Prognostic value of movement quality parameters derived from wrist sensor data in a large population
Prognostic value of movement quality parameters derived from wrist sensor data in a large population
Lay summary
Falls, dementia, depression and Parkinson's disease are common in older age and often lead to long term disability, premature institutionalization and significant socio-economic burden. However, timely diagnosis and intervention are difficult and often relate to some older people's inability to report their signs and symptoms. Previous studies have demonstrated that clinical mobility tests, such as walking speed and sit-to-stand time, can predict the onset of geriatric conditions. This study aims to investigate whether more precise predictions can be achieved through using wearable sensor data. Compared to clinical tests, wearable sensors record usual performances in everyday living, collect more data on movement quality and document a longer exposures. UK Biobank has seven-day wrist-worn accelerometry data from around one hundred thousand participants with over three-years follow up on health outcomes. Through further processing and analysis of data, we aim to recognize daily activities, derive movement quality parameters, and predict onset of adverse events. These prediction models will form the basis for large scale screening programs, which will facilitate prevention and early intervention of age-related syndromes and diseases.
Scope extension:
Research question(1): to what extent can movement quality parameters derived from wrist sensor data predict falls, dementia, depression and Parkinson's disease
Aims:
1) To identify activities (e.g. walking, sit-to-stand) with wrist-accelerometer data with machine learning
2) To estimate movement quality parameters (e.g. cadence, harmonic ratios, sit-to-stand power) within the activity-matched acceleration data
3) To predict falls and onset of dementia, depression and Parkinson's disease with estimated movement quality parameters
Research Question(2): What are the differences in objectively-measured daily-life mobility between individuals with or at-risk of developing chronic diseases or adverse health events and healthy controls?
Aims:
4) To provide age- and gender-specific normative values of movement quality parameters
5) To compare the differences of movement quality parameters between people with chronic health conditions (including cardiopulmonary diseases, chronic obstructive pulmonary diseases, chronic pain, cognitive impairments, depression, frailty, motor neurone diseases, Parkinsonian diseases and sarcopenia etc.) and healthy controls
6) To compare the differences of movement quality parameters between people at-risk of falls or developing chronic health conditions and healthy controls
7) To evaluate the predictive values of movement quality parameters in falls and mortality, chronic health conditions
Research Question(3): Do/How genetic factors influence the movement quality parameters?
Aims:
8) To run genome wide association studies on objectively measured movement quality parameters
9) To construct polygenic score for movement quality parameters and estimate its predictive value
10) To estimate genetic correlation between movement quality parameters and related phenotypes such as dementia, depression and Parkinson's disease.