Wandering Risk Assessment for Clients With Dementia at Home

At 6:40 a.m., the Bed Was Empty

A caregiver called the coordinator to report that a client's bed was empty and the back door was unlocked. He was found forty minutes later, two blocks away, in his pajamas, disoriented and cold. The care plan on file described his diagnosis as "moderate dementia" and nothing else — no note on exit-seeking behavior, no record of triggers, no documented plan for what overnight staff should watch for or do differently. When the family asked what had been done to prevent this, there was no record to answer with, and no dated history showing when the plan had last been reviewed.

That gap is what a wander risk assessment exists to close — not just to reduce the chance of an incident, but to leave a dated, specific record of what the plan said, what mitigations were in place, and when they were last checked. By the end of this piece, you'll have a structured way to identify wander and exit risk, score it consistently, and turn it into a documented, versioned section of the care plan rather than a diagnosis label with nothing behind it.

Why Wander Risk Needs Its Own Line in the Plan

Dementia is common enough, and growing common enough, that treating wander risk as an afterthought inside a general safety note is no longer defensible. The Alzheimer's Association estimates that 7.2 million Americans age 65 and older were living with Alzheimer's in 2025, with prevalence rising sharply by age — about 5% at ages 65–74, 13.2% at 75–84, and 33.4% at 85 and older. Alzheimer's deaths increased 142% between 2000 and 2022, per the same source. Those numbers describe scale, not any specific wandering incidence — there's no reliable published figure to cite for how often clients with dementia attempt to leave a home unsupervised, so that part of the picture stays qualitative. What's not in question clinically is that exit-seeking, disorientation, and nighttime confusion are widely recognized features of dementia progression, and a plan that only names the diagnosis gives a caregiver nothing to act on at 2 a.m.

A general home safety checklist and a falls risk assessment cover hazards and mobility. Wandering is a different kind of risk — behavioral and situational, not just physical — and it deserves its own documented assessment inside the broader home safety evaluation.

What to Look For: Behavioral, Environmental, and Historical Signals

A useful wander assessment pulls from three categories of observation, each dated and attributed to whoever made the observation:

Behavioral signals — increased confusion or agitation in the late afternoon and evening, repeated statements about needing to "go home" or "go to work," pacing near doors or windows, checking locks repeatedly, or attempts to leave that were caught before they escalated.

Environmental factors — how many exits the home has and whether they're secured, whether door or window alarms exist, whether the client has access to car keys, and whether the layout makes an exit easy to reach unnoticed.

Historical signals — any prior episode of leaving unsupervised, getting lost while walking or driving, or disorientation in a previously familiar place. A single past episode is the strongest predictor caregivers usually have, and it should never be missing from the record just because the family didn't think to mention it at intake.

None of this is a proprietary clinical instrument, and it isn't a substitute for a physician's or care manager's clinical judgment about a client's cognitive status. It's a structured way to make sure the same three categories get checked every time, by everyone on the team, instead of depending on which caregiver happens to notice something and remember to say so.

Scoring Wander Risk: A Worked Example

Scoring gives the assessment a number that can be tracked over time instead of a paragraph that gets reread differently by every caregiver. Here is a simple structure, offered as a worked example of the method rather than a validated clinical scale:

Score each of the three domains from 0 (no indicators observed) to 2 (clear, repeated indicators):

• Behavioral signals: 0–2
• Environmental exposure: 0–2
• Historical episodes: 0–2

Total score 0–6. A commonly used tiering for a scale like this: 0–1 as lower risk, 2–4 as moderate risk requiring active mitigation, 5–6 as high risk requiring immediate environmental and supervision changes. If a client scored a 1 on behavior, a 2 on environment (multiple unsecured exits, no alarms), and a 2 on history (one prior elopement), the total is 5 — high risk — even though the behavioral signals alone looked mild. That's the value of scoring across all three domains instead of relying on gut impression from the most recent shift.

Mitigations to Document in the Plan

Once a score is assigned, the plan needs specific, checkable mitigations tied to it — not a general note to "monitor closely." Common entries include: door and window alarms on exits identified as high-risk, a secured or supervised exit at night, an ID bracelet or a location device the family has arranged, enrollment in a local wandering-registry program if one is available in the area, a documented supervision schedule for the specific hours the client is most likely to attempt an exit, and a clear, current emergency contact tree so whoever is on shift knows exactly who to call, in what order, within minutes rather than after searching for a number.

Each mitigation should be dated to when it was put in place and who is responsible for maintaining it — the caregiver checking the door alarm at shift start, the family member responsible for keeping a location device charged. A mitigation nobody owns is a mitigation that quietly stops happening.

Why a Timestamped Version Matters Here Specifically

The family didn't ask what the plan said today. They asked what it said the week before the incident.

That's the moment wander risk documentation earns its keep. A plan that's been edited in place, with the old wording gone, can't answer that question. A plan that produces a dated, immutable snapshot every time it's finalized can. CareWorkbook builds this in structurally: when a plan is finalized, it's saved as a dated version in a portable .cwbplan file, and earlier versions stay intact rather than being overwritten. Because the plan data itself is never stored on CareWorkbook's servers — the architecture is built so client and patient information stays with the agency and the family, not in a vendor's database — there's no server-side record for a licensing reviewer to subpoena from a third party, and no Business Associate question to sort out with a vendor. That's a fact about how the product is built, not legal advice; confirm your own documentation obligations with your state licensing authority.

Keeping the Assessment Current as Dementia Progresses

Wander risk isn't static, and a plan reviewed once at intake and never again is close to useless for this specific risk. Review cadence should generally be shorter than a standard annual or semi-annual check — many agencies find a 30- or 60-day cycle more appropriate once any wander indicators are present, tightening further after any incident. Review-cycle reminders that use non-identifying labels, and an org-wide dashboard showing every client's next review date at a glance, make it realistic to actually keep to that cadence across a full caseload rather than relying on memory.

For a fuller picture of how a wander risk section fits into the rest of a client's documentation, see the complete home care plan guide. And if you'd rather start from a structured template than a blank page, the Home Safety Assessment Pack includes a scored wander and exit-risk worksheet alongside the falls and hazard sections, ready to adapt to your own intake process.