Why check-in procedures fail
Almost every organisation with lone workers has some check-in arrangement, and most of them quietly don't work. The common failure isn't technology — it's design. Intervals are set once, uniformly, without reference to risk, so they feel arbitrary. Missed check-ins go to an inbox nobody watches after 5pm. Workers learn that nothing happens when they forget, so forgetting becomes normal. Then an incident happens, and the procedure that existed on paper turns out to have stopped operating months earlier.
A check-in procedure is a promise running in both directions: the worker promises to confirm they're safe on a schedule, and the organisation promises that a missed confirmation triggers a real response, every time, at any hour. Design either half loosely and the other half decays. The rest of this guide takes the two halves in turn, then the harder problem — making the habit stick.
Set intervals by risk, not by habit
The interval question — how often should people check in? — has no single answer, and pretending it does is the first design mistake. The HSE's expectation is proportionality: the monitoring should match the risk. In practice that means working from your lone worker risk assessment and asking, for each role or task: if something went wrong at the start of an interval, how long is it acceptable for nobody to know?
That framing produces sensible spreads. A high-risk task — chainsaw work, confined spaces, a visit to a client with a known history of aggression — might justify check-ins every 30 to 60 minutes, or continuous monitoring with automatic fall detection carrying part of the load. Moderate-risk mobile work — home visits, inspections, site surveys — commonly lands on check-ins at each arrival and departure plus a mid-shift confirmation. Low-risk lone working might need only shift start and end. Write the interval into the risk assessment for each role, with the reasoning — an assessor will ask why you chose the number, and 'it's what the app defaulted to' is not an answer.
Two refinements pay for themselves. First, tie check-ins to natural moments — arriving, leaving, finishing a task — rather than raw clock intervals where possible; a check-in that maps to something the worker is already doing gets done. Second, let the schedule flex with the task: a worker doing paperwork at base and the same worker felling trees that afternoon should not be on the same interval.
Define the missed check-in response to the minute
The response to a missed check-in is the procedure. Write it as a numbered, time-boxed sequence with named roles, and be honest about the hard cases — nights, weekends, the supervisor's holiday. A workable pattern: at the deadline, an automatic reminder prompts the worker (most misses are forgetfulness, and a nudge resolves them in seconds); a few minutes later, unanswered, the system attempts the worker directly by a second channel; a few minutes after that, escalation begins — first responder contacted by push, SMS and voice, then a second responder, then a decision point with a defined owner: welfare check, site visit or emergency services.
Every step needs a number of minutes and a name (a role, not a person — people leave). Test it quarterly the way you test fire alarms: a deliberate missed check-in, unannounced, timed end to end. If your platform automates the sequence, use its dry-run capability first — Vygard, for example, runs missed check-ins into a multi-step cascade across push, SMS, WhatsApp and voice with a test mode that verifies dispatch without paging anyone — and then still run the live drill, because the automation only covers the machine half of the promise. The human half — does the responder actually pick up and act at 6am on a Sunday — is the half that fails.
Log everything: the miss, each contact attempt, the acknowledgement, the outcome. The log is your evidence that the control operates, and it's also your improvement data — a role generating daily false alarms has an interval problem you should fix, not tolerate.
Beat alert fatigue before it beats you
Alert fatigue is the quiet killer of check-in systems. Set intervals too aggressively and workers drown in prompts they resent; flood responders with false alarms and acknowledgements become reflexive rather than considered — the safety equivalent of clicking 'accept cookies'. Both failure modes end the same way: the day a real alert arrives, it's treated like all the false ones.
The fixes are structural. Match intervals honestly to risk so every prompt has a reason the worker can respect. Use pre-deadline reminders so a busy worker gets a nudge before a miss, not an escalation after one. Make cancelling an accidental alarm quick and blame-free — a worker who fears triggering a fuss will disable the system, mentally or literally. Review the false alarm rate monthly per role: it's the single best health metric a check-in procedure has. And when a real escalation happens, close the loop — tell the team what fired, what the response was, and how fast. Nothing builds trust in a system like watching it work.
Make it a habit, not a policy
Compliance with a check-in procedure is cultural before it's technical. The strongest predictor of whether workers check in isn't the app's design — it's whether they believe anyone is genuinely watching and would genuinely come. That belief is built by small, visible behaviours: responders acknowledging promptly, managers treating a missed check-in as concern rather than misconduct ('are you OK?' lands very differently from 'you forgot again'), and leadership using the system themselves when they work alone.
Ban the workarounds explicitly and early — 'I texted my partner instead' is not a check-in, because a partner has no escalation path, no location, and no log. Fold the procedure into induction, revisit it after every incident and near-miss, and put the system's own data — miss rates, response times, false alarm trends — in front of the team periodically. A check-in procedure that people follow isn't the one with the strictest policy document; it's the one where checking in feels like the organisation having your back, and where nobody can remember the last time a missed check-in went unanswered.
Frequently asked questions
- How often should lone workers check in?
- Proportionate to risk — there is no universal interval. High-risk tasks like chainsaw work or confined spaces might warrant check-ins every 30–60 minutes, often alongside automatic fall detection; moderate-risk mobile work commonly uses arrival and departure confirmations plus a mid-shift check; low-risk lone working may need only shift start and end. Set the interval per role in the risk assessment, and record the reasoning.
- What should happen when a lone worker misses a check-in?
- A defined, time-boxed sequence, not an unwatched notification: an automatic reminder to the worker first (most misses are forgetfulness), then contact attempts by additional channels, then escalation through named responder roles with minutes attached to each step, ending at a clear decision point — welfare check, site visit or emergency services. Every step should be logged, and the whole sequence tested regularly with a deliberate missed check-in.
- Can lone workers just text a colleague or family member instead?
- No — informal arrangements fail exactly when they're needed. A text to a partner has no defined escalation if it goes unanswered, no location data, no out-of-hours cover and no audit trail. Check-ins need to route into a system or process where a missed one reliably triggers a defined response and every step is recorded.
- How do you stop workers ignoring or resenting check-in prompts?
- Design against alert fatigue: match intervals honestly to risk so prompts feel justified, send reminders before the deadline rather than escalations after it, make cancelling an accidental alarm quick and blame-free, and monitor false alarm rates per role monthly. Then build trust culturally — respond fast, treat misses as concern not misconduct, and share real examples of the system working.
Last updated 2026-08-01