Short answer. Wandering in dementia is rarely aimless. It is usually goal-directed behaviour — the person is trying to do something (go to work, find a child, leave for "home," relieve boredom, escape discomfort) — filtered through disorientation that strips the goal of context (Alzheimer's Association, 2024). The research-backed approach has two parts that must go together: reduce the drive to wander by addressing the underlying unmet need (Gitlin, Kales, & Lyketsos, 2012), and reduce the risk if wandering does occur through environmental modification and layered safety measures. Restraint and locked confrontation are neither safe nor effective; redirection and prevention are.
What the research says about wandering
Wandering is one of the most anxiety-producing behaviours family caregivers face, because the worst-case outcome — an unaccompanied person with dementia leaving the home and becoming lost — is genuinely dangerous. The Alzheimer's Association (2024) notes that a large share of people with dementia will wander at some point, and that becoming lost carries real risk of injury, exposure, and death, especially when it is not detected quickly.
But the behavioural intervention literature (Gitlin, Kales, & Lyketsos, 2012) reframes wandering away from "a symptom to be contained" and toward "a behaviour that communicates something." Under the unmet-need model, most wandering is an attempt to meet a need the person can no longer articulate or satisfy directly:
Goal-directed searching — looking for a person, a place, or an object from an earlier life stage ("I have to pick up the children," "I need to get to work").
Wanting to go "home" — often not the current house but a home from decades ago, expressing a need for safety and familiarity rather than a literal address.
Restlessness and under-stimulation — pacing and exit-seeking that rise when the day is empty and the body has unspent energy.
Discomfort or unmet physical need — pain, hunger, a full bladder, or overheating expressed as agitated movement.
Environmental triggers — a coat by the door, keys in view, or watching others leave, all of which cue the over-learned act of "going out."
The clinical implication is that why the person is wandering determines what helps. A person searching for their long-dead mother and a person pacing from boredom need opposite interventions.
Reduce the drive: address the unmet need
The most durable prevention works upstream of the front door.
Decode the pattern first
Wandering usually has a time signature and a trigger. Track when it happens (late afternoon overlaps heavily with sundowning), what preceded it, and what the person says they are trying to do. The stated goal — even a confused one — is the clue. The same decoding pass used for agitation episodes applies here.
Meet the need the wandering points to
Structured daytime activity and movement. Regular walks, purposeful tasks, and outdoor time reduce the restless-energy driver more reliably than any lock. A tired body at a predictable time wanders less.
Address the "home" statement as an emotion, not a location. Arguing "you are home" almost always escalates. Acknowledging the feeling — "You miss home. Tell me about it" — and redirecting tends to defuse the exit-seeking.
Rule out physical precipitants. Pain, constipation, a full bladder, and overheating all raise agitated movement. Address body basics before assuming a behavioural cause.
Reduce the risk: environmental modification and safety layers
Because no upstream intervention is perfect, the research-backed approach layers safety so that a single lapse is not catastrophic. This environmental-modification approach is the same family of home-based, caregiver-delivered strategies tested in the REACH II intervention (Belle et al., 2006), whose home-safety component reduced risk and caregiver distress.
Modify the environment
Camouflage and de-cue exits. Removing coats, shoes, keys, and bags from view near the door reduces the visual cue to leave. Some families find a curtain over the door or a dark mat reduces exit-seeking, though evidence here is mixed and individual.
Reduce visible triggers to "go." Turning off news programmes showing people leaving, and keeping the person's own outdoor gear out of sight, lowers the cue rate.
Make the current environment satisfying. A clear, calm, well-lit space with something to do reduces the pull toward the door.
Layer safety measures
Door awareness. Alarms or chimes that signal when an exterior door opens give the caregiver a chance to intervene without a locked confrontation. Locks placed out of the usual line of sight (high or low) exploit the fact that the over-learned reach is at handle height.
Identification. A medical ID bracelet or a sewn-in garment label with a name and contact number dramatically shortens the time to reunification if the person is found by a stranger.
Location technology. GPS-enabled wearables or shoe inserts allow rapid location if the person does leave. Enrolment in a local wandering-response or safe-return programme, where available, adds a community layer.
Neighbour and community awareness. Letting trusted neighbours and nearby businesses know reduces the "no one recognised them" gap that turns a short absence into a crisis.
What does not work
Confronting the person at the door or physically blocking them. This reliably escalates agitation and can turn wandering into a fight. Redirection — walking with them, then gently circling back — is safer and more effective.
Reasoning them out of the goal."You don't work anymore," "your mother died years ago" re-traumatises and rarely stops the behaviour.
Relying on a single measure. A lock alone, an alarm alone, or an ID bracelet alone each has a failure mode. Safety comes from layers.
Physical or chemical restraint as a first response. Restraints increase agitation and injury risk and do not address the underlying drive; guidelines and the behavioural literature (Gitlin, Kales, & Lyketsos, 2012) place them well after nonpharmacologic approaches.
How families ask this
The question arrives in many phrasings, all pointing at the same fear:
"How do I stop my mom with dementia from wandering off?"
"My dad keeps trying to leave the house — what do I do?"
"He says he has to go to work and won't stop trying the door."
"She keeps saying she wants to go home even though she is home."
"I'm terrified he'll get out at night while I'm asleep."
These are not different problems. They are the same goal-directed, unmet-need behaviour presenting under different life histories — and they respond to the same two-part approach: reduce the drive, reduce the risk.
Gitlin, L. N., Kales, H. C., & Lyketsos, C. G. (2012).Nonpharmacologic management of behavioral symptoms in dementia. JAMA, 308(19), 2020–2029.
Belle, S. H., Burgio, L., Burns, R., et al. (2006).Enhancing the quality of life of dementia caregivers from different ethnic or racial groups: a randomized, controlled trial (REACH II). Annals of Internal Medicine, 145(10), 727–738.