A Probabilistic Finite-State Automata Framework for Monitoring Long-Term Activities of Daily Living

Aging is an increasingly challenging healthcare issue with long-term sociopolitical implications, requiring more sophisticated management of healthcare services for the elderly. A key element in managing such services is the knowledge about a typical elderly person's Activities of Daily Living (ADLs). However, there is little study on the model-based objective and standardized ADLs assessment. This paper presents the concept of tracking the health status of the elderly by monitoring their individual patterns of ADLs: the specific pattern of the elderly model as probabilistic state-transition structures, and the transition probabilities in such a probabilistic model are considered as a representation of the health status of a typical older person. The typical pattern of ADLs changes with the health status due to changes in the residential environment, progressive aging, or the onset of certain diseases. Such changes are reflected (over time) in the model, and therefore monitoring the transition probabilities to assess ADLs represents a possible way to track the health status of an older person and to alert relevant health service professionals if some change that warrants subsequent careful medical attention.