Predicting falls
Funded Solo Fellowship projectPredicting falls and hip fractures using routinely collected care data in high-risk populations: Understanding uncertainty quantification and digital twins methods
Solo Fellow
Project core team
Dr Aishwaryaprajna
Dr Fliss Guest
Professor Umesh Kadam
Professor Krasimira Tsaneva-Atanasova
Hub Research theme
Frailty, fall prediction and fall prevention
Hub Researcher(s) in Residence
Emily Nielsen
Partners
Background
As people get older there is an increasing risk of falls which may result in serious injury, such as a fracture. The risk of falls is due to a range of different factors. Examples include physical fitness, confusion, chronic diseases, medications, as well as the environment such as where a person lives.
Many of these factors are recorded as data in routine electronic healthcare records. Bringing them together and looking at how they change over time, may help to identify a person who is at a higher risk of falls in the future. Risk changes over time, often increasing but can decrease with earlier interventions, so these data offer not only the reduction of risk but also of which ones might be modifiable earlier.
Intended outcomes
Dr Owusu’s team have obtained a national database of 1 million people with heart problems that brings together data from primary care, hospital, accident & emergency, and outpatients, without identifying patients. They want to use these data to develop and test prediction risk models for falls and fracture and understand how this risk might vary with different levels of health severity, and by gender, socio-economic status and ethnicity.