duration, according to regression models. Cognitive scores and sleep length had a positive correlation, and clustering identified different sleep health profiles by socioeconomic status (SES), with a Silhouette Score of 0.61. Sleep inequality is a measurable and powerful predictor of chronic illness in aging Canadians.

Sleep Inequality and Chronic Illness in Aging Canadians

Chronic conditions may develop as a result of a variety of reasons, such as genetics, lifestyle choices, or a combination of factors. Socioeconomic status has been found to be a major factor in the development of multiple chronic diseases. This study aimed to examine the socio- economic factors linked with cardiovascular and musculoskeletal chronic diseases and hypertension using multi-component analysis (MCA). Also, we probed further into the relationship between socioeconomic factors and chronic musculoskeletal diseases (CMDs). In contrast to other similar studies conducted with respect to Canadian seniors, which had used inferential analysis techniques, this study focused on the impact of socioeconomic factors through the use of logistic regression and random forest models. Both models found education level and mental health to be the primary predictors of CMDs in seniors. While education level could be the cause, mental health status can be both a cause and effect, as the diseases’ development might affect seniors’ mental well-being, which can have adverse effects on pre-existing symptoms. Also, lower education levels were found to be a key predictor for all three categories. These findings lead us to recommend policies which: address financial disparities in accessing treatment, support senior immigrants in culturally-aware ways, and encourage mental-wellbeing through counselling and community-building activities.

Socio-economic Predictors of Chronic Musculoskeletal Diseases

Air pollution is a significant contributor to adverse health outcomes, particularly among older adults. Chronic exposure to fine particulate matter (PM2.5), nitrogen dioxide (NO₂), and ozone (O₃) has been linked to elevated risks of respiratory diseases, impaired lung function, and premature mortality. This study investigates the relationship between long-term pollutant exposure and respiratory health in aging populations by integrating data from the National Health and Nutrition Examination Survey (NHANES), the U.S. Environmental Protection Agency (EPA), and satellite-based monitoring systems. A combination of statistical modelling, machine learning algorithms, and geospatial analysis in R and Python was applied to identify regions where older individuals are most at risk. The findings aim to support targeted public health policies and improve understanding of environmental health disparities across geographic and demographic lines.

Spatial and Demographic Disparities in Respiratory Disease Risk among Older Adults

Background: Parkinson’s Disease (PD) is a progressive neurodegenerative disorder primarily affecting adults 60 years of age and older. It is typically diagnosed after motor symptoms appear by which point extensive neuronal damage has often already occurred. Earlier detection is key to improve prognosis and enable preventative strategies. Objective: This research study aims to develop a diagnostic process focused on factors like motor and non-motor symptoms, cardiometabolic factors, demographic factors, and medical history to detect early manifestations of PD or establish strong risk associations for the development of PD to aid in early treatment. Methods: A Random Forest classifier was applied to a dataset of individuals aged 50-90 years, using features including motor symptoms, non-motor symptoms, cardiometabolic indicators, and demographic features. Results: The model achieved high predictive accuracy (ROC AUC = 0.75) with motor and non- motor features being the strongest individual predictors. Models trained soley on demographic, lifestyle, or cardiometabolic factors performed poorly (ROC AUC = 0.51-0.53). Non-motor symptoms such as sleep and cognitive decline demonstrated moderate standalone predictive value, supporting their role in early PD screening. Conclusion: Machine learning models integrating symptom-based data outperform those based on demographic or risk factor profiles alone. While the findings suggest valuable utility of non- motor features, limitations such as the use of a single dataset may affect generalizability. Future work should validate these findings across diverse populations to support real-world clinical applications.

The Best Predictors of Parkinson’s Disease: A Subgroup Analysis Using Random Forests

Non-Alcoholic Fatty Liver Disease (NAFLD) is a critical contributor to aging worker health concerns. While metabolic and lifestyle factors are recognized influences, psychosocial stress factors such as job strain are inadequately explored. This meta-analysis studies the association between job strain and NAFLD risk in adults aged 30 and older. Ten peer-reviewed studies were selected through a thorough database screening. Five studies with adjusted odds ratios were incorporated in a quantitative synthesis using a random-effects model. The findings demonstrate a positive association between job strain and increased NAFLD risk. Job strain is measured through factors such as long working hours, shift work, and occupational stress. Odds ratios across studies ranged from 1.12 to 2.33, indicating elevated risk in job-strain-exposed individuals. Older adults were particularly more vulnerable, suggesting that aging increases the consequences of liver damage. Funnel and forest plot analyses showed minimal publication bias and a statistically significant pooled effect (p = 0.0236), though high heterogeneity (I² = 78.17%) was observed. These results highlight job strain as an important risk factor for NAFLD in aging populations. This understanding of the likelihood of NAFLD chances in aging workers offers an opportunity for intervention strategies. Public health efforts should consider job-related stress in NAFLD risk assessments and promote workplace health policies to mitigate this hidden burden.

The Hidden Burden of Work: Job Strain and NAFLD in Aging Populations

Alzheimer’s disease is a progressive neurodegenerative disorder and the leading cause of dementia, marked by cognitive decline, emotional disturbances, and neuropathological changes. Despite extensive research, accurately predicting Alzheimer progression remains challenging. This study introduces a tri-model machine learning framework that explores how music features, electroencephalography (EEG) signals, and cognitive measures could forecast disease progression. Four publicly available datasets were analyzed: EEG responses during music listening tasks, resting state EEG recordings and Mini-Mental State Examination (MMSE) scores, longitudinal cognitive assessments with dementia status classifications, and EEG paired with working memory tasks under varying music conditions. Three sequential models were created. Model A tested if acoustic features from music could predict EEG activity, finding no link and showcasing the difficulty of decoding music driven neural responses without disease specific markers. Model B assessed the relationship between EEG features and cognitive status using MMSE, presenting limited predictive power, with non-linear models like Random Forest only marginally outperforming linear approaches. Model C predicted Alzheimer’s progression using longitudinal cognitive and demographic data, showing moderate success in predicting cognitive decline and dementia status change. Overall, these findings emphasize both the potential and limitations of using multimodal data and machine learning approaches in Alzheimer research. While predictive accuracy was modest, this study provides a foundation for future research aiming to enhance early detection and support the development of personalized, non-pharmacological interventions for Alzheimer’s disease management.

Tri-model Forecasting of Alzheimer’s Progression using Music, EEG, and Cognition

Polypharmacy, commonly defined as the concurrent use of five or more medications, is frequently observed in older adults and may be associated with adverse health outcomes. This study aimed to describe patterns of polypharmacy and related characteristics in a nationally representative sample of adults aged 65 and older using data from the 2017–2018 cycle of the National Health and Nutrition Examination Survey (NHANES). Demographic, clinical, and prescription drug data were extracted and merged using participant identifiers. Continuous variables were standardized using the StandardScaler. Categorical variables such as sex, race and ethnicity, and medical conditions were one-hot encoded to ensure appropriate treatment in clustering. Prescription medications were mapped to therapeutic category codes based on the Multum Lexicon system. Each participant’s drug profile was embedded into a 16-dimensional vector using a skip-gram model trained on categorical co-occurrence. Unsupervised clustering was conducted using the KMeans algorithm. The optimal number of clusters was selected as three based on elbow plot and silhouette score analysis. Cluster labels were reassigned post hoc according to the average number of medications per participant. Descriptive analyses were performed to compare demographic and clinical features across clusters. Independent t-tests were conducted on the standardized income-to-poverty ratio to assess variation in economic status between groups. Additional categorical comparisons were examined for patterns related to income, education, chronic conditions, and depressive symptoms. This approach combines demographic, clinical, and medication data to describe distinct groupings of older adults based on polypharmacy and associated variables.

Downloads

Next from Aging Powerfully: AI Blueprint for Chronic Disease Management and Social Determinants of Longevity

Predicting Chronic Illness in Aging Populations Using Physiological Markers

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES