Fall Detection Devices Compared: Pendants vs Watches vs Sensors

Four approaches to the same problem

Every fall detection device on the market is trying to solve the same problem: detect when an elderly person falls and get help to them quickly. But the approaches are fundamentally different, and each comes with trade-offs that manufacturers do not always make clear.

This guide compares the four main categories of fall detection devices side by side, with honest assessments of accuracy, cost, compliance rates, and real-world performance. Whether you are a family choosing for one parent or a care provider evaluating options at scale, the goal is to help you make the right decision for your specific situation.

Category 1: personal emergency response systems (PERS/pendants)

The oldest and most familiar form of fall detection. A wearable device (usually a pendant necklace or wristband) with an SOS button that connects to a 24/7 monitoring centre. Newer models include a basic accelerometer for automatic fall detection.

How detection works

Traditional models rely entirely on the wearer pressing a button. Automatic models use an accelerometer to detect sudden impacts consistent with a fall: a rapid acceleration followed by deceleration. If a potential fall is detected, the device either automatically connects to the monitoring centre or prompts the wearer to confirm/cancel the alert.

Real-world performance

• Detection accuracy (when worn): manual button is 100% (if the person can press it). Automatic detection hits 60-75% for hard falls, much lower for slow collapses or slides.
• Compliance rate: 20-30%. Studies consistently find that most falls occur when the pendant is not worn.
• False alarm rate: low for manual (the person decides). Moderate for automatic (sudden movements can trigger alerts).
• Response time: under 60 seconds to reach a monitoring centre operator (from button press or automatic detection).

Cost

£0–£150 for the device (some providers include it free). £15–£50 per month for 24/7 monitoring. Annual cost: £180–£600 per person.

Category 2: smartwatches

Consumer smartwatches with built-in fall detection, primarily the Apple Watch (Series 8+, Ultra 2+), Samsung Galaxy Watch (5+), and Google Pixel Watch (2+). These use sophisticated combinations of accelerometers, gyroscopes, and barometric pressure sensors to detect fall signatures.

How detection works

The watch continuously samples motion data from its sensors. Machine learning algorithms (trained on hundreds of thousands of real and simulated falls) identify movement patterns consistent with falling: rapid wrist acceleration, impact signature, and post-fall immobility. If a fall is detected and the wearer does not respond within 30-60 seconds, the watch calls emergency services and sends the user's GPS location.

Real-world performance

• Detection accuracy (when worn): 80-90% for hard falls (Apple Watch Series 10 performs best in independent tests). 40-60% for slow falls and collapses. Much lower accuracy for "soft" falls onto carpet or beds.
• Compliance rate: 30-40%. Better than pendants because watches are a normal accessory, but charging requirements (daily for most models) create overnight gaps.
• False alarm rate: moderate to high. Aggressive hand gestures, slamming a car door, or dropping heavily into a chair can trigger alerts. Apple has improved this since the initial release, but false positives remain common.
• Response time: 30-60 second countdown, then automatic call to emergency services. GPS location shared.

Cost

£250–£500 for the watch. £5–£15/month for a cellular plan (required for standalone calling without a nearby phone). Annual cost: £310–£680 per person, assuming the watch lasts 3–4 years.

Category 3: camera-based AI

Indoor cameras paired with artificial intelligence algorithms that analyse video feeds in real time to detect falls. Systems from various providers use computer vision to track human poses and identify sudden transitions from upright to ground-level positions.

How detection works

The camera captures continuous video. AI models trained on fall footage identify human skeletal poses (limb positions, body angle) and flag rapid changes consistent with falling, specifically a standing or sitting pose transitioning to a prone position on the floor within a short time window. Some systems process video locally (edge computing); others send it to a cloud server.

Real-world performance

• Detection accuracy (where visible): 85-95% in ideal conditions (good lighting, clear view, single person). Drops with poor lighting, occlusion by furniture, or multiple people in the frame.
• Compliance rate: not applicable (passive), but acceptance rate is extremely low. 78% of over-65s refuse cameras in their home (Age UK, 2024).
• False alarm rate: low to moderate. Algorithms occasionally mistake controlled movements (yoga, picking up objects from the floor) for falls, but visual AI is generally good at distinguishing intentional vs unintentional ground-level positions.
• Response time: seconds for detection, then requires human review and action. No automatic emergency calling in most systems.
• Coverage gap: cannot be placed in bathrooms or bedrooms for legal and ethical reasons. These are the highest-risk rooms.

Cost

£100–£300 per camera. £30–£80/month for cloud AI processing and monitoring. Multi-room coverage requires multiple cameras. Annual cost: £460–£1,260 per home (2–3 cameras). For care providers, GDPR compliance adds significant administrative overhead.

Category 4: radar sensors

Wall-mounted sensors using 60GHz FMCW radar to detect falls through radio wave reflections. No cameras, no microphones, nothing to wear. Systems like HomeCare use this technology to provide passive, privacy-preserving fall detection and activity monitoring.

How detection works

The sensor emits continuous low-power 60GHz radio waves that bounce off people and objects. By analysing the reflected signals (distance, velocity, and signal characteristics), the system builds a real-time motion profile. Machine learning algorithms trained on fall data identify specific signatures: rapid downward velocity, height transition from standing to floor level, impact pattern, and post-fall immobility. A confirmed fall triggers an automatic alert to family members or a care provider dashboard. For a detailed technical explanation, see Radar Fall Detection: How It Works and Why It Matters.

Real-world performance

• Detection accuracy: over 95% in peer-reviewed studies, including slow collapses and slides that challenge other technologies. Effective in darkness, through moisture (bathrooms), and regardless of what the person is wearing.
• Compliance rate: not applicable, fully passive. The sensor works whether the person knows about it or not (with their consent). No wearing, charging, or interaction required.
• False alarm rate: low (under 1% in controlled studies). Full-body motion analysis and post-fall immobility confirmation reduce false triggers considerably compared to wrist-based accelerometers.
• Response time: seconds from fall to alert. Automatic notification to family phones or care dashboard.
• Coverage: works in every room including bathrooms and bedrooms, the highest-risk locations.

Cost

HomeCare: €400 for hardware (covers an entire home). €29/month for monitoring, alerts, and dashboard. Annual cost: €748 per home. Other radar-based systems vary. For care providers, per-home pricing means predictable costs regardless of how many residents a home has.

The complete comparison table

*Effective detection rate = detection accuracy x compliance/coverage rate. This is the number that actually matters because it represents the real-world probability that a fall will be detected.

Recommendations by situation

There is no single best device. The right choice depends on the person, their circumstances, and the operational context.

For an active, tech-comfortable parent

Best option: Smartwatch (Apple Watch or Samsung Galaxy Watch)

If your parent already wears a watch, uses a smartphone, and is comfortable with technology, a smartwatch adds fall detection without introducing a "medical" device. The additional health features (heart rate, irregular rhythm, blood oxygen) are a bonus. The risk is overnight charging: consider a second charger in the bedroom and encourage wearing it to sleep.

For a parent who refuses to wear anything

Best option: Radar sensor

If your parent has refused a pendant, finds wearables uncomfortable, or has mild cognitive impairment that makes wearable compliance unreliable, a passive radar system is the only option that provides actual fall detection around the clock. Nothing to wear, nothing to charge, nothing to remember. Works in every room, including the bathroom at 3am. See our guide: Fall Detection Without a Pendant: Alternatives That Actually Work.

For a high-risk parent (frequent falls, frailty, cognitive decline)

Best option: Radar sensor + pendant

For the highest-risk individuals, combining a passive system with a wearable provides both automatic detection (the radar) and a proactive help button (the pendant). The radar catches falls when the pendant is not worn. The pendant gives the person agency to call for help proactively. This combination closes the most gaps.

For a care agency managing multiple homes

Best option: Radar sensors with centralised dashboard

Per-home passive monitoring scales better than per-person wearables. Advantages for agencies:

• No chasing pendant compliance across dozens of clients
• One dashboard showing all homes, with alerts, activity patterns, and response times
• Per-home pricing rather than per-person pricing
• Fast installation (15 minutes per home) allows setup during standard home visits
• GDPR simplicity: no visual data, no footage management, no data subject requests for camera recordings
• Consistent evidence base for CQC inspections and quality reporting

Systems like HomeCare offer a care provider dashboard specifically designed for multi-home management, with alerts routed to on-call staff and activity reports exportable for compliance documentation.

When budget is the primary constraint

Best option: Pendant + motion sensors

If cost is the main factor, a basic pendant alarm (from £15/month, some local authorities provide them free) supplemented with inexpensive motion sensors (£100-£200) provides a reasonable baseline. The pendant gives immediate help-button access for conscious falls. The motion sensors provide routine monitoring that can flag prolonged inactivity. This is not the best option, but it is far better than nothing, and Attendance Allowance (£72.65-£108.55/week, not means-tested) can help cover the costs.

What to consider beyond the device

Whichever device you choose, fall detection is only useful if there is a response plan behind it. Ask yourself:

• Who receives the alert? You? A monitoring centre? A neighbour? The carer on shift?
• How quickly can someone reach your parent? A 5-minute response is very different from a 45-minute drive. Local contacts are essential.
• Is there a key safe or entry plan? If your parent is on the floor and cannot open the door, how does help get in?
• Does everyone in the response chain know the plan? An alert that goes to three family members who all assume someone else will handle it is worse than an alert to one person who knows it is their responsibility.

The best fall detection device in the world is only as good as the response it triggers. Plan the response before you choose the device.

For the full breakdown of every fall detection option, including cost analysis and a decision framework, see our guide: Fall Detection for Elderly: Every Option Compared (2026).