Bariatric Program Coordination: The Clock Nobody Owns

Bariatric program coordination is mostly calendar work. A patient who wants an operation in March has to finish a payer required run of supervised weight management visits. Three or four outside offices have to clear them. Every piece of that paperwork has to stay inside its own expiry window. Miss one month and the run can restart. Nobody on the schedule finds out until a coordinator opens the chart to build the packet.

Every bariatric program has a coordinator and every coordinator has a spreadsheet. It tracks who is on month two of six. It tracks whose psychology report has not come back, whose labs are about to age out, and who sat through the seminar in January and was never called. The spreadsheet exists because scheduling systems track appointments. They do not track a sequence that runs for months.

That spreadsheet is also where the money sits. A patient who completes four supervised visits and then drifts has used up four visits worth of coordinator time and produced no surgery. The work happened. It never turned into the case it was supposed to protect.

What makes it frustrating is how little of it is medicine. The clock, the packet, the fax chase, and the post op series are logistics. That is the part a program should be able to hand off. It is also the hardest part to hand off, because the rules live in one person’s head and the deadlines live in four different documents.

The supervised visit clock restarts on a gap

Payers disagree about how long the supervised period runs, and that is the first operational problem. Illinois Medicaid requires six consecutive months of participation in a medically supervised weight loss program. It has to be finished within one year before the prior approval request. Other policies run shorter. A coordinator with a mixed payer panel holds two or three clocks at once and has to know which patient is on which.

The word doing the damage is consecutive. A patient who books month four and cancels the night before has not lost an appointment. They have put the whole run at risk. Whether it restarts is a question about that specific plan. Most programs do not ask it until someone checks the chart against the policy weeks later.

A scheduling template cannot express any of this. A template holds slots. It has no idea that a patient owes a visit inside a named month, and it will never tell you the visit was cancelled and not rebooked.

Outbound automation fits that shape well. It works from the athenaOne order or follow up task that opened the episode. The AI calls the patient inside each month’s booking window, offers the appointment type the program actually uses against the right provider group, books it, and confirms it. When a visit cancels, it calls back the same day instead of waiting for anyone to notice. What it will not do is rule on whether a broken run restarts. That is a coverage question, and it goes to the coordinator with the dates attached.

Every item in the clearance packet has its own expiry date

The submission cannot go out until the packet is complete. Most of the packet is made of documents the program does not generate. Psychology, cardiology, pulmonary, and sleep reports arrive as faxes from outside offices on those offices’ schedules. While everyone waits, the clocks keep running. Illinois Medicaid wants the history and physical finished within six months of the prior approval request, and the psychosocial behavioral evaluation within twelve. Two documents, two windows, one submission date that both have to survive.

So a delay anywhere in the workup is expensive rather than annoying. Finished work ages out and gets repeated. The patient who has already given the program five months is asked for more.

Those requirements are measurably where patients disappear. One review looked at 1,475 patients evaluated for bariatric surgery across twelve major payors. Longer required diet duration, a primary care letter of necessity, cardiology evaluation, and advanced lab testing each independently predicted a patient would drop out. The ones who did reach surgery waited an average of 5.8 months from their first visit, and gained weight over that stretch.

Everything an automation does here is a chase. The AI calls the outside office to ask where the report is. It re requests the report when nothing was sent. It calls the patient when the outside appointment was never booked at all. Where the program has a relationship that allows it, the AI books that outside visit, files the returned document to the right work queue, and tells the coordinator which item is now the long pole. Reading the report and deciding a patient is cleared stays with the surgeon.

The seminar to consult handoff leaks before anyone is in the funnel

Most programs still open with an information seminar, in a room or on a webinar. Attendance is the real top of the funnel. The handoff from that attendance list to a booked surgical consult is almost always manual, which means it happens when somebody has a free afternoon.

The leak is large and it has been measured. At one academic program, 484 patients attended a bariatric information session in a single year. Among the candidates still eligible after multidisciplinary review, 30% dropped out before surgery. Three months of supervised weight loss documentation, endocrinology clearance, urine drug screening, and extra dietitian sessions were each independently linked to that dropout.

Read that as an access problem rather than a motivation problem. Every one of those requirements is an errand the patient has to run between calls that nobody is making.

A seminar attendance list is the warmest outbound calling list a bariatric program owns, and it cools within a week. The AI calls each attendee inside a couple of days. It answers the logistics questions people actually have about cost, timing, and what happens first. Then it books the consult against the correct athenaOne appointment type and marks the dead numbers so the list gets shorter instead of longer. Anything that becomes a question about whether surgery is right for this person goes to the program’s clinical staff, with the call history attached.

The second appointment is the one that silently never happens

Multi stage bookings fail the same way in every specialty, and the failure is quiet. Where paired visits are the norm, a booking can complete the first appointment cleanly while the follow up task that should create the second is never generated. The second half of the workflow simply does not exist, and it surfaces in review rather than at the time.

Bariatrics carries more of those pairs than most programs. The supervised visit and the nutrition visit attached to it. The clearance appointment and the consult that depends on it. The surgery date and the pre op visit that has to land inside the window in front of it.

Reminders have a related trap. Native reminders commonly fire on the first appointment of the day. A patient with two bookings gets reminded about one and misses the other. On a morning built out of a lab draw and a program visit, one of those quietly does not happen.

The fix is boring. Treat the pair as the unit of work. Book both legs or neither, confirm both, and send a reminder per appointment rather than per day. Then run the report that finds patients with only one leg on the books. Most programs have never seen that report.

One more reason those gaps survive. A follow up task and an open order can both be alive on the same patient at once. Satisfying the task does not always close the order, so somebody calls the patient again about something already handled. Staff also move items between queues by hand, so the queue an automation reads is not always the queue the item ended up in. Writing that policy down, per program, comes before automating any of it.

The post op series runs for years and no template holds it

Follow up after surgery runs on a schedule that outlasts most staff tenures. Two weeks, six weeks, three months, six months, a year, then annually for as long as the program can hold on to the patient. That is six or seven bookings, each triggered by the calendar rather than by anything the patient does. They are spread across years in which people move, change jobs, and change plans.

Attrition is well documented and it starts early. One review followed 554 consecutive patients eligible for one year follow up. Self pay status more than tripled the odds of being lost to follow up. Men, and members of racial and ethnic minority groups, were roughly twice as likely to be lost. The accreditation program bariatric centers report to defines loss to follow up at one year as a measured outcome. That rate lands on the program’s scorecard whether or not anyone is working the list.

Recall outreach at this cadence lives entirely in the margin, and the margin is the first thing a short staffed office gives up. The list also rots. Numbers change and patients move, and a share of the list has already been seen somewhere else. Half of any recall campaign is finding out which entries are still real.

Start the automation there. The AI works the recall list on the cadence the program sets and calls at the six month, one year, and two year marks. It books against the correct athenaOne appointment type and provider group, confirms the booking, and flags entries that are no longer valid so the list shrinks. Calls that turn into questions about how the patient is doing go to clinical staff with the history attached. That conversation was never the automation’s to have.

What to hand over, and where it has to stop

The programs that get value out of this are not buying a bariatric scheduling product. They hand over a list of gaps and get a team that fills them, in the order they name, against the systems they already run. That posture matters more than any feature list. A bariatric program’s rules are specific enough that a fixed product will cover most of them and hand the rest back.

Start with the surfaces already sitting in athenaOne. The order and follow up task queue that opens the episode. The appointment types the program uses for supervised visits, consults, pre op, and each post op interval. The provider groups and departments those types are bookable against. The document classes the outside clearances land in. None of that has to be built. But somebody does have to write down which slot type can actually hold which visit, per provider and per department. That is usually the longest part of an implementation.

Then draw the line and keep it visible. The AI books, rebooks, confirms, chases paperwork, and reports what it could not do. It does not read a report. It does not decide a patient is ready. It does not tell anyone whether an operation is a good idea. All of that goes to a person, with the call and the dates attached so nobody starts cold.

The honest version of the pitch is that none of this removes a coordinator. It removes the dialing, the voicemails, and the refaxing. What it gives back is the part of the week that is judgment about a specific patient in a specific plan.

Key takeaways

Track the supervised run as an episode with a per patient deadline, not as six unrelated appointments. The cancelled month is the failure, and a template will never surface it.
Put every packet item on its own countdown. The history and physical, the psychology report, and the labs expire on different clocks. Whichever expires first sets your real submission date.
Call the seminar attendance list within days rather than when someone has a free afternoon. It is the warmest outbound list the program owns and it cools fast.
Book linked visits as a pair or not at all, and send one reminder per appointment rather than per day. Then run the report that finds patients with only one leg booked.
Automate the top of post op recall so the list shrinks. Marking a dead number is worth as much as booking a live one, and accreditation scores the follow up rate either way.
Write down which athenaOne appointment types each generic template slot can actually hold, per provider and per department, before automating anything. Nobody has ever had to say it out loud before.

Bariatric programs rarely lose patients at the front door. They lose them in month four. Or in the six weeks a psychology report sat in somebody’s fax queue. Or eighteen months after an operation that went fine. Every one of those is a call that nobody had time to make.

The work is not complicated. It is relentless, it is administrative, and it runs on a clock nobody in the building owns. If your program runs on athenaOne, the order queue, the appointment types, and the provider groups are already there. Somebody just has to work them every week, forever, without getting bored.

Sources: