Night rounds without waking anyone

The physical night round is one of the oldest practices in residential care, and one of the least examined. A staff member walks the building, opens doors quietly, checks that each resident is breathing and settled, and moves on. It is well intentioned and it has an obvious cost: it wakes people.

Sleep in older adults is already lighter and more fragmented. A door opening at one, three and five in the morning is not a neutral event, and the cumulative effect on rest — and therefore on daytime confusion, falls and mood — is not trivial. Passive sensing offers a way to get most of the reassurance without most of the disturbance.

What a night round is actually checking

Worth being precise, because the sensors need to answer the same questions.

Is the resident in bed? A bed occupancy sensor answers this continuously and more reliably than a glance in the dark.

Did they get up and not come back? The single most valuable overnight signal, and the one a three-hourly round is most likely to miss entirely — someone can be on the floor for two and a half hours between checks.

Are they moving around unusually? Repeated night-time activity is clinically meaningful and worth recording, whether it indicates pain, infection, medication effects or distress.

Have they left the apartment? A door contact answers this instantly.

Is the room a safe temperature? A sensor answers this better than a person who has just walked in from a warm corridor.

Notice that sensors answer four of those five better than a physical check, and continuously rather than three times a night. The one thing a round does that sensing cannot is see the resident — which matters, and is the reason rounds do not disappear entirely.

The sensor set for a bedroom

• Bed occupancy sensor. A pressure strip or mat under the mattress. Unobtrusive, nothing worn, no charging, and it is the foundation of everything else here.
• A single motion or presence sensor in the room, to distinguish “out of bed and moving about” from “out of bed and no longer detected”.
• Door contact on the apartment door.
• Temperature sensor.
• A call button within reach of the bed, on the regulated call system.

What we would not fit is a camera. There is no version of a bedroom camera in a care setting that is worth the loss of dignity, and radar or pressure sensing answers the clinical question without producing an image. If a supplier proposes bedroom video, ask them to justify it against a bed sensor and a presence sensor, and be sceptical of the answer.

Designing the overnight rules

Out of bed too long

The core rule. If the bed is unoccupied for longer than that resident's normal pattern — and the room shows no movement — alert. Two refinements matter enormously. Use a per-resident threshold, because someone who reliably reads for forty minutes at two in the morning will otherwise generate an alert every night. And require the absence of movement as well as the absence of bed occupancy, so a resident sitting in their chair does not trigger a search.

Left the apartment

Door open overnight, with no return within a few minutes, is worth an immediate alert. Not a siren in the building — a notification to staff, so the response is a person arriving calmly rather than an alarm frightening a disoriented resident.

Unusual restlessness

Repeated bed exits, or many hours out of bed across a night, may not need a response at three in the morning but should appear in the handover. This is genuinely useful clinical information that a physical round records only as an impression.

Never got into bed

If the bed has not been occupied by, say, two in the morning, something is unusual. Whether that warrants a check depends on the resident, which is exactly why per-resident configuration matters.

The reasonable version of “fewer rounds”

The strongest case is not eliminating rounds but making them responsive rather than scheduled. Instead of checking every resident three times a night regardless, staff check the residents the system flags, plus a reduced schedule for those whose care plan requires it.

The benefits compound. Residents who are settled are not woken. Staff time moves from walking corridors to responding to actual signals. And the resident who does need help is found in minutes rather than at the next scheduled check. That last point is the real argument: continuous monitoring with a fast response is clinically better than intermittent observation, and it is worth making that case explicitly rather than presenting the change as a staffing saving.

Getting it accepted

Two groups need convincing, and they need different arguments.

Residents and families. Lead with the sleep. “Nobody will open your door at three in the morning unless there is a reason” is a proposition most people welcome immediately. Be explicit about what is sensed and what is not — no cameras, no microphones, a mat under the mattress and a sensor on the wall.

Night staff. Their concern is usually that the technology will be used to cut staffing while responsibility stays with them. That concern is legitimate and should be answered honestly. Involve them in setting thresholds; they know which residents are up at night and why, and they will spot unworkable rules in minutes.

Practical cautions

Allow a month of tuning per resident. Night patterns are highly individual and the first fortnight will produce alerts that teach you the pattern rather than indicating problems.

Watch for the mattress problem. Pressure sensors can be defeated by very thick mattresses, mattress toppers or air mattresses. Test each bed rather than assuming, and re-test if bedding changes.

Handle empty apartments. A resident in hospital generates a night of alarming silence. Have a documented way to suspend monitoring for an absence, and to reinstate it, so nobody is chasing a resident who is on a ward.

Keep the data proportionate. Bed occupancy over time is intimate information. Restrict access to those who need it clinically, set a retention period, and resist the temptation to keep it indefinitely because it might be interesting.

An implementation that will hold up

1. Pick one wing and one month. Fit bed sensors, a presence sensor, door contacts and temperature. Log everything, change nothing, run rounds as normal.
2. Compare the log with the round records. You will find events the rounds missed and quiet nights the rounds interrupted. That comparison is your evidence.
3. Set per-resident thresholds with night staff in the room.
4. Run alerts alongside unchanged rounds for a further month, and measure how often each found something first.
5. Only then propose a change to care plans, per resident, through whatever clinical governance process applies.

That is slower than a supplier will suggest and it is the version that survives contact with a regulator, a family and a night shift.