There are two ways an agency learns that this morning's visit did not happen. In the first, a coordinator's screen flags it at 8:47, the office phones the client by 8:55 with an apology and an ETA for cover, and the family hears about the problem in the same call that announces its solution. In the second, the client's daughter phones at 6 p.m., because her mother mentioned that nobody came today, and the coordinator's stomach drops while they search for records that were never made. The care failure is identical. The relationship damage differs by an order of magnitude, and which version an agency lives in is not luck. It is the presence or absence of a specific operational discipline, and it is worth describing precisely, whether or not you ever buy software for it.
The miss has a timeline, and the office is racing it
A missed visit is not an event so much as a window. It opens when the caregiver fails to arrive and closes when someone in charge knows and acts. Everything about detection is about shortening that window. The mechanics that do it: a definition of "missed" in minutes, a start time plus a configurable grace window, because home care runs on travel and honest slippage; and an automated sweep, every fifteen minutes on the agency's own clock, comparing the schedule against the punches. In Carelyst that sweep also watches the adjacent failure modes in the same pass: the assigned shift with no clock-in after start, which is the critical one, the shift starting within 24 hours that nobody is assigned to, the visit running long past its planned end, and the visit that never clocked out at all, judged with overnight-aware slack so a legitimate all-day shift ending near midnight is not flagged while the caregiver is still punching out.
Detection is only respectable if the signal stays clean, so alerts self-heal: the late clock-in clears its own missing-clock-in alert, the eventual clock-out clears the running-late ones, and the alert screen holds only items still needing a human decision, deduplicated per shift, with an acknowledge-and-resolve trail. A true miss never self-clears. Somebody owns it, and the record shows who.
The family should never be the sensor
Here is the uncomfortable truth in the second version of the story: the family was functioning as the agency's monitoring system. That is backwards, and fixing it is partly a portal design question. Families with portal access to their parent's schedule see the truth as it changes: a cancelled visit shows as cancelled, and the cancellation pings them; a moved time pings them; a hospital-stay hold marks the affected visits as paused with a plain-language note that nobody will attend. The portal boundary matters as much as the transparency: internal dispatch churn, offers and reassignments, stays internal, so the family sees a calm, truthful schedule rather than the office's bloodstream. An agency that communicates this way converts its worst operational moments into demonstrations of competence, which is the only good thing a missed visit can ever become.
The same courtesy runs in the other direction, toward the field. When the office cancels an assigned shift, the caregiver should be pinged immediately, so nobody drives across town to a dead visit or waits outside a door that was never going to open; that failure is a missed visit too, just with the roles reversed. And within visits that do happen, the early-warning layer keeps listening: wellness questions answered at the visit can flag automatically when an answer crosses a threshold, so the deterioration that would otherwise cause next month's crisis surfaces as this week's quiet alert.
The fifteen minutes after a true miss
Detection buys the office a head start; what the head start is for deserves to be written down as a playbook, because a flustered improvisation at 8:47 wastes it. First, call the client or their contact, before anything else, and establish safety, particularly where the visit carried medications, insulin timing, or a transfer someone may have attempted alone; a miss on a companionship visit and a miss on a morning personal-care visit are different emergencies. Second, dispatch cover: broadcast the shift as an open offer to eligible caregivers, first accept wins, and tell the family the ETA. Third, record the truth: if harm or a near-miss occurred, open an incident with its notifications and follow-up; if it turns out the caregiver actually attended and the record is what failed, correct the punch through the audited flow rather than shrugging the alert away. Ten minutes of playbook converts the agency's worst moment into its most convincing one.
Prevention is a scheduling discipline wearing a monitoring costume
Most missed visits are conceived days earlier: the shift nobody was ever assigned to, the caregiver who was double-booked or sent somewhere they were excluded from, the sick call that scattered a day's work into voicemails. So the deepest fixes sit in the scheduling layer, and the FAQ below walks through them: unfilled-shift visibility, the one-action sick call that returns shifts to the fill queue and notifies everyone touched, offer broadcasts that skip anyone absent or unavailable rather than letting them accept into a rejection, and exclusion lists that hold at assignment time. The sweep then guards the residue, and the residue is small.
The last mile is memory. Every miss should leave a record that feeds the questions worth asking monthly: which clients, which time slots, which caregivers, which causes? A daily summary showing scheduled against completed and missed, and a missing-punches report separating never-came from came-but-never-clocked, turn the anecdotes into a pattern the office can actually fix. That loop, detect fast, tell the family first, prevent upstream, learn monthly, is the whole discipline, and it is what Carelyst automates end to end, from the fifteen-minute sweep to the family's plain-language ping. If your agency still finds out about Tuesday from a phone call on Tuesday evening, start a free 14-day trial and move to the version of the story where the office calls first.
Frequently asked questions
Define it in minutes or you cannot detect it in minutes. A workable definition: a visit is late the moment its start time passes without a clock-in, and missed once a grace window, set by the agency, expires with nobody having arrived. The grace window matters because home care runs on travel between homes and honest fifteen-minute slippage is normal; alerting on every late start buries the office in noise, while alerting only at end of day buries the client. The other definitions worth writing down: an unfilled shift, no caregiver assigned as the day approaches, is a miss you can see coming; and a visit where the caregiver attended but never clocked in looks identical to a miss in the record until someone checks, which is why punch discipline and missed-visit detection are the same project.
Within minutes of the grace window expiring, and automatically. The mechanism is a monitoring sweep: an automated check running every fifteen minutes against the schedule in the agency's own timezone, raising a flag when a shift's start plus grace has passed with no clock-in. The same sweep should catch the adjacent failures: an assigned shift starting within 24 hours that has no caregiver, a visit running well past its planned end, and a visit that was never clocked out. Anything slower than this is not monitoring, it is history; a hungry or frightened client does not experience the miss at the end of the day, they experience it at 8:15 in the morning, and the office's response clock starts then whether the office knows or not.
They should heal themselves whenever reality answers them, and demand a human only when it does not. If a caregiver clocks in late, the missing-clock-in alert should clear itself, because the thing it warned about is over; the record still shows the lateness for pattern-spotting. If a visit finally clocks out, the running-late alerts about it should close. What must never self-clear is a true miss: nobody came, and someone in the office has to own the recovery, call the client, dispatch cover, record what happened. The design principle is that coordinators should only ever see alerts that still need a decision. Offices whose alert screens fill with stale noise learn to ignore the screen, and an ignored alert screen is more dangerous than none at all.
Told, yes; told by the agency, ideally; and told before they discover it, always. The trust-destroying event is not the miss itself, families understand that caregivers get sick and cars break down, it is the silence after it, and the suspicion that nobody upstream noticed either. The clean pattern is a family portal that reflects the truth automatically: a cancelled visit shows as cancelled and pings the family, a changed time pings them, and care put on hold shows as paused with a plain-language note that nobody will attend, so no one waits at a window for a caregiver who was told not to come. Routine internal churn, which caregiver was swapped for which, deliberately stays internal. Transparency about events, discretion about operations: families want to know what affects them, not how the sausage is dispatched.
Most misses are born days earlier as staffing facts, so prevention lives in the scheduling layer. Unfilled shifts should be visible as a count the whole office watches, with anything starting soon flagged loudly. A caregiver calling in sick should trigger one action that unassigns their shifts back into the fill queue and notifies everyone affected, not a manual hunt through the week. Open shifts should broadcast to eligible caregivers with first-accept-wins, and eligibility should be enforced: someone marked absent or outside their availability should never even receive the offer they cannot honestly take. And exclusions should hold at assignment time, so the miss that is really a refusal, a caregiver sent where they should never have been sent, cannot be scheduled at all. After all that, the sweep catches what slips. Prevention plus detection, in that order.
Related reading
See Carelyst with your own clients and caregivers.
Start your 14-day free trialNo credit card to start · Cancel anytime