An ERCP does not end when the patient leaves the endoscopy unit. Sedation wears off on the drive home, the throat is sore, the abdomen feels full and strange, and somewhere between six and twelve hours later the patient is sitting on the edge of a bed at eleven at night trying to decide whether what they are feeling is normal. The office closed hours ago. The discharge sheet says to call with concerns, which is exactly the instruction people are least able to act on when they are not sure whether they have a concern.
One question from PrepQ’s gastroenterology answer library captures that moment better than any other, and it is written the way a real patient wrote it: “After the ERCP I had severe pain and was being sick — how do I know if it’s just gas or actually pancreatitis?”
Here is the physician-written answer PrepQ delivers, by text or by voice, at whatever hour it is asked: “Please get this checked now rather than assuming it’s just gas. Severe belly pain with vomiting after an ERCP can be a sign of pancreatitis, the most common significant complication, and it needs prompt medical attention. Contact your care team right away, and if the pain is severe or you feel very unwell, emergency care may be needed. It’s much safer to be evaluated.”
Notice what that answer refuses to do. It does not weigh the patient’s symptoms and render a verdict. It does not reassure. A patient describing severe post-ERCP pain with vomiting is the textbook case in which software has no business producing a clinical judgment, and PrepQ is built so that it cannot. Messages carrying urgent symptoms are escalated to the practice or to 911 rather than answered by the AI, and questions with no physician-approved answer route the patient to the office instead of being improvised. The system’s job here is not to diagnose post-ERCP pancreatitis. It is to make sure that at eleven at night, a frightened patient is told plainly and immediately to be evaluated, instead of talking themselves into waiting until morning.
Every answer in the library is written by physicians before it enters the system. When a practice subscribes, its own doctors review and approve the answer set for their procedures — and can edit any answer to match their protocols — before a single patient receives one. The platform is HIPAA-eligible, and a Business Associate Agreement is available to subscribing practices.
Why the after-hours load falls so heavily on GI
Gastroenterology carries an unusual combination: a large procedure catalogue, prep instructions that are detailed and strictly time-bound, and recovery guidance that differs meaningfully between a screening colonoscopy, an ERCP, an EUS with fine-needle biopsy, and a PEG tube placement. PrepQ’s gastroenterology library covers more than one hundred procedures — colonoscopy, upper endoscopy, ERCP, EUS and EUS-FNA/FNB, esophageal dilation, esophageal stent, liver biopsy, paracentesis, PEG tube placement, hemorrhoid banding, manometry, pH monitoring, and more — behind hundreds of physician-approved answers.
ERCP concentrates the problem. It is the GI procedure with the most consequential complication profile, the most medication questions beforehand, and the most stent-related follow-up afterward — which means it generates questions in three separate windows: the night before, the night of, and weeks later when a stent is still in place. Those questions arrive on a practice’s voicemail in whatever order the patients happened to think of them.
When the answers are instead available by text and voice at any hour, the pattern that follows is one every GI office recognizes from its own phone log. Patients arrive better prepared, because they got a correct answer at the moment of confusion rather than a callback the following afternoon. Staff start the morning with patients rather than a voicemail queue. Fewer cases are lost on the day to a misread fasting window or a medication held that should not have been. Any dollar figure attached to that is a modeled estimate, not a measured customer result — but the operational shape of it is familiar.
The distinguishing feature was never that software can generate an answer. Anything can generate an answer. What a procedural practice needs is its answer — the one its own physicians stand behind — in the patient’s hand at eleven at night, with the urgent cases pulled out and handed to a human being. PrepQ was built by practicing physicians, Keith Schmidt MD and Ryan Kuta DO, around that standard: correct, approved, and available 24 hours a day, 365 days a year.