Night shift staffing in a typical memory care unit runs at roughly half the coverage of a daytime shift — one CNA for every 12 to 15 residents, against a daytime ratio closer to 1 in 7.¹
This is not a staffing failure. It is a staffing norm. And it would be manageable — except that the hours between 10 PM and 6 AM are precisely when dementia wandering risk is at its highest.
These two facts do not operate independently. They collide, every night, in every memory care facility in the country. Most care systems have no infrastructure designed for that collision.
What Sundowning Actually Is
The term "sundowning" describes a predictable pattern of increased confusion, agitation, and restlessness that emerges in many people with dementia as daylight fades.
It is not a rare complication. It affects between 20% and 45% of people with Alzheimer's disease, according to Alzheimer's Association research.² The underlying mechanism points to progressive deterioration of the suprachiasmatic nucleus — the brain region governing circadian rhythm. As ambient light diminishes in the evening, the dementia brain loses its primary environmental anchor.
The result is not simply increased agitation. It is a specific behavioral state: heightened disorientation, a drive to move, and — critically — the loss of recognition of where "safe" is.
A resident who was calm at dinner may be attempting to leave the building by 11 PM, oriented not to the memory care unit, but to a place she needed to be forty years ago.
The Staffing Arithmetic
Federal minimum staffing requirements ensure a Registered Nurse on duty overnight. Beyond that threshold, overnight coverage in most 40–80 bed memory care units follows a predictable pattern:
- Day shift: approximately 1 CNA per 6–8 residents
- Evening shift: approximately 1 CNA per 8–10 residents
- Night shift: approximately 1 CNA per 12–15 residents¹
The same square footage. The same corridors. The same exits. Staffed, after midnight, at roughly half the human coverage of the shift that ended eight hours earlier.
Night staff are not underprepared. They are structurally outnumbered — and they know it.
The Moment When Both Curves Cross
Between 10 PM and 2 AM, these two realities intersect.
Sundowning-driven agitation is near its peak. A resident who has been restless since late afternoon has now been awake for hours, growing more disoriented, more persistent, more directed toward the door.
The night-shift CNA has 12 residents. She is performing scheduled checks, administering medications, responding to call lights. She is not neglecting the wandering resident. She simply cannot provide continuous observation across a 12-person assignment while also managing everything else the shift demands.
The wandering resident is not announcing her departure. She has learned, through repetition, which door is less alarming, which corridor leads toward the feeling she is chasing. She is quiet. She is purposeful. And she is moving.
The alert — if it fires at all — fires after the door closes.
Why Exit-Triggered Systems Fail at This Hour
Most detection systems in current use share a design assumption: alert on departure.
Door alarms, threshold sensors, perimeter monitors — all trigger at exit. The logic is intuitive: if someone is leaving, alert.
But in a minimum-staffing overnight environment, an exit alert does not solve the problem. It announces that the problem has already occurred.
The one CNA on duty must now leave her remaining 11 residents unsecured to respond. The 15-minute SAR window — the critical threshold after which the probability of finding a cognitively impaired person in safe condition begins to decline sharply — starts counting the moment the door closes.³
Exit-triggered systems were designed to slow wandering. They were not designed for a staffing environment where the first responder is also the only responder, managing a full assignment at midnight.
Alert Fatigue Makes It Worse
Overnight staff in memory care facilities already operate under significant alarm burden. Call lights, bed sensors, fall-prevention monitors — every resident-facing alert routes to a floor with skeleton coverage.
The consequence is well-documented: when alert volume is high and false positives are frequent, staff become desensitized. The alarm that fires twenty times without consequence will eventually not receive an urgent response.
This is not a human failing. It is what happens when systems designed for daytime environments are deployed unmodified into the overnight context.
A night-shift CNA does not need more alerts. She needs fewer, better ones — signals that distinguish genuine escalation from routine nighttime movement, early enough to act before departure, specific enough to cut through the noise.
What Earlier Detection Requires
The facilities that perform best on overnight wandering outcomes share one structural feature: they have moved the detection window upstream of the exit.
Rather than alerting on departure, their systems identify behavioral escalation — changes in movement pattern, restlessness duration, agitation signals — before the resident reaches the door.
This is not continuous video surveillance. It is pattern intelligence: learning what normal looks like for each resident and flagging meaningful deviation from that baseline. A resident who typically sleeps through until 4 AM but has been active and moving since 11 PM is exhibiting a signal. That signal is actionable — but only if someone receives it before she exits.
Pre-departure detection solves two problems simultaneously. It gives night staff a longer response window when coverage is at its thinnest. And it creates a documented record of the intervention — which matters considerably in the event of a surveyor review.
The Compliance Dimension
CMS F-tag 689 applies to accident hazard reduction, and elopement is explicitly covered. After an overnight wandering incident, a surveyor's inquiry will not stop at policy documentation. It will ask: what did your system tell staff before the resident left the building?
A door alarm timestamp reading 11:47 PM does not demonstrate a prepared response posture. It demonstrates a reactive one.
A behavioral escalation alert at 11:23 PM — logged, acknowledged, and acted upon — is a fundamentally different record. It shows that the care environment was built to detect risk before crisis, not merely to document crisis after the fact.
The Night Shift Is Not the Problem. The Gap Is.
Memory care overnight staffing is not going to double. The demographic pressure on the long-term care labor market runs in the opposite direction.
What can change is the intelligence layer between the resident and the responder. Not more alerts — the floor cannot absorb more volume at 1 AM. Earlier alerts. Smarter prioritization. Context-aware escalation that reaches the right person at the moment when intervention is still possible.
The collision between peak wandering risk and minimum staffing is structural. But structural problems have structural solutions.
The night shift does not need to be redefined. It needs to be better equipped for the hours it already owns.
Sources
¹ Centers for Medicare & Medicaid Services. Nursing Home Staffing Study. cms.gov
² Alzheimer's Association. Sundowning. alz.org
³ Koester, R.J. Lost Person Behavior: A Search and Rescue Guide. dbs-sar.com, 2008; International Search and Rescue Incident Database (ISRID).