Annual Visit Outreach Across a Whole Panel

Annual visit outreach looks like a mail merge and behaves like a call center. A preventative medicine practice decides to reach its whole panel about the yearly visit, sends the campaign, and discovers that the hard part was never writing the message.

The hard part is what comes back. Send four thousand messages on a Monday and a share of those people answer on Monday, all of them at once, to a front desk that was already at capacity.

So practices throttle by doing less. The campaign becomes a quarterly batch, then an annual one, then a list that somebody works when there is time. Coverage drops to whatever the current staffing can absorb, which is a staffing decision wearing the costume of an outreach strategy.

There is a second failure that is quieter and more expensive. A panel campaign sent on the calendar rather than on each patient’s own clock reaches most people at the wrong moment. Some are not eligible yet. Some have been due for four months. The message is identical for both, which teaches the panel that these messages are not really about them.

And then there is the reply that lands on the wrong chart, which nobody notices for a week.

The clock is per patient, not per calendar

The eligibility date is the trigger, and it moves independently for every patient.

Medicare makes this concrete. The annual wellness visit is billable once in a 12-month period, with one code for the first visit and another for subsequent ones, and it cannot be billed within 12 months of the initial preventive physical exam for the same patient. Claims that ignore the interval come back saying the patient reached the benefit maximum for the period. That is not a billing curiosity. It is the outreach schedule, written by the payer.

A campaign built on that clock is a rolling daily list rather than a quarterly blast. Every day, a small set of patients crosses their own eligibility line and enters the queue. The volume is even, the message is accurate for the person receiving it, and nobody gets a reminder for a visit they had six weeks ago.

The operational effect of rolling the list is the part practices notice first. A flat daily volume can be staffed. A quarterly spike cannot, which is why quarterly campaigns quietly become annual ones.

Send where the patient already is

Channel choice decides reach, and most panels split three ways.

Portal adoption is real but partial. About three in five individuals nationwide reported being offered and accessing their online medical record or patient portal in 2022, a 50% increase since 2020. That is a solid majority of the country and still leaves a large minority who will never open a portal message, plus a group who have an account and do not read it.

So the campaign runs on more than one rail. A secure message to the patient inside athenaOne for the portal users, with the inbox read back so a reply is picked up rather than sitting unread. Voice for the patients who answer the phone and will not touch a portal. Text for the ones who reply to a text and never call back. The patient record already carries the contact preferences, which means the routing decision does not need a human.

Language is part of channel, not a feature bolted on afterward. Auto-detection works, but bilingual patients mix languages mid-sentence and a system that flips back and forth reads as broken. Practices generally want certainty before switching, and beyond the two most common languages the honest answer is a human language line. Say that plainly in the design instead of claiming coverage the workflow does not have.

Somebody has to staff the reply

Outbound is cheap. Inbound is where the money goes.

Message volume into practices has been climbing for years while phone volume has held roughly flat, and the load lands on the same people. One analysis of electronic health record use found medical subspecialists spending 9.9% more time in the record during their off hours, which is the sound of an inbox that grew without anyone deciding it should.

A panel outreach program that ignores this is transferring work, not removing it. The reply has to be handled by the same system that sent the message: read it, match it to the patient and the appointment type, offer real open slots, book, confirm, and write the result back so the chart and the calendar agree. Anything that does not fit that pattern gets routed to a named queue with the thread attached, rather than dropped into a general inbox for someone to sort.

The volume test is worth running before launch. Take the daily rolling list, apply a realistic reply rate, and ask who is answering those messages at 4pm. If the answer is the front desk, the campaign is capped at whatever the front desk has left over, which is the situation the practice was trying to escape.

The shared phone number will misfile a chart

Panel campaigns produce this failure at a rate no single-call workflow ever will, for the simple reason that a panel contains families.

A practice reported a mother calling about her daughter, both patients there. The phone number was tied to the mother’s account, so both conversations were documented in the mother’s chart and staff had to move the messages by hand afterward. It is now treated internally as a defect class rather than an incident, covering parent and child as well as spouse and spouse.

The two fixes are dull and they work. Ask at the top of every contact whether this is for the patient or for someone else, and detect the mid-conversation switch when a person hands the phone to their spouse. Both belong in the outreach script itself, not in a cleanup process afterward, because the cleanup is manual and it happens after the note is already in the wrong place.

The same discipline applies to third parties. A caregiver or a service company calling on a patient’s behalf is an exception to identity verification, not a feature of it, and the practice has to say in advance who is allowed and what gets captured about them.

One more small thing that decides whether the whole campaign reads as human: make the confirmation language conditional on the booking. A community health center asked for the standard 48-hour cancellation notice line to be removed from same-day confirmations, where it makes no sense. Details like that are the difference between a message a patient trusts and one they file as spam.

Where the person takes over

The automation sends on the rolling eligibility list, picks the channel, handles the language it can handle, reads the reply, offers real slots, books, confirms, and writes it back.

It hands off on anything clinical, immediately and without a second attempt. A patient who replies with a question about their health, a new concern, or anything that is not about a date goes to clinical staff with the full thread attached. Requests to speak to a person are honored, with the small amount of friction practices actually ask for: acknowledge the wait, say what can be handled now, then transfer.

What this leaves the practice with is the thing operators keep describing when they say they want an extra team rather than another tool. The campaign runs every day at a volume the building can absorb, the replies get worked by something that does not go home at five, and the staff who used to run the list spend the hour on the patients in front of them.

Measure booked visits per hundred patients who became eligible, by month. Message counts and open rates will look good long before that number moves.

Key Takeaways

Run panel outreach off each patient’s own eligibility date, since the Medicare annual wellness visit is billable once in a 12-month period.
Convert quarterly blasts into a rolling daily list so the reply volume is flat enough to staff.
Expect roughly two in five of your panel to be unreachable by portal message alone, and run voice and text alongside it.
Design language support honestly, including where the workflow hands off to a human language line, rather than claiming coverage it lacks.
Budget for inbound, because the reply is where a panel campaign either books visits or transfers work to the front desk.
Ask whether the contact is for the patient or someone else at the top of every conversation, and detect the mid-call handoff to a spouse.
Decide in advance which third parties may act for a patient and what gets captured about them, since that is an exception to identity verification.
Make confirmation language conditional on the booking, so a same-day appointment does not get a 48-hour cancellation warning.

Panel-wide annual outreach only works when it stops being a campaign and becomes a daily process. Trigger on each patient’s own clock, run the channels the panel actually uses, staff the reply with something that scales, and fix the family-chart problem in the script rather than in cleanup. The measure is booked visits per hundred eligible patients, and nothing else on the dashboard is worth arguing about.

Sources