Developing Machine Learning Models for Cognitive Disorder Classification Using Administrative Health Data and Specialist‐Confirmed Diagnoses
Bibliographic record
Abstract
BACKGROUND: Mild cognitive impairment (MCI) and dementia affect a growing proportion of the aging population, placing significant demands on healthcare systems. While administrative health records (AHRs) capture broad healthcare data, their usefulness in identifying MCI and dementia is limited by a lack of expert-confirmed diagnostic information. The Prospective Registry for Persons with Memory Symptoms (PROMPT) provides specialist-confirmed diagnoses and detailed clinical information, creating a unique opportunity to validate machine learning (ML) models in real-world settings. By linking PROMPT's registry data with population-level AHRs, this study addresses current gaps in early detection and staging of cognitive decline, with implications for both clinical practice and health policy. METHOD: We analyzed data from 1,901 adults referred to a tertiary memory clinic in Calgary, Canada (2010-2023), with records linked to provincial Administrative Health Records (AHRs). Cognitive status was classified as "no objective impairment" (NOI), mild cognitive impairment (MCI), or dementia through comprehensive multidisciplinary assessments, including neuropsychological testing and specialist evaluations. Predictor variables derived from AHRs encompassed diagnostic codes, medication dispensations, and healthcare utilization patterns. We developed and compared multiple machine learning (ML) algorithms, including logistic regression, random forests, and ensemble methods, against previously established rule-based approaches from a Canadian study. Model performance was evaluated using the F1 score to balance sensitivity and specificity, along with overall accuracy metrics. RESULTS: Of the 1,901 participants (mean age 73 years, 52% male), 34% (n = 646) were diagnosed with dementia, 44% (n = 836) with MCI, 12% (n = 228) with NOI, and 10% (n = 191) received an 'uncertain' diagnosis. Initial analyses demonstrated expected patterns in cognitive assessment scores across diagnostic groups, including the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA). Complete model performance metrics, including sensitivity, specificity, and area under the receiver operating characteristic curve (AUC), are currently being finalized and will be presented. CONCLUSION: Forthcoming results will clarify how specialist-confirmed data from PROMPT can enhance ML-based classification of MCI and dementia in AHRs. These findings have the potential to improve surveillance of cognitive disorders, inform healthcare resource allocation, and guide earlier interventions. By integrating advanced analytics with validated clinical information, this project could shape future approaches to dementia care in Canada and beyond.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.011 | 0.023 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".