Adding live video visits to an async telehealth program

Add live visits only where they’re required

Add synchronous video visits to an async telehealth program only where the patient’s state, the treatment or the stage of care requires one, and keep everyone else async. At 400 visits a month with a quarter needing a live visit, correct routing costs $20,500 on pay-as-you-go; making every visit live costs $28,000 and loses about four times as many patients.

Picture an async program at 400 visits a month that enters new states or adds a stricter treatment. Suddenly a quarter of patients need video before a provider can prescribe.

What makes a visit live: state, treatment and stage

Three things create the requirement, and they stack:

• The state. Telehealth modality laws vary. Some states require a real-time interaction, usually video, to establish care or before prescribing certain medications. Others allow asynchronous care for appropriate treatments. See what you can prescribe without a video visit.
• The treatment. Controlled substances carry federal rules under the Ryan Haight Act and DEA telemedicine rules, revised more than once; check the DEA telemedicine map rather than hard-coding a date. Some protocols require a live visit on clinical grounds.
• The stage. A state may require a live first visit, then allow async follow-ups. Still choosing states? Weigh this in which states to launch first.

Route async where allowed, sync where required

Don’t skip the last row: a licensed provider can decide any case needs a conversation.

On Lithos, an encounter’s modality is async or sync, and sync means the patient’s state requires a live video visit at this stage. Call GET /v1/encounter_precheck before intake, so the patient learns their visit is live before they pay.

Each 100 live visits loses about 14 patients

A sync patient has to pick a time, show up with a working camera and wait for the provider, and every step loses people. These figures are assumptions, including a 5% async drop; swap in your own.

Per 100 paid intakes, about 95 async encounters complete. On the sync path, 85 book, 76.5 attend the first booking at a 10% no-show rate, and 4.25 more attend after rebooking, so about 81 complete.

Multiply those 14 by your first-year value per patient: that, more than the price gap, is the real cost of a live visit. Route all 400 patients to sync and you lose about 56, most of whom could lawfully have been seen asynchronously.

How much does the live-visit premium cost?

Hold completed visits at 400 a month. On pay-as-you-go, async is $45 and sync $70, all in the first band.

Correct routing saves $7,500 a month, about 27%, before counting lost patients. Rates are graduated by volume, and the sync premium shrinks from $25 a visit in the first band to $20 from 500 visits and $15 from 2,500.

Live visits take half the provider time

Async review is a queue that can wait a few hours. A live visit is an appointment with a provider licensed in the patient’s state, often in their evening. Assume an async review takes 8 minutes. A live visit takes a 15-minute slot plus 5 minutes of notes, and at 10% no-shows, 100 completed visits need about 111 slots.

• Async: 300 × 8 minutes = 40 hours
• Sync: 100 × 20 minutes + 11 × 15 minutes = about 36 hours

A quarter of the volume takes almost half the provider time, in fixed blocks by state (clinician capacity planning covers the licensure matrix). On Lithos, unused slots aren’t on your bill.

The appointment flow in five calls

1. Find slots. GET /v1/appointment_slots for the patient’s state, up to 14 days out. Lead with the next two or three days.
2. Hold one. POST /v1/slot_reservations holds a slot during checkout; release it if the patient leaves.
3. Book it. Create the encounter with modality sync and the reservation_token. For an existing encounter, use POST /v1/encounters/{id}/appointments.
4. Remind and join. Drive the join button from patient_join (opens_at, closes_at, url), which opens 10 minutes before the start by default. Remind at booking, the day before, and just before opens_at with a device check.
5. Handle the outcome. appointment.completed, then encounter.completed, bring the decision and orders. On appointment.no_show, fetch the appointment to see whether it’s patient_no_show or clinician_no_show.

Let patients reschedule and cancel themselves within your notice window; a patient who can move a visit in two taps is less likely to miss it.

Rebook a no-show within minutes

For a patient no-show, send fresh slots as soon as the event arrives, follow up once the next day, and hand off to a person after a second miss. If you charge a missed-visit fee, say so at booking. For a clinician no-show, apologize, offer the earliest slots and raise it in your shared channel.

Track no-shows by days between booking and visit and by time of day; if far-out bookings miss most, offer earlier slots. A no_show webhook can land after a reschedule you’ve processed, so act on the appointment’s current status.

Frequently asked questions