Alzheimer’s wandering prevention: why GPS alone is not enough in a care home

Introduction

The prevention of uncontrolled wandering has become, in a few years, one of the most sensitive topics for care home directors. Cognitive disorders evolve, resident profiles are becoming more complex, and care teams must balance safety, dignity, freedom of movement and the facility’s legal responsibility.

In this context, a widespread assumption circulates in the sector: “it’s enough to equip at-risk residents with a GPS wristband to solve the problem.” On paper, the promise is attractive. In practice, directors who have deployed a 100 % GPS setup almost always run into the same limits — degraded indoor accuracy, insufficient battery life, refusal to wear the device, delayed alert notifications.

This article offers a technical and operational analysis of these limits, without dramatisation, and details the complementary building blocks that a facility benefits from combining to build a genuinely effective wandering prevention system — one that protects the residents concerned without turning their living environment into a surveillance perimeter.

Why wandering is a specific issue in a care home

Pathological wandering is not a marginal behaviour. According to estimates published by reference associations on Alzheimer’s disease (France Alzheimer, Ligue Alzheimer asbl in Belgium), around 60 % of people living with Alzheimer’s disease or a major neurocognitive disorder are concerned, at some point in the progression, by episodes of wandering or spatial disorientation. The orders of magnitude referenced in the specialist literature are as follows:

• The prevalence of major neurocognitive disorders in the population over 85 is around 20 to 30 %.
• In a Belgian care and nursing home, the proportion of residents with significant cognitive impairment frequently exceeds 50 to 70 % of the population accommodated.
• Uncontrolled departures (“unaccompanied exits by disoriented residents”) are not all serious events, but each episode mobilises the team, delays other care and can engage the facility’s responsibility.

Care home directors therefore reason not in terms of isolated incidents but in terms of a permanent at-risk population, with profiles that evolve week by week as pathologies progress.

This is what makes the technology question delicate. A single, single-component solution applied uniformly to all residents concerned cannot respond to this heterogeneity — neither clinically, ethically nor practically. GPS alone is the most visible illustration.

The 4 fundamental limits of GPS alone

Insufficient indoor accuracy

The Global Positioning System was designed for outdoor navigation. Its operation relies on receiving signals from a satellite constellation — signals that pass very poorly through building structures.

In a typical care home, the GPS accuracy observed inside the building commonly ranges from 10 to 50 metres, and can degenerate to complete signal loss. Several factors come into play:

• load-bearing concrete walls and reinforced floors,
• technical shafts and stairwells,
• insulated roofs with metal vapour barriers,
• L- or U-shaped buildings that create satellite shadow zones,
• electromagnetic interference from lifts and technical rooms.

Concretely, this means that a purely GPS-based device cannot reliably locate a resident on a specific floor, corridor or room. It indicates at best “somewhere in the building” — information hardly actionable for a care team seeking to intervene quickly. Accuracy returns to acceptable levels as soon as the resident has crossed the entrance and is outdoors. In other words, GPS effectively detects the outcome of an uncontrolled departure, but not its onset.

Battery life of wristbands

Satellite geolocation is energy-intensive. A GPS module attempting to obtain a position fix queries the constellation, which consumes significantly more than a short-range radio transmitter.

The GPS wristbands currently available on the market show, in real use, a battery life of between 24 and 48 hours, sometimes 72 hours on the most recent models in power-saving mode. In operational practice, this implies:

• a daily recharge rotation to organise, resident by resident,
• a non-negligible risk of discharged wristbands at the time of a departure,
• additional workload for care teams, who must integrate battery-level checking into their rounds,
• dependence on procedural adherence — a wristband forgotten on its charger protects zero residents.

By way of comparison, low-power radio technologies (Bluetooth Low Energy, LoRa, active RFID) last several months on a battery, which radically changes the operational equation.

Acceptance by the resident

This is the most often underestimated limit when buying a GPS device. The wristband, however discreet, remains a constraining, visible object, sometimes perceived as a restraint. The clinical profiles concerned by wandering are precisely those who:

• no longer fully understand the function of the object,
• may associate it with an uncomfortable foreign object,
• regularly try to remove, chew or take it off,
• tolerate so-called “tamper-proof” wristbands poorly, experiencing them as a form of restraint.

Legally, the wearing of a geolocation device by a resident falls within a logic of consent and, failing that, of rigorous clinical and ethical oversight. Practically, a wristband that is not worn protects no one. A device anchored in the resident’s environment (beacon in the room, floor-exit gateway, presence detection) often offers much better acceptance, because it does not materialise any physical constraint on the body.

Alert notification latency

A GPS wristband, to save its battery, does not transmit its position continuously. Depending on the model, position refresh ranges from 30 seconds to several minutes. Added to this are:

• transmission delay via the GSM network (which may be degraded in rural areas),
• processing of the position by the supervision platform,
• alert routing to the care team, sometimes by SMS or email,
• the human time for an available carer to become aware.

On a realistic cumulative basis, a “perimeter breach” GPS alert can take 2 to 5 minutes to concretely reach the person capable of acting. During this time, a resident walking at 4 km/h has already covered 130 to 330 metres. Latency, in a wandering prevention system, is not a secondary metric: it is the central metric.

Useful complementary technologies

None of these limits disqualifies GPS. Simply, GPS should be thought of as the outer edge of a broader system whose backbone plays out inside the building.

Bluetooth Low Energy beacons (BLE beacons). These small battery-powered beacons are deployed at regular intervals in corridors, vestibules, living areas and sensitive rooms. A transmitter worn by the resident (wristband, pendant, sewn-in badge) makes it possible to reconstruct an indoor path with accuracy of around 2 to 5 metres, without calling on GPS. The wearer’s battery life then becomes compatible with months of use.

RFID gateways or zone-crossing detection. At each critical door (night exit from a secure unit, emergency stairwell, car park access), a gateway detects the passage of a tag worn by the resident. The “zone crossing” event is generated instantly, before the resident has even reached the outside. This is the building block that turns late detection into early detection.

AI cameras with automatic blurring. Several recent solutions allow local video analysis without nominative recording: presence detection in an exit vestibule, fall detection, alert on crossing of a defined zone. Native face blurring and limited stream retention align the video block with GDPR requirements and Data Protection Authority recommendations. The use must remain proportionate, disclosed and documented in the processing register.

Connected door locking. Sensitive access doors (unit exits, outdoor vestibule) can be equipped with smart locks that automatically lock outside authorised time slots, or that trigger temporary locking when an at-risk resident is detected nearby. This block must of course respect fire and life-safety regulatory constraints.

Motion detection and environmental sensors. Infrared presence sensors in the rooms of at-risk residents allow detection of unusual night-time rising, upstream of any movement towards a door. Coupled with supervision, this information steers the night team before an event begins.

Alert buttons for carers. Ground-level carers must be able to trigger a general alert themselves in case of visual detection of a disoriented resident. A mobile nurse call button connected to supervision turns every carer into an active sensor in the system.

The combination of these building blocks produces a detection mesh where GPS alone produces only an exit point.

The indispensable human layer

No technical architecture replaces clinical knowledge of profiles and team presence. An effective wandering prevention system always rests on a structured human layer.

Ongoing team training. Carers must be trained in recognising precursor signs of wandering (evening agitation, repeated requests to “go home”, temporal disorientation). This clinical reading remains the first line of detection.

Written, tested and revised protocols. Every facility benefits from formalising: who is alerted first, within what time, what actions are taken in the first 5 minutes, when management is informed, when the police are contacted. Regular scenario exercises (quarterly simulation) move these protocols from the binder to the reflex.

Reinforced night organisation. Wandering events frequently occur outside high-staffing slots. Night staffing ratios, structured rounds and technical coverage must be thought through together.

Link with family referents. The family referents designated for each resident remain essential contacts: they know habits, old routes, meaningful places. A clear contact protocol with these referents is part and parcel of the system, without the care home having to play a technology-interface role with them.

Internal ethics committee. The question of using geolocation, video analysis or door locking is as much an ethical question as a technical one. A multidisciplinary committee (management, coordinating doctor, care coordinator, quality referent) that validates uses by resident profile secures the facility legally and preserves the dignity of those accommodated.

Example of a “80-bed care home” wandering prevention stack

By way of illustration, here is an example of a typical architecture for an 80-bed facility where a 20-bed secure unit accommodates residents with major cognitive disorders.

Indoor detection layer

• 25 to 35 BLE beacons deployed in corridors, vestibules and common areas of the secure unit.
• 4 to 6 RFID gateways on sensitive passage points (unit exit, stairwell, vestibule to the therapeutic garden).
• Infrared presence sensors in the rooms of residents identified as high-risk.

Outdoor detection layer

• 1 compact GPS transmitter per resident going on an accompanied walk outside the enclosure, returned on return (loan logic rather than permanent wearing).
• Geofencing around the building and the garden.

Video layer

• 2 to 4 AI cameras with automatic blurring at the main exit points, without nominative recording.
• GDPR-compliant configuration, up-to-date processing register, posted information.

Passive security layer

• Connected locks on external access doors and secure unit exits, tied to time slots and real-time events.

Human and software layer

• Supervision platform centralising all notifications (nurse call, BLE beacon, gateway, video analysis, GPS geofencing).
• Mobile terminals (professional rugged smartphones or DECT) for carers, with alert acknowledgement and traceability.
• Written protocols, quarterly exercises, annual audit of the system.

This architecture is not the only valid one. It is a realistic order of magnitude, to be adapted to the building configuration, the resident typology and the human resources available.

What Healthcall offers in wandering prevention

Healthcall is a Belgian B2B nurse call software publisher, active since 2017 with around twenty care homes (21 facilities equipped to date, from 30 to 150 beds, for a total volume of approximately 1,250 supervised residents). The company is based in Frasnes-lez-Anvaing, self-funded, and provides 4 updates per year of its premium hosting platform.

On the wandering prevention scope, Healthcall offers a module integrated into the nurse call solution, rather than an additional standalone block to operate. Concretely:

• centralised supervision of presence events, zone crossings and alerts,
• open architecture enabling interfacing with various suppliers of BLE beacons, RFID gateways, AI cameras and connected locks,
• triple database storage compliant with GDPR, controlled hosting, documented processing register,
• pricing entry from 50 € excl. VAT per call button, consistent with multi-floor deployments,
• operational stability reflected in one single client departure in 8 years of activity.

The aim is not to sell another wristband, but to equip the supervision layer on which the detection building blocks already in place, or to come, in the facility rely.

Checklist to evaluate your current wandering prevention system

To quickly audit the maturity of the system in place in your facility, here are ten points to review.

1. Do you have an up-to-date mapping of residents concerned by a cognitive disorder with wandering risk, reviewed at least quarterly by the clinical team?
2. Does the detection system cover the inside of the building (BLE beacons, gateways, sensors) or does it rely solely on a GPS exit notification?
3. What is the observed latency between a sensitive zone crossing and the effective receipt of the alert on the terminal of an available carer?
4. Do the wristbands or transmitters used offer battery life compatible with daily use without service interruption (organised recharge, backup batteries)?
5. Is the wearing acceptance rate tracked? How many residents regularly remove or refuse their device, and what alternative is offered to them?
6. Are the sensitive exit points of the building equipped with passive security (connected lock, time-slot control) compatible with fire safety constraints?
7. Do any cameras present comply with GDPR (blurring, purpose, retention period, posted information) and are they entered in the processing register?
8. Is there a written protocol for managing a wandering episode, known to all teams, tested in simulation at least once a year?
9. Do the night teams have the same supervision and alert acknowledgement means as the day teams?
10. Are wandering-prevention events logged, analysed and presented in quality review to feed continuous improvement of the system?

A score below 7 validated points signals, in the vast majority of cases, an architecture undersized relative to the facility’s real risk level.

Conclusion

The prevention of uncontrolled wandering is not solved with a GPS wristband. It is built with a coherent system that layers indoor detection, outdoor detection, passive security, proportionate video analysis and — above all — a trained human team equipped with reliable supervision tools.

GPS retains its place, but in its proper place: one building block among others, useful in the context of accompanied walks or as the last link after a perimeter breach.

If you wish to assess the maturity of your facility’s wandering prevention system and identify the priority building blocks to reinforce, the Healthcall teams are available for a technical conversation based on your care home plan and the clinical profile of the population accommodated. An initial no-commitment audit makes it possible to map the existing detection zones, identify blind spots and cost out an upgrade trajectory compatible with your budget constraints.

Get in touch to arrange a scoping meeting.