A new study from the University of Oxford demonstrates how clinicians could use data to predict the risk of falls in people with Parkinson's disease (PD) to help improve effective longer-term care planning.

Falls are a common problem for people living with Parkinson's. A recent review estimated that some 60% of all people living with PD have experienced at least one fall. These can lead to injury and hospitalisation, as well as reduced mobility, quality of life, and life expectancy.

Accurate fall risk assessment is crucial for effective care planning for Parkinson's disease patients, but traditional assessments are often subjective and time-consuming. This new study set out to see whether data gathered by wearable sensors during a brief test in a clinic can predict the risk of falling in people with Parkinson's over a period of five years. This could facilitate more effective and longer-term care planning.

The team from the NeuroMetrology lab at the Nuffield Department of Clinical Neurosciences collected data from 104 people with Parkinson's disease, without prior falls, using six wearable sensors. Patients were asked to carry out tasks during short data collection sessions, including a two-minute walk and a 30-second postural sway task. Alongside this the team used several commonly used questionnaires and clinical scales to assess disease severity and the patient's own perception of their decline in mobility.

The article, "Predicting future fallers in Parkinson's disease using kinematic data over a period of five years," is published in npj Digital Medicine.

Machine learning methods were used to analyse the sensor measurements taken from participants at their first study visit, together with follow-up data taken at 24 and 60 months. From this data, they were able to identify the key features that distinguish people with Parkinson's with and without the risk of falling. The analysis revealed significant differences in features related to walking and posture between those who went on to have falls and those who did not.

The study contributes to the understanding of the risk of falls in PD, showing that wearable sensors can deliver accurate predictions of fall risk, but using the novel method of just a short three-minute assessment which minimises the drain on resources for clinicians and the burden on people living with PD.

If we are able to predict these falls, the natural next step is to offer a way to prevent them from happening. Being able to make earlier predictions about who is likely to fall could pave the way for more targeted and effective care programs, ultimately helping to prevent life-threatening falls.

The ability to predict who does or does not need to be offered such a program could also help in wider population health and social care resource planning, saving time and money. Finally, the study findings could also improve participant selection for clinical trials looking at medicines to prevent falls. Resources can be focused on the population with a significant risk of falling within the study timeline, leading to the most accurate results.

Lead author, Professor Chrystalina Antoniades, said, "I am absolutely delighted to see this work published. It is well documented that Parkinson's does increase the risk of falling. This is work in the making from the last few years following patients from our OxQUIP cohort and shows great promise in accurately assessing falls and therefore giving us the opportunity to start thinking of effective care planning. This is a great opportunity for enhancing PD management and starting to develop realistic and effective prevention strategies."