Telehealth Scheduling and the Visit Nobody Can Connect To

Telehealth scheduling gets blamed for a technology problem it did not cause. The appointment says 2:15. At 2:22 the provider is looking at an empty video room, the patient is in a parking lot tapping a link that will not open, and the front desk finds out about it at 2:31 when the patient gives up and calls the main line. Nothing in that sequence is a software failure. It is a workflow with three moments and nobody assigned to any of them.

Practices tend to file connection failures under technology, which is why they never get fixed. The fix is not a better platform. It is deciding who owns the minutes before, during and after.

Before the visit, somebody has to confirm the patient can actually join. Right device, working link, correct portal login, a phone number that reaches them if it fails.

At the visit, somebody has to notice within about two minutes that the patient has not appeared, and reach them on the channel that still works, which is almost always the phone.

After a failure, somebody has to decide what the visit becomes. Rebooked as video, converted to a phone visit if the practice allows it and the payer covers it, or moved to in person. Whichever it is, the appointment in the calendar has to change to match, or the practice bills for something that did not happen the way the record says it did.

Most practices staff none of those three moments. They staff the complaint that arrives afterward.

Modality is a per-appointment-type rule, not a checkbox

The first thing to correct is the mental model. A virtual visit is not an in-person visit with a video flag turned on. In the calendar it is a different appointment type, and the rules attached to it are specific.

A real provider setup packet for one practice listed eight distinct appointment identifiers for what a patient thinks of as an appointment: office intake, office first visit, office follow-up, and the telemedicine equivalents of each, plus in-person and virtual variants of a study consult. Booking the wrong one is not cosmetic. It changes duration, eligibility and what can be billed.

Slot semantics are worse than the names suggest. One practice was told about a telehealth slot type that, despite the name, did not block non-telehealth bookings, which made it useless for online scheduling. Another practice required in-person modality for every wellness exam type, blocking the virtual path for those visits even where slots existed.

Provider rules layer on top. At one practice a single provider does all telehealth and those visits are almost always started by the clinic rather than the patient, so routing a virtual request to any other physician produces a visit nobody agreed to. These are configuration facts, they differ per practice, and they are exactly the kind of thing an out-of-the-box scheduler cannot express.

The link is a delivery problem, and delivery is measurable

Once the right appointment exists, the remaining work is getting a working link into the patient’s hands and confirming it landed.

Portal reach is better than practices assume and worse than they need. ONC data for 2022 found that about 3 in 5 individuals nationwide were offered and accessed their online medical record or patient portal, and that among those who accessed it, 48% used a website only while 19% used an app only and 32% used both. Roughly two in five people are not reliably reachable through the portal at all, and the ones who are may be on a device the practice never checked.

So delivery has to be plural and confirmed. POST /patients/{patientid}/securemessage/topatient puts the link in the portal, a text carries it to the device the patient is holding, and a short confirmation the day before establishes whether they can open it while there is still time to fix it.

That pre-visit confirmation is the highest-value call in the whole workflow and the one practices skip. It is thirty seconds of administrative work that prevents a fifteen minute provider hole, and it has nothing to do with the patient’s health. Can you open this link, is this the best number if it does not work, do you know it is video and not phone.

Recovery is a booking, not an apology

When a visit drops, the practice’s instinct is to apologize and offer to call back. The correct response is to put a new time on the calendar before the patient hangs up.

What the visit converts into is a rules question with real money attached. Medicare describes telehealth as services delivered using audio and video communications technology, or audio-only communication in some cases, and covers telehealth from anywhere in the United States including the patient’s home through December 31, 2027. Which of those paths this particular practice, this provider and this plan will support is configuration the practice supplies once.

The front office applies it, and it never decides whether a visit should be virtual in the first place. That belongs to the clinical team, and practices routinely hand over both lists: the situations to steer toward a virtual visit, and the ones to keep in the office. An automated front office follows those lists and does not extend them.

One more trap worth naming. A small practice found their virtual platform would not open a patient’s chart at all unless insurance was active, which turned an eligibility gap into a cancelled visit rather than a billing problem discovered later. Checking coverage before a virtual visit is not revenue hygiene in that setup. It is the difference between the appointment happening and not happening.

The calendar has to end up telling the truth

The quiet damage from failed virtual visits is not the lost fifteen minutes. It is that the schedule stops matching reality, and every downstream number inherits the error.

GET /appointments/booked is how a practice sees what the day actually contained rather than what it intended. If a video visit became a phone call and the appointment type never changed, the utilization report, the modality mix and the claim all disagree with each other.

That is why the rebooking step has to update the appointment rather than sit in a note. The visit that happened is the record, and the record is what the practice manages from.

GET /appointments/open closes the loop by putting the vacated slot back where somebody can use it. A fifteen minute hole created at 2:22 is fillable the same afternoon if anyone knows it exists within a few minutes. It is unfillable if it surfaces in a report next week.

What stays with a person

The line here is unusually easy to hold, which is why virtual visits are a good early workflow for a practice that is nervous about automation.

The automation books the correct telehealth appointment type for that provider and that plan, then confirms the patient can join while there is still time to fix it. It delivers and re-delivers the link, and calls when the patient does not appear. It converts the visit to the modality the practice has approved, updates the appointment so the record is accurate, and releases the empty slot.

It does not decide whether a patient’s situation suits a video visit, it does not evaluate anything the patient reports during the readiness call, and it does not advise anyone about being seen. Those questions go to staff, immediately, with the context already captured.

For a practice administrator the promise is narrow and checkable. Fewer virtual visits that quietly fail, and a calendar afterward that says what actually happened.

Key Takeaways

Treat a failed video visit as a workflow gap in the minutes before, during and after, rather than as a platform problem.
Book the specific telehealth appointment type for that provider, since modality carries duration, eligibility and billing consequences.
Deliver the link through both the portal and text, and confirm the day before that the patient can actually open it.
Keep the clinical lists of when to steer toward or away from a virtual visit with the clinical team, and have the front office follow them without extending them.
Convert a dropped visit into a new booking on the call, not into a promise to follow up.
Update the appointment when the modality changes, or the schedule, the utilization report and the claim will all disagree.
Verify coverage before a virtual visit, because some platforms will not open the chart at all without active insurance.

A virtual visit that nobody can connect to is not a technology story. It is thirty seconds of confirmation that never happened, and a rebooking that turned into a callback. Both are front-office work, and both run better when something is watching the clock.

Sources