Every year, thousands of older adults across both Australia and New Zealand experience falls that lead to hospitalisation. Many live alone, and the longer they remain on the floor without help, the greater the chance of complications such as hypothermia, dehydration, pressure injuries, or even death. While pendant alarms and help buttons provide reassurance, the challenge is that after a fall, the person may not be able to press the button at all, because they’ve lost consciousness, are in shock, or are just physically unable to reach it.
This is where the technology behind fall detection algorithms becomes critical.
Why Robust Algorithms Make a Difference
A fall detector is only as good as its ability to tell the difference between an actual fall and everyday movement. If the algorithm is too sensitive, false alarms can frustrate users and erode trust in the device. If it’s not sensitive enough, the risk can be disastrous as a fall may go undetected and the person doesn’t get the help they need.
For older adults and their families, trust is everything. They want the confidence that if the worst happens, help will come. When this technology works as it should, it quietly ensures that support is alerted quickly, without the wearer having to do a thing.
Older adults across Australia and New Zealand want to age at home, surrounded by familiar routines and communities. The better fall detection systems perform, the longer older adults can maintain this independence. Families then feel reassured, and health systems benefit from fewer prolonged hospital stays linked to undetected falls.
That’s why we’ve been focused on improving our falls detection technology.
Improved Accuracy of Falls Detection
It’s been a challenging journey. Simulating a fall to match an unintentional real-life fall is tricky. Fall detection devices are usually tested using young people pretending to fall. Whilst simulations do provide useful data, young people just seem to fall differently to older adults and, no matter how good an actor the test subject is, an unintentional fall has a lack of control that is near on impossible to fake. Real life testing is also limited, quite rightly, by the ethical considerations in getting older people to participate in a study where they pretend to fall.
Accuracy of the data about a fall also depends on testing a broad range of activities being performed when a fall occurs. Most studies only test a person falling from standing upright, rather than the many smaller falls disabled or older people might have, like losing balance getting out of a chair or when bending down.
This means that much of what we know about fall detection is extrapolated from data that isn’t as real life as we’d like it. Luckily, a study from the University of Antioquia in Colombia tested things in a different way, capturing more accurate data on falls and movement detection than any other study before. In their testing, they used a larger number of participants, including older adults (albeit with strict parameters), and studied a broader range of daily activities and types of falls for more realistic results.
Armed with this more specific data we were able to improve the fall detection software in the GO G4, focusing on both the sensitivity of the device in alerting a fall and the accuracy of the device in detecting a real fall.
The results from our improvements are exciting! The GO G4 is now twice as sensitive at detecting when a fall happens, giving our users greater confidence their GO G4 device will more reliably activate after a fall.
In our practical trials, we found a considerable improvement in the GO G4’s reliability for detecting an actual fall meaning a much lower false activation rate. This has a significant impact for our customers, reducing activation demands on monitoring staff and reducing unnecessary workload.
An unexpected result of our testing was an 89% increase in the accuracy of detecting those smaller falls, like falling from the knees or from a chair. While these types of falls are common, without useful data it has been harder to achieve the right sensitivity within a device. This increase in accuracy gives users peace of mind their GO G4 is reliable for a broader range of fall situations.
Making a Difference with Smarter Falls Detection
Overall, we are delighted with the results. The new version of software in GO G4 is now much more reliable than its previous version. It is significantly better at detecting when a fall happens, with much fewer false activations, and it is also far more likely to accurately detect smaller falls.
This is a win for both our customers and for the people in the community who rely on the support from these devices. We continue to be committed to supporting healthier and safer communities, as well as reducing the financial burden on our health systems.
You may also like
Summer Safety for Seniors: How to Protect Loved Ones from Heat Risks
Summer Safety for Seniors: How to Protect Loved Ones from Heat Risks Summer across Australia and New Zealand is a season we all look forward
Aged Care Reform: Why Technology Must Be at the Heart of Change
Aged care at a turning point: why technology-enabled care must sit at the heart of New Zealand’s reform by: Abby Moore, CEO, Chiptech Coming out
#ChipFest25
#ChipFest25 We had the privilege of hosting some of our key Australian and New Zealand customers at our very first #ChipFest event in October. It