Data-driven decision making for the screening of cognitive impairment in primary care: a machine learning approach using data from the ELSA-Brasil study.

The systematic assessment of cognitive performance of older people without cognitive complaints is controversial and unfeasible. Identifying individuals at higher risk of cognitive impairment could optimize resource allocation. We aimed to develop and test machine learning models to predict cognitive impairment using variables obtainable in primary care settings. In this cross-sectional study, we included 8,291 participants of the baseline assessment of the ELSA-Brasil study, who were aged between 50 and 74 years and were free of dementia. Cognitive performance was assessed with a neuropsychological battery and cognitive impairment was defined as global cognitive z-score below 2 standard deviations. Variables used as input to the prediction models included demographics, social determinants, clinical conditions, family history, lifestyle, and laboratory tests. We developed machine learning models using logistic regression, neural networks, and gradient boosted trees. Participants' mean age was 58.3±6.2 years, 55% were female. Cognitive impairment was present in 328 individuals (4%). Machine learning algorithms presented fair to good discrimination (areas under the ROC curve between 0.801 and 0.873). Extreme Gradient Boosting presented the highest discrimination, high specificity (97%), and negative predictive value (97%). Seventy-six percent of the individuals with cognitive impairment were included among the highest ranked individuals by this algorithm. In conclusion, we developed and tested a machine learning model to predict cognitive impairment based on primary care data that presented good discrimination and high specificity. These characteristics could support the detection of patients who would not benefit from cognitive assessment, facilitating the allocation of human and economic resources.

Data-driven decision making for the screening of cognitive impairment in primary care: a machine learning approach using data from the ELSA-Brasil study

The systematic assessment of cognitive performance of older people without cognitive complaints is controversial and unfeasible. Identifying individuals at higher risk of cognitive impairment could optimize resource allocation. We aimed to develop and test machine learning models to predict cognitive impairment using variables obtainable in primary care settings. In this cross-sectional study, we included 8,291 participants of the baseline assessment of the ELSA-Brasil study, who were aged between 50 and 74 years and were free of dementia. Cognitive performance was assessed with a neuropsychological battery and cognitive impairment was defined as global cognitive z-score below 2 standard deviations. Variables used as input to the prediction models included demographics, social determinants, clinical conditions, family history, lifestyle, and laboratory tests. We developed machine learning models using logistic regression, neural networks, and gradient boosted trees. Participants' mean age was 58.3±6.2 years, 55% were female. Cognitive impairment was present in 328 individuals (4%). Machine learning algorithms presented fair to good discrimination (areas under the ROC curve between 0.801 and 0.873). Extreme Gradient Boosting presented the highest discrimination, high specificity (97%), and negative predictive value (97%). Seventy-six percent of the individuals with cognitive impairment were included among the highest ranked individuals by this algorithm. In conclusion, we developed and tested a machine learning model to predict cognitive impairment based on primary care data that presented good discrimination and high specificity. These characteristics could support the detection of patients who would not benefit from cognitive assessment, facilitating the allocation of human and economic resources.

Ideal vascular health and cognitive performance in the Brazilian Longitudinal Study of Adult Health.

BACKGROUND AND PURPOSE: Most evidence for the association between ideal vascular health (IVH) and cognitive performance comes from high income countries. The aim was to investigate this association in the Brazilian Longitudinal Study of Adult Health. METHODS: Cognition was assessed using the word list, verbal fluency and trail making tests. The IVH score included ideal metrics for body mass index, smoking, physical activity, diet, blood pressure, fasting glucose and total cholesterol. Poor, intermediate and optimal health were characterized in those presenting 0-2, 3-4, 5-7 ideal metrics, respectively. To determine the association between IVH score and cognitive performance, linear regression models adjusted for age, sex, education, race, alcohol use, depression and thyroid function were used. RESULTS: In 12 271 participants, the mean age was 51.3 ± 8.9 years, 54% were women, 57% White and 53% had poor vascular health. Participants with intermediate (ß = 0.064, 95% confidence interval 0.033; 0.096) and optimal health (ß = 0.108, 95% confidence interval 0.052; 0.164) had better global cognitive Z-scores. In addition, interactions of IVH score with age, education and race were found, suggesting a better cognitive performance with higher IVH in older adults, Black/Brown participants and those with lower levels of education. CONCLUSION: Ideal vascular health was associated with better cognitive performance. Older, Black/Brown and low-educated participants had better cognition in the presence of higher IVH scores.

Association between cognitive performance and self-reported glaucoma in middle-aged and older adults: a cross-sectional analysis of ELSA-Brasil.

Recent evidence suggests that glaucoma and Alzheimer's disease are neurodegenerative diseases sharing common pathophysiological and etiological features, although findings are inconclusive. We sought to investigate whether self-reported glaucoma patients without dementia present poorer cognitive performance, an issue that has been less investigated. We employed cross-sectional data from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) and included participants ≥50 years of age without a known diagnosis of dementia and a self-reported glaucoma diagnosis. We excluded those with previous stroke, other eye conditions, and using drugs that could impair cognition. We evaluated cognition using delayed word recall, phonemic verbal fluency, and trail making (version B) tests. We used multinomial linear regression models to investigate associations between self-reported glaucoma with cognition, adjusted by several sociodemographic and clinical variables. Out of 4,331 participants, 139 reported glaucoma. Fully-adjusted models showed that self-reported glaucoma patients presented poorer performance in the verbal fluency test (ß=-0.39, 95%CI=-0.64 to -0.14, P=0.002), but not in the other cognitive assessments. Thus, our results support the hypothesis that self-reported glaucoma is associated with poor cognitive performance; however, longitudinal data are necessary to corroborate our findings.

Association between cognitive performance and self-reported glaucoma in middle-aged and older adults: a cross-sectional analysis of ELSA-Brasil

Recent evidence suggests that glaucoma and Alzheimer's disease are neurodegenerative diseases sharing common pathophysiological and etiological features, although findings are inconclusive. We sought to investigate whether self-reported glaucoma patients without dementia present poorer cognitive performance, an issue that has been less investigated. We employed cross-sectional data from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) and included participants ≥50 years of age without a known diagnosis of dementia and a self-reported glaucoma diagnosis. We excluded those with previous stroke, other eye conditions, and using drugs that could impair cognition. We evaluated cognition using delayed word recall, phonemic verbal fluency, and trail making (version B) tests. We used multinomial linear regression models to investigate associations between self-reported glaucoma with cognition, adjusted by several sociodemographic and clinical variables. Out of 4,331 participants, 139 reported glaucoma. Fully-adjusted models showed that self-reported glaucoma patients presented poorer performance in the verbal fluency test (β=-0.39, 95%CI=-0.64 to -0.14, P=0.002), but not in the other cognitive assessments. Thus, our results support the hypothesis that self-reported glaucoma is associated with poor cognitive performance; however, longitudinal data are necessary to corroborate our findings.