Using wearable sensors to monitor gait and predict falls in Parkinson’s disease
Falls are a common and serious concern for people living with Parkinson’s disease. Changes in balance, stride length, walking speed and turning can develop gradually, then become more pronounced during medication “off” periods, fatigue or busy movement environments. A short clinical assessment may miss these fluctuations, particularly when symptoms vary across the day.
Wearable sensors offer a practical way to observe movement in daily life. Small devices worn on the wrist, ankle, waist or footwear can record steps, acceleration, postural transitions and gait variability. Used carefully, this information may help clinicians identify changing mobility patterns and support earlier, more personalised interventions.
For Australian health services, the value lies in connecting home-based data with neurological care, physiotherapy, occupational therapy and falls prevention programs. The strongest results will come from partnerships that treat sensor readings as one part of a broader clinical picture, rather than as an automatic diagnosis.
What wearable gait monitoring can measure
Inertial measurement units use accelerometers and gyroscopes to detect how the body moves. A sensor may estimate walking speed, cadence, step-time asymmetry, stride regularity and the number of steps taken each day. Some systems can also identify sit-to-stand movements, prolonged inactivity and changes in turning performance.
Freezing of gait is particularly important in Parkinson’s disease. A person may feel that their feet are stuck to the floor when approaching a doorway, turning or starting to walk. Algorithms can search for movement signatures associated with freezing, although real-world detection remains difficult when devices are loose, incorrectly positioned or exposed to interference from other activities.
Why everyday data matters
A person may walk reasonably well during a ten-metre clinic test but struggle when carrying shopping, moving through a crowded Brisbane railway station or walking on uneven paths. Continuous monitoring can reveal patterns that a single appointment cannot, including reduced mobility late in the day or increased instability after medication effects wear off.
Longitudinal data can also support conversations between appointments. A physiotherapist may compare gait variability before and after a balance program, while a neurologist may examine whether mobility changes coincide with medication timing. The goal is useful clinical context, not an endless stream of numbers that increases anxiety for patients and carers.
Turning sensor readings into fall-risk insight
Fall prediction is more complex than counting steps. Risk may be associated with shorter strides, greater variation between steps, slower turns, repeated freezing episodes and difficulty rising from a chair. A useful model combines these signals with previous falls, cognitive changes, vision, footwear, home hazards and medicines that can affect blood pressure or alertness.
Machine-learning systems can identify patterns across large datasets, but their performance may vary between populations. A model trained in a European research centre may not work equally well for older Australians, people in regional Queensland or Aboriginal and Torres Strait Islander communities. Local validation, transparent reporting and clinician oversight are essential before automated risk scores influence care.
Building evidence through collaboration
Wearable technology sits at the intersection of engineering, neuroscience, allied health and community care. Collaborative research can establish consistent sensor protocols, assess usability and determine whether alerts actually reduce falls or improve confidence. The Brisbane Diamantina network provides an example of the type of connected environment needed to translate health research into practice.
Partnerships also help researchers focus on outcomes that matter to patients and families. A university team may develop an algorithm, while a hospital tests workflow, physiotherapists assess function and community organisations identify barriers to adoption. Lessons from a heart failure partnership show how shared clinical and research expertise can turn promising ideas into service improvements.
Designing technology for Australian care
The local wearable market includes smartwatches, smartphone applications, research-grade inertial sensors and connected insoles. Consumer devices may be affordable and familiar, but their algorithms are often proprietary and may not provide raw data or clinically validated fall-risk measures. Research systems can be more precise, yet they may require charging, technical support and specialist analysis.
Access also differs across Australia. A participant in inner Brisbane may have regular appointments and reliable internet, while someone in regional Queensland may face long travel distances, limited specialist services or patchy connectivity. Programs should offer simple instructions, offline data storage where possible and support for carers. Involving people with Parkinson’s in design can improve comfort, adherence and relevance; community participation should shape the research from the beginning.
Privacy, safety and regulation
Movement data can become sensitive health information when it is linked to an identifiable person. Australian projects need clear consent, secure storage and transparent explanations of who can access the data, how long it will be retained and whether it will be used for future research. The Privacy Act 1988 and the Australian Privacy Principles are important considerations, alongside Queensland health-service policies and institutional ethics approval.
If software makes diagnostic or treatment recommendations, developers may need to consider whether it falls within the Therapeutic Goods Administration’s medical device framework. A risk alert should support professional review rather than trigger a frightening message without context. Services should also plan for false positives, missed events, device failure and the possibility that a person stops wearing the sensor.
Moving from pilot studies to routine support
A successful program needs a clear clinical pathway. A sensor might flag repeated freezing, prompting a physiotherapy review, medication discussion or home safety assessment. The response should be documented, timely and proportionate, with escalation arrangements for serious concerns. Without this pathway, monitoring can generate information without improving outcomes.
Implementation teams can use established digital health and research expertise to test whether a system fits Australian workflows; a digital health platform may assist with broader questions about data access and service integration. Evaluation should measure falls, near-falls, hospital presentations, confidence, walking participation and patient burden, rather than relying on technical accuracy alone.
Wearable gait monitoring is most valuable when it helps people remain active and independent. Researchers, clinicians and health services can advance this work by co-designing studies with people living with Parkinson’s disease, testing systems in Australian communities and publishing practical results. Queensland organisations interested in translating movement research into safer care can connect through Brisbane Diamantina Health Partners and develop evidence that benefits patients, families and carers.
