Artificial intelligence-based automated scoring of the Clock Drawing Test in older adults in primary care: association with Mini-Mental State Examination scores, discriminative performance, and agreement with clinician rating.
Аннотация
BACKGROUND
Cognitive impairment is common in older adults, and its early detection in primary care remains challenging. The Clock Drawing Test (CDT) is a practical tool, but variability in its scoring limits standardization.
OBJECTIVE
To evaluate artificial intelligence (AI)-based automated CDT scoring in primary care by examining its association with Mini-Mental State Examination (MMSE) scores, discriminative performance for cognitive impairment, and agreement with clinician rating.
METHODS
In this cross-sectional study, 207 adults aged ≥65 years were assessed in a primary care setting. CDT drawings were scored manually by a neurologist and automatically using a multimodal generative AI system based on the Manos and Wu 10-point method. Associations with MMSE were analyzed, and discriminative performance was assessed using receiver operating characteristic analysis. Agreement was evaluated using intraclass correlation coefficient and Bland-Altman analysis.
RESULTS
AI-based CDT scores demonstrated a moderate positive correlation with MMSE (ρ = .437, P < .001). Both methods significantly discriminated MMSE-defined cognitive impairment, although manual scoring yielded a significantly higher area under the receiver operating characteristic curve than AI-based scoring (.779 vs .715; DeLong P = .028). AI-based scoring provided higher sensitivity but lower specificity. Agreement between methods was good (intraclass correlation coefficient = .721), with AI tending to assign slightly higher scores (mean difference: 1.13).
CONCLUSION
AI-based CDT scoring is associated with global cognitive performance and offers meaningful discrimination of cognitive impairment in primary care. However, given lower accuracy and individual-level variability, it should be considered a supportive tool rather than a replacement for clinician assessment.
Предметные рубрики MeSH
Подготовка и перевод публикации
Обработка статьи · этапы, ошибки и управление
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Тип публикации указан по данным источника. Уровень доказательности автоматически не присваивается.