The Order Queue Is Where GI Referrals Go to Die

A referral arrives, somebody creates the order, and everyone treats that as the finish line. It is not. The order queue is a holding pen, and a gastroenterology practice can carry hundreds of patients in it who were referred, entered correctly, and never scheduled.

The reason this hides so well is that nothing about it looks like a failure. The referring office sent it. Your staff received it. The order exists in the chart. Every party can point at their own completed step.

What nobody owns is the gap between the order and the appointment. That gap is where a colonoscopy referral turns into a patient who assumed somebody would call, and a practice that assumed the patient would.

The loop is not closed as often as anyone thinks

Practices tend to estimate their referral conversion from memory, and memory is generous. The published work is not.

An analysis of primary care referrals to specialists in a large health system looked at 103,737 referral scheduling attempts and found that only 36,072, or 34.8%, resulted in documented complete appointments. Nearly four in 10 scheduling attempts, 38.9%, lacked an appointment date entirely. That study covers referrals across specialties rather than gastroenterology alone, and the mechanism it describes is the one every GI practice recognizes.

A procedure-heavy specialty makes the gap worse, not better. A referral to an office visit competes with one calendar. A referral for a screening colonoscopy competes with a prep conversation, an anesthesia question, a coverage question, and a patient who has to arrange a driver and a day off.

Every one of those is a reason to put the phone down and deal with it later. Later is where the queue comes from.

Storing the referral well is not the same as working it

Most groups have already solved the storage problem and think they solved the whole thing. When MGMA asked how practices manage patient referrals, more than three out of four, 76%, use their EHR at 66% or dedicated referral management software at 10%, while about one in five, 21%, still rely on manual tracking. The poll had 309 applicable responses.

That is a good result for data integrity and a neutral one for conversion. A referral sitting in a system is legible, searchable, and just as unbooked as a referral sitting in a fax tray. The system records the work. It does not do the work.

What converts an order is somebody making contact, repeatedly, at times a working adult can answer. That is a labour problem, and it is exactly the labour a practice never has enough of, which is why the queue grows in the first place.

On athenaOne the order and tickler queues are the surface where this lives. They are a genuinely rich list, they carry the class and the age of every open item, and in most practices nobody has ever been assigned to them as a daily job.

Work the queue by age and by what is blocking it

The instinct is to work the queue newest first, because new items feel urgent. That is backwards. A referral three days old will usually convert on the first call. One that is six weeks old needs a different conversation and a different offer.

Sort by age, then by blocker. Some open orders are waiting on a coverage answer. Some are waiting on a prep instruction the patient did not understand. Some are waiting on a slot that did not exist when the patient last called. Those are three different outbound calls, and treating them as one generic reminder is why generic reminders convert badly.

The automation can carry all three. It knows the order’s age and class, it can check what coverage is on file before it dials, and it can offer real openings rather than promising a callback. Multiple attempts across different hours cost it nothing, which is the part a person cannot match.

Set a stop rule and honour it. After a defined number of attempts across defined windows, the item stops being an outreach problem and becomes a decision. Somebody has to choose whether to write to the patient, tell the referring office, or close the order. That decision belongs to a person.

The complication: the referral is rarely one appointment

Here is the thing that makes GI order queues different from a general referral backlog. What arrives as one referral is frequently two or three linked bookings, and nothing in the scheduling template enforces the link.

A practice may book a procedure and a provider visit as a pair with specific spacing between them. The pairing exists as convention rather than as a rule the system knows. The patient just calls asking to see the doctor. The failure mode is booking one leg and not the other, and the second half of the workflow silently does not exist until a human catches it in review.

Native reminders make this worse in a specific way. They fire on the chronologically first appointment only, so a patient with a morning study and a later provider visit gets reminded about one of them. The other one is a no-show waiting to happen, and it will be recorded as a patient problem.

So the automation confirms the set rather than the appointment. It checks that every leg the order implies actually exists, reminds against each leg separately, and raises the ones where a leg is missing. It does not decide what the patient clinically needs, and it does not create the second appointment on its own judgment. It flags the incomplete set and a scheduler completes it.

Report it as a funnel or it will not improve

Practices measure referrals received and procedures performed, and those two numbers sit at opposite ends of a process nobody sees the middle of.

The useful report has four stages. Referrals received. Orders created. Patients reached. Appointments booked. Each drop between stages has a different owner and a different fix, and collapsing them into a single conversion rate hides which one is broken.

Age the open items alongside that. Orders open more than 14 days, more than 30, more than 60. A queue that is growing at the 60-day end is not a communication problem anymore, it is a capacity problem, and the answer is a schedule change rather than more calls.

Run the report weekly and read it in that order. The point is not the total. It is finding the one stage where your practice loses more patients than the stage before it.

Key Takeaways

Treat the created order as the start of the work rather than the end, because a stored referral converts no better than a filed one.
Work the athenaOne order and tickler queues by age and by blocker, since a three-day-old order and a six-week-old order need different calls.
Check what coverage is on file before dialling, so the outbound call can answer the question that stalled the patient.
Confirm the whole set of linked appointments a GI order implies, because native reminders fire on the first one only and the second leg goes missing.
Set an explicit stop rule after a defined number of attempts and hand the decision to close, write, or notify the referring office to a person.
Report referrals as a four-stage funnel from received to booked, and age the open items, so you can see which stage is actually leaking.

A gastroenterology practice with a large order queue is not disorganized. It is doing the intake correctly and stopping one step short of the outcome. Work the queue daily by age and blocker, dial with the coverage answer already in hand, confirm every leg of a linked procedure rather than just the first, and stop cleanly when it becomes a judgment call. The referrals are already yours. They are just sitting one phone call away from being patients.

Sources