He Wasn't Wandering. He Was Going to the Store.

Most missing incidents start with an ordinary trip in daylight, not an unlocked door overnight. That changes what actually protects someone.

Is this wandering, or did he just go out and not come back?

Most missing incidents do not begin with someone slipping quietly out a back door. They begin with a normal trip the person was allowed to take, made alone, that did not end where it was supposed to.

A 2011 analysis in BMC Geriatrics looked at 325 US cases of people with dementia reported missing in the community. The primary antecedent — the thing that happened right before — was becoming lost while doing a routine, caregiver-sanctioned activity by themselves. The walk. The errand. The drive he has made a thousand times.

That finding rearranges the problem. Pacing the hallway, rattling the door handle, the restless circuit at dusk — those are visible, and families act on them. The walk to the mailbox is invisible, because it has always worked.

So "she doesn't wander" is not the reassurance it sounds like. In a Canadian study of 560 missing incidents among MedicAlert subscribers, 91.9% of people were on foot and 67.1% went missing from their own homes. A 2026 analysis of police records in England and Wales found 78% of the people with dementia had gone missing from their own home rather than from a facility or a hospital.

Does everyone with dementia eventually wander?

The figure you will hear is that six in ten people with dementia will wander. It is repeated on nearly every caregiving site, and it does not trace back to a study.

The 60% number comes from the Alzheimer's Association rather than from a published cohort, and researchers have flagged that nobody knows what population it describes. It is not a fabrication so much as an orphan — a number with no parent.

What has actually been measured is lower and more useful. A 2011 study in Dementia and Geriatric Cognitive Disorders put the rate of missing incidents among community-dwelling people with dementia at 0.65 per year. Once it happens, it tends to happen again: 22.5% of the Canadian incidents were repeats, and up to 15% of the people in the English and Welsh police data went missing more than once.

That combination — uncommon, then recurrent, occasionally fatal — is why the standard advice fails. "Watch her more carefully" is not a plan anyone can sustain against an event that has never happened. What works is deciding in advance what you will do in the first hour of the day it does.

What actually happens when someone goes missing?

Almost all of them are found, and almost all are found alive. The frightening number that circulates — that roughly a third die — comes from newspaper reports, and newspapers do not run the story about the man who was back in his kitchen in forty minutes.

The best population-level figures available:

• Texas, 2017–2022. A cross-sectional study in JAMA Network Open analyzed statewide Silver Alert activations for missing adults with dementia. 94.7% were found.
• Canada, 2015–2021. Among 560 missing incidents involving MedicAlert subscribers, 90.6% were returned home safely. Injuries were reported in 9.2%. There was one death, 0.2%.
• England and Wales, 2015–2021. Of 414 dementia-related missing incidents recorded by two police forces, 9.4% ended in significant harm, including one recorded death.

The 30% mortality figure that gets quoted comes from that same 2011 BMC Geriatrics paper, and the authors say plainly what their sample was: 325 newspaper reports. A death gets written up. A recovery does not.

Knowing the real odds is not comfort for its own sake. It changes how you make the call. A caregiver who believes a missing person is probably already dead panics and searches badly. A caregiver who knows the person is very likely alive and very likely nearby calls 911 in the first ten minutes and tells the dispatcher something useful.

One more number from the Canadian data, because it changes who you telephone: 47.7% of people were located by first responders and 46.1% by ordinary members of the public who happened to see them. The neighbor two doors down is roughly as likely to find him as the police are.

Where do search teams actually find people?

Close. Uncomfortably close. In an analysis of 93 deaths, 87% of the people were found in natural, secluded, unpopulated places — woods, fields, ditches, bodies of water — and 73% of those who died were within half a mile of where they were last seen.

The people found alive were farther out: 87% within a five-mile radius. Put those two findings side by side and you get a cruel piece of geometry. The search expands outward, hour by hour, while the person in the most danger is inside the first half mile, in cover thick enough that a line of searchers can pass within a few feet and see nothing.

Drainage ditches. Culverts. Briars. Creek beds. The strip of trees behind a subdivision. A neighbor's shed. Retention ponds.

This is worth saying out loud when you make the call, because you know the terrain and the responding officer may not. Ask that the initial effort cover the dense vegetation and the water within half a mile before the perimeter widens.

Drivers are a different problem with different geometry. A 2012 study in the Journal of the American Geriatrics Society reviewed 156 Florida Silver Alert records for missing drivers with dementia. Only 15% were found while driving — most were found in or near a parked car. They had traveled an average of 116 miles from home. More than 70% were men, and most had gotten lost on ordinary trips their families had approved.

Do GPS trackers work?

They find people. There is no trial showing they prevent harm, and that is a different claim from the one printed on the box.

A 2024 integrative review in the American Journal of Alzheimer's Disease and Other Dementias screened 1,429 studies and found 10 worth including. The devices did give people more independence, let them go outside feeling safer, and lowered caregiver worry. What the literature does not contain is a randomized trial measuring whether anyone came to less harm.

The programs built around the same idea have not been tested either. A 2013 critical review in The Gerontologist examined Silver Alert and the other locate-the-missing-elder programs and found no independent evaluation of any of them — the only efficacy data available were media reports. Project Lifesaver, the radio-frequency bracelet run through many sheriff's offices, is in the same position: widely deployed, never independently evaluated.

None of that means the devices are useless. It means the reason to own one is narrower and more honest than the marketing suggests.

Three limits worth knowing before you spend the money:

• People take them off. A 2024 implementation study published in Dementia reported a 25% refusal rate and 33% attrition within six months. The failure mode is human, not technical.
• They need charging every day. That is a new daily task in a life that has no room for one, and a dead tracker is indistinguishable from no tracker.
• They fail where they matter most. A cellular GPS tracker needs signal. The places where people are found dead — wooded, low-lying, unpopulated, at night — are the places with the thinnest coverage. The technology is weakest exactly where the stakes are highest.

Geofence alerts have a related problem. Set the boundary tight and it fires every time he walks to the mailbox, until you stop reading the alerts. Set it loose and it tells you nothing you did not know. And no geofence is relevant to a man 116 miles down the interstate in a parked car.

So buy it for the one thing it does well. It collapses the interval between noticing and locating. Time is the variable that predicts harm, and a tracker attacks that variable directly. That is a real benefit. It is just not the same as safety.

Will an AirTag do the job?

No. An AirTag has no GPS inside it. It reports a location only when a stranger's iPhone happens to pass close enough to hear it.

Apple built the AirTag for luggage and says so. Two failure modes matter here.

The first is how it locates. An AirTag is a Bluetooth beacon riding on other people's phones. In a mall or a dense city block it may resolve in seconds. On a rural roadside at two in the morning, or in a drainage ditch behind a subdivision, nothing walks past it, and it goes silent — which is precisely the scenario the deaths come from.

The second is that Apple built in anti-stalking alerts. If the person carries an iPhone, the phone will eventually tell them that a tracker they do not own is traveling with them. The feature exists for good reasons and it works against you here.

An AirTag in a coat pocket is better than nothing in a parking lot. It is not a safety device, and treating it as one buys a family a false sense of coverage in exactly the terrain where coverage decides the outcome.

What do I do in the first hour?

Call 911 immediately. There is no 24-hour waiting period for reporting a missing person, and the first 24 hours are when nearly everyone is recovered alive.

• Call 911 first, then look. Say the words "adult with dementia," "at risk," and "cannot ask for help." Those phrases trigger a different response than "my husband is late."
• Ask about a Silver Alert. Most states have one, under various names, with different criteria and different agencies that activate it. Ask who does it and whether he qualifies.
• Hand over a recent photo and describe what he is wearing, shoes included.
• Tell them where the cover and the water are within half a mile. Woods, ditches, culverts, ponds, sheds.
• If a car is gone, give the plate, make, model and color, and say he may be far away and may be parked rather than driving.
• Search the house and the property first — closets, behind doors, vehicles, sheds, under and behind things. People are found at home more often than anyone expects.
• Call the neighbors now, not later. Roughly half of recoveries come from ordinary people who spotted someone.
• Leave somebody at the house. Some people walk home.

What is worth setting up before it ever happens?

Prepare for the day it happens instead of trying to guarantee that it never will. The evidence supports shortening the search far better than it supports preventing the exit.

Write the one page now. Current photograph, height, weight, build, the name he answers to, medical conditions, medications, and the addresses he might be heading for — the house he grew up in, the church, the plant where he worked, the lake. Keep a copy on your phone and a printed copy by the door. On the worst morning of your life you will not be able to recall his weight.

Enroll in MedicAlert Safe & Found. This is the program most people still call Safe Return; the name changed and the old one is still printed on brochures in doctors' offices. It pairs ID jewelry with a 24-hour emergency line that connects responders to the person's information and to you.

Ask the sheriff's office about Project Lifesaver. It has no published effectiveness data, and it has one real advantage over a consumer tracker: radio frequency penetrates woods and terrain where cell service dies, and the people holding the receiver are trained and local. Availability and cost vary by county.

Put a tracker in the car before you take the keys. Given how far drivers travel and how rarely they are found while still moving, a vehicle tracker is the highest-value device in this whole category, and one of the cheapest.

Use door chimes, not visual tricks. A Cochrane review of subjective barriers — mirrors on doors, camouflage, grids and stripes of tape on the floor — found no evidence that any of them prevent wandering, and could not rule out that they cause psychological harm. A twelve-dollar chime tells you a door opened. A mirror on the door tells you nothing and may frighten someone in his own hallway.

Tell the neighbors today. Not after. Give two or three of them your cell number and a sentence: if you ever see Dad out on the street alone, please call me before you do anything else. That is the cheapest and best-evidenced intervention available to you, because half the people who go missing are found by somebody exactly like them.