Identifying Mild Cognitive Impairment from Routinely Collected Physical, Functional and Psychosocial Measures in Community-Dwelling Older Adults: A Comparative Machine-Learning and Regression Modelling Study in the Northern Japanese ORANGE Registry.
5 affiliations
- 1Department of Occupational Therapy, Graduate School of Medicine, Akita University, 1-1-1 Hondo, Akita 010-8543, Japan.
- 2Advanced Research Center for Geriatric and Gerontology, Akita University, 1-1-1 Hondo, Akita 010-8543, Japan.
- 3Clinical Research Promotion and Support Office, Future Cooperative Research Organization, Akita University, 1-1-1 Hondo, Akita 010-8543, Japan.
- 4Integrated Community Support Center, Public Health and Welfare Department, City Hall of Yokote, Yokote, Akita 010-8543, Japan.
- 5Omori Municipal Hospital, Yokote, Akita 010-8543, Japan.
Abstract
Background: Community health check-ups routinely collect non-cognitive data. Whether these support mild cognitive impairment (MCI) triage, and whether machine learning outperforms regression, is unclear. Methods: Cross-sectional analysis of 2650 community-dwelling adults aged ≥ 65 in Akita, Japan (2018-2025). MCI was defined by the NCGG-FAT (1.5 SD, age/education-adjusted norms). Thirty-four pre-specified non-cognitive predictors were modelled using tree ensembles, penalised regression, other machine-learning families and simple reference scores, with stratified 70/30 splitting, DeLong tests, 100 repeated splits, training-derived thresholds, calibration metrics and decision-curve analysis. Results: MCI prevalence was 28.6%. Random forest discriminated best (ROC AUC 0.722, 95% CI 0.681-0.761; PR-AUC 0.542) and was well calibrated (slope 1.048; intercept -0.024), but its advantage over LASSO regression was marginal (ΔAUC +0.031, p = 0.044) and absent versus gradient boosting (p = 0.059). All multivariable models greatly outperformed the Kihon Checklist (AUC 0.550, p < 0.001). Neither operating point was deployable: at 0.5, sensitivity was 0.269 (166/227 cases missed); at the training-derived Youden threshold (0.321), sensitivity was 0.604, specificity 0.699, and PPV 0.445. Net benefit exceeded assess-all/assess-none between threshold probabilities 0.15-0.60. Conclusions: Routine non-cognitive data support moderate, well-calibrated MCI discrimination, but machine learning added little over regression. This is an exploratory triage aid requiring external validation.
Related
- Comparative Value of Cognitive and Functional Assessments for Predicting 24-Month Progression from Mild Cognitive Impairment to Alzheimer's Disease: An ADNI Cohort Study
- Menopausal symptoms and work impairment: the moderating role of psychosocial work factors
- Psychosocial Distress and Radiation Treatment Interruptions in Patients With Cancer Using Machine Learning and Regression: Retrospective Cohort Study.
- Comparative efficacy of exercise modalities for mild cognitive impairment in older adults: A network meta-analysis
- Comparative efficacy of non-pharmacological interventions for amnestic mild cognitive impairment: a network meta-analysis
BibTeX
@article{kodama2026identifying,
title = {{Identifying Mild Cognitive Impairment from Routinely Collected Physical, Functional and Psychosocial Measures in Community-Dwelling Older Adults: A Comparative Machine-Learning and Regression Modelling Study in the Northern Japanese ORANGE Registry.}},
author = {Ayuto Kodama and Takako Ohnuma and Kana Sasaki and Kaoru Sugawara and Nobuhiro Fujiyama and Youko Umetsu and Tsuyoshi Ono and Hidetaka Ota},
journal = {Geriatrics (Basel, Switzerland)},
year = {2026},
month = oct,
doi = {10.3390/geriatrics11050138},
url = {https://europepmc.org/article/MED/42930017},
note = {CC BY},
}
RIS
TY - JOUR
TI - Identifying Mild Cognitive Impairment from Routinely Collected Physical, Functional and Psychosocial Measures in Community-Dwelling Older Adults: A Comparative Machine-Learning and Regression Modelling Study in the Northern Japanese ORANGE Registry.
AU - Ayuto Kodama
AU - Takako Ohnuma
AU - Kana Sasaki
AU - Kaoru Sugawara
AU - Nobuhiro Fujiyama
AU - Youko Umetsu
AU - Tsuyoshi Ono
AU - Hidetaka Ota
PY - 2026
DA - 2026/10/01
JO - Geriatrics (Basel, Switzerland)
DO - 10.3390/geriatrics11050138
UR - https://europepmc.org/article/MED/42930017
AB - Background: Community health check-ups routinely collect non-cognitive data. Whether these support mild cognitive impairment (MCI) triage, and whether machine learning outperforms regression, is unclear. Methods: Cross-sectional analysis of 2650 community-dwelling adults aged ≥ 65 in Akita, Japan (2018-2025). MCI was defined by the NCGG-FAT (1.5 SD, age/education-adjusted norms). Thirty-four pre-specified non-cognitive predictors were modelled using tree ensembles, penalised regression, other machine-learning families and simple reference scores, with stratified 70/30 splitting, DeLong tests, 100 repeated splits, training-derived thresholds, calibration metrics and decision-curve analysis. Results: MCI prevalence was 28.6%. Random forest discriminated best (ROC AUC 0.722, 95% CI 0.681-0.761; PR-AUC 0.542) and was well calibrated (slope 1.048; intercept -0.024), but its advantage over LASSO regression was marginal (ΔAUC +0.031, p = 0.044) and absent versus gradient boosting (p = 0.059). All multivariable models greatly outperformed the Kihon Checklist (AUC 0.550, p < 0.001). Neither operating point was deployable: at 0.5, sensitivity was 0.269 (166/227 cases missed); at the training-derived Youden threshold (0.321), sensitivity was 0.604, specificity 0.699, and PPV 0.445. Net benefit exceeded assess-all/assess-none between threshold probabilities 0.15-0.60. Conclusions: Routine non-cognitive data support moderate, well-calibrated MCI discrimination, but machine learning added little over regression. This is an exploratory triage aid requiring external validation.
SN - 2308-3417
ER -