What a digital supervision pilot should measure

At the end of a pilot, a care team should be able to explain what changed for the people living in the rooms and the people working the night shift. A dashboard full of events cannot answer that on its own.

In August 2026, KS Digital shared municipal experiences that put clear responsibility and measurement before implementation at the centre of successful digitalisation. Those lessons apply directly to a first digital supervision pilot. Read KS Digital's account.

For Stillsense, that suggests a practical starting point: agree what would make the service better before placing the first sensor. Here is the evaluation we would bring to a pilot conversation.

Give the pilot one clear question

A useful question might be: can this ward reduce unnecessary night-time room entries while maintaining the follow-up each resident needs?

That gives the team something specific to evaluate. It also keeps the resident's individual care plan central. Some visits provide medication, reassurance or personal care; knowing someone is present in a room does not remove those needs.

Choose a defined group, a review date and a named service lead. Record the current pattern of visits, staff time and reported disturbances before changing the workflow. Include ordinary variation across shifts, rather than comparing a quiet pilot week with an unusually difficult week beforehand.

Follow five things through the night

Our suggested measures cover both the technology and the work around it:

• Room entries and disturbances. Record why staff entered and whether the resident was disturbed. Fewer entries only count as an improvement when necessary care still happens.
• The path from alert to help. Separate the time an event is detected, the time a member of staff acknowledges it and the time any required assistance arrives. A quick notification tells only part of the story.
• Alerts that require no action. Review why they occurred and how much attention they took. A correctly detected movement may still produce an unhelpful alert if the rule does not fit that resident.
• Events the system missed. Compare staff observations and incident records with the system's history. Review gaps in coverage and periods when equipment was unavailable. An empty alert log cannot establish that nothing happened.
• The resident's experience. Ask whether the arrangement feels understandable and acceptable, and whether nights feel less interrupted. Include relatives where appropriate, while keeping the resident's own voice central.

Use comparable units, such as alerts per resident-night, and record changes to staffing, occupancy or alert settings. Otherwise a lower total may simply reflect fewer people or fewer nights.

Count the work that moves elsewhere

A visit avoided can release time. Reviewing alerts, adjusting equipment, training colleagues and documenting events also take time. Include those tasks in the evaluation, with initial setup separated from recurring work.

Then describe what the released time made possible. More time with a resident and a lower operating bill are different outcomes. Helsedirektoratet's national roadmap includes service quality, users' experience and increased care capacity among the intended benefits of welfare technology. See the roadmap for welfare technology.

Set the decision before the results arrive

Agree which findings would justify expanding, adjusting or stopping the pilot. Include unresolved missed events, excessive alert workload and residents' concerns in that decision. A small pilot with no recorded falls cannot establish fall-detection accuracy or prove that falls were prevented.

Stillsense offers presence, movement and bed-exit detection without cameras or wearables. Whole-home coverage depends on layout. Fall detection is available under independent validation, with no safety-grade guarantee. Those distinctions belong in the evaluation plan from the start; our evidence and trust page sets out the capability boundaries.

A good pilot leaves the service with a decision it can explain, including what still needs testing.

Book a pilot conversation to define the question, the measures and the first setting together.