Wearable vs Camera Fall Detection

Two very different promises

A wearable detects the body it is attached to. A camera detects the space it can see. Those sound like variations on the same idea, and they behave nothing alike. Almost every disappointing fall-technology project we hear about chose the wrong one for the setting rather than the wrong vendor.

The right question is not which is better. It is where your incidents happen and what you can rely on at that moment.

What a wearable does well

A good pendant or watch alarm is the most direct way to summon help when someone is conscious, alone and can press a button. It is cheap per resident, it travels with the person, and it works in the bathroom and the garden, where fixed sensors often do not.

• Works anywhere the resident goes, including outside the building.
• Direct, personal and easy for a resident to understand.
• No camera and no coverage map to design.

Where a wearable fails

The failure mode is behavioural, not technical. Devices are removed for showers, left on chargers, forgotten in a drawer, or simply refused by a resident who does not want to wear an alarm. A device in a drawer detects nothing, and a home usually finds out at the worst possible moment.

Automatic detection on a wearable also depends on the wearer responding to a confirmation prompt. A resident who is unconscious, confused or simply asleep cannot confirm, which is exactly the case where detection matters most.

What a camera-based system does well

A zone-based camera system watches a fixed area rather than a person, so nothing has to be worn and nothing has to be confirmed. It can flag a fall-like event in a corridor, a lounge or a washroom zone and alert staff, and it produces a timestamped record of the event as a side effect.

Where the system is built on skeletal detection rather than facial recognition, and processes at the edge, the resident can be monitored without a video recording leaving the building. We describe that design in more detail in privacy-preserving vision in aged care.

Where a camera-based system fails

Coverage is the weakness. A camera sees what it is pointed at, and no home wants a camera in every private space. Occlusion, low light and an unusual posture can also produce a false alert or a miss, which is why the alerting and acknowledgement design matters as much as the detector.

This is the honest limit of the technology: it is a response-time tool for covered zones, not a guarantee of safety everywhere, and a vendor who describes it otherwise is overselling it.

Combining them without duplication

The two approaches overlap less than they appear to. Wearables cover the resident who moves and can press a button. Cameras cover the shared zones where an unwitnessed fall is most likely and where nobody is wearing anything because they are asleep or undressed.

If you run both, decide which one owns which zone and which one owns the alert, so staff do not receive two alerts for one event. Start from your own incident history, not from a product brochure, as we suggest in choosing a fall detection system.

How to decide in one afternoon

Map your last twenty incidents onto a floor plan. Mark each as witnessed or unwitnessed, and note whether the resident was wearing anything at the time. The map usually answers the question on its own, and it gives you a fair basis for judging any pilot. Our FAQ covers the deployment and privacy questions that follow, and the resident-level groundwork is in our fall risk assessment guide.

A fair way to test both

Whichever way you lean, test with your own residents rather than with a demo. Ask for a month, name the zones and the residents in advance, and write down what you will measure before the trial starts.

For a wearable, the number that matters is wear time, not detection accuracy. A device that is worn forty per cent of waking hours detects forty per cent of what it could. For a camera system, the number that matters is alert volume per zone per week, together with the time from event to a staff member knowing.

Ask the same closing question of both: how long until someone knew? That single measure is comparable across the two approaches, and it is the one that maps onto a resident's night.

• Wearable trial: measure wear time first, detection second.
• Camera trial: measure alerts per zone per week, and false alerts separately.
• Both trials: measure time from event to acknowledgement, on the same zones if possible.

What the comparison usually finds

The exercise rarely produces a winner and a loser. It produces a map: this corridor and this washroom are covered by the camera; this resident, who walks to the garden and refuses to wear anything, is covered only by the pendant she does wear. The map is the procurement decision, and it is also the thing to re-check every year as your residents change.

Where the two overlap, and why that is fine

Some zones will be covered twice, and that is not waste. A resident who wears a pendant and walks a covered corridor has two chances of being found quickly, and the two systems fail for different reasons, which is precisely why redundancy helps.

What you should avoid is two alerts arriving for one event, because that is how staff learn to dismiss the alert before reading it. Decide which system owns which zone, and suppress the other where they overlap.