Abstract:Abstract Background Existing literature highlights the need for better risk assessment of mild cognitive impairment (MCI) to mitigate the onset and burden of Alzheimer’s Disease and Related Dementia (ADRD) through earlier interventions. Hispanics/Latinos are the largest US ethnic/racial minority, at higher risk for developing ADRD compared to Non‐Hispanic Whites, but underrepresented in dementia‐related research. This study develops prediction models for targetable protective and risk factors of MCI in diverse, middle‐aged Hispanics/Latinos in the US. Method We use data (n = 4246, average baseline age = 56‐years) from the Hispanic Community Health Study/ Study of Latinos (HCHS/SOL; 2008‐2011), a multisite prospective cohort study of diverse Hispanic/ Latinos, and its ancillary study, the SOL‐Investigation of Neurocognitive Aging (SOL‐INCA). Our outcome is prevalent MCI at SOL‐INCA visit (average 7‐years after the baseline) operationalized based on National Institutes of Aging‐Alzheimer’s Association criteria. Using nearly 40 baseline indicators representing (1) sociodemographic characteristics, (2) childhood factors, (3) acculturation factors, (4) biological and (5) behavioral markers, and (6) mental and (7) functional health, we compared a supervised machine learning (Random Forest; RF) and a standard statistical technique (logistic regression; LR) in identifying predictors. Additionally, to evaluate RF as a feature selection tool, we estimated Area Under the Curve (AUC) from LR using all factors and a subset of factors identified as important by RF. Result LR identified socioeconomic conditions and mental and physical health scores as among the leading factors predicting MCI. RF identified a larger subset of predictors and differed from LR in ranking the relative importance of common predictors. Removal of factors identified as insignificant in predicting MCI by RF (n = 17) slightly improved classification performance in LR (AUC all factors = 0.692, AUC subset by RF = 0.697). Conclusion Supervised machine learning (ML) identified different predictors for MCI compared to standard methods and was compelling as a feature selection method. This suggests potential value of ML to focus on significant predictors for MCI prediction. Substantively, our findings emphasize the importance of accounting for life‐course risk and protective factors for predicting cognitive health. Further investigations are needed to test additional ML modalities and better understand how complex interactions of life‐course multi‐domain factors can be addressed through interventions.

A supervised machine learning approach to predicting mild cognitive impairment among diverse Hispanics/Latinos: The SOL‐INCA Study

Sociodemographic Factors Predict Incident Mild Cognitive Impairment: A Brief Review and Empirical Study.

Predicting mild cognitive impairment among Chinese older adults: a longitudinal study based on long short-term memory networks and machine learning

Machine Learning Approach for Improved Longitudinal Prediction of Progression from Mild Cognitive Impairment to Alzheimer’s Disease

Establishment of a Mild Cognitive Impairment Risk Model in Middle-Aged and Older Adults: a Longitudinal Study

MACHINE-LEARNING-BASED INTERPRETABLE PREDICTION FOR MILD COGNITIVE IMPAIRMENT: FROM ADNI DATABASE

Personalized screening and risk profiles for Mild Cognitive Impairment via a Machine Learning Framework: Implications for general practice

Machine learning-based prediction of mild cognitive impairment among individuals with normal cognitive function

Machine Learning Predicts Conversion from Normal Aging to Mild Cognitive Impairment Using Medical History, APOE Genotype, and Neuropsychological Assessment

Using Machine Learning to Predict Cognitive Impairment Among Middle-Aged and Older Chinese: A Longitudinal Study

Machine learning algorithms to predict mild cognitive impairment in older adults in China: A cross-sectional study

Analysis of Risk Factors in Dementia Through Machine Learning

Machine learning for the prediction of cognitive impairment in older adults

Selecting the most important self-assessed features for predicting conversion to mild cognitive impairment with random forest and permutation-based methods

A Risk Prediction Model Based on Machine Learning for Cognitive Impairment Among Chinese Community-Dwelling Elderly People With Normal Cognition: Development and Validation Study

A Risk Prediction Model Based on Machine Learning for Cognitive Impairment Among Chinese Community-Dwelling Elderly People With Normal Cognition: Development and Validation Study (Preprint)

Subjective cognitive decline and cognitive change among diverse middle-aged and older Hispanic/Latino adults: Results from the Study of Latinos-Investigation of Neurocognitive Aging (SOL-INCA)

Diabetes, cognitive decline and mild cognitive impairment among diverse Hispanics/Latinos: Hispanic Community Health Study/Study of Latinos (HCHS‐SOL) investigation of cognitive aging results

Predicting Alzheimer’s Disease from Cognitive Footprints in Mid and Late Life: How Much Can Register Data and Machine Learning Help?

Diabetes, Cognitive Decline, and Mild Cognitive Impairment Among Diverse Hispanics/Latinos: Study of Latinos–Investigation of Neurocognitive Aging Results (HCHS/SOL)

Predicting Alzheimer's disease from cognitive footprints in mid and late life: How much can register data and machine learning help?