Decoding Prescribing Decisions in Ulcerative Colitis

A Patient chart audit report across seven gastroenterology cases — what seven physician conversations revealed about prescribing logic in Ulcerative Colitis

About This Report

This report presents findings from seven Patient Scribe conversations with gastroenterologists managing Ulcerative Colitis patients. Rather than asking physicians to select from predetermined categories, Patient Scribe conducts a voice-first conversation — capturing clinical reasoning in the physician’s own words, probing when something unexpected surfaces, and structuring the output without compressing the underlying logic. The conversations captured current prescribing logic across the IL-23, JAK inhibitor, and anti-TNF classes — including the reasoning behind drug selections, near-prescriptions that did not happen, and the nonclinical factors that determined intra-class outcomes. What they surfaced is the kind of intelligence that checkbox audits record too late and too incompletely to act on.

WHAT THIS REPORT COVERS

• 7 gastroenterologist conversations via Patient Scribe
• Drugs in scope: Skyrizi (risankizumab) · Tremfya (guselkumab) · Rinvoq (upadacitinib) · Humira / adalimumab biosimilar · Remicade / infliximab · Entyvio (vedolizumab)
• Patient profiles ranging from biologic-naive to post-anti-TNF, moderate to severe disease
• Includes cases where Skyrizi was clinically evaluated and not prescribed — and the case where it was the only defensible option
• Covers both the induction architecture dynamic and the safety-driven class selection logic

 Patients range from biologic-naive to post-anti-TNF, moderate to severe disease, across commercial and managed care insurance. Three of seven had Skyrizi in their physician’s consideration set — and were not prescribed it. Drugs in scope: Skyrizi, Tremfya, Rinvoq, Humira, Remicade, and Entyvio.

What a Traditional PCA Would Have Captured

Below is a direct comparison for two of the seven cases: what a checkbox-based PCA would have recorded versus what the Patient Scribe conversation actually revealed. The structured data is not wrong — it is structurally incomplete in ways that generate the wrong commercial conclusions.

The checkbox records the prescribed drug. The conversation captures the consideration set, the exclusions, and the nonclinical factors that determined which drug made it to the chart — and which did not.

What Patient Scribe Revealed

Across seven conversations, four distinct insight categories emerged that a traditional PCA would have flattened, missed, or recorded under the wrong driver.

1. Skyrizi Lost on Induction, Not Evidence

Three of seven patients in this dataset had Skyrizi in their physician’s consideration set. None were prescribed it. The clinical profile was not the issue in any of these cases. The common thread was induction architecture: Skyrizi requires IV induction, while Tremfya offers subcutaneous induction. For patients who want at-home administration, and for physicians navigating insurance authorization, that difference determined the outcome. A traditional PCA records the drug that was prescribed. It does not record the drug that was considered and excluded — or why.

• P1: Biologic-naive patient wanted at-home induction. Skyrizi evaluated and excluded on this basis alone. The chart records a Tremfya initiation; the clinical consideration for Skyrizi is invisible.
• P3: Travel schedule made infusion center commitments impractical. An oral agent solved the administration problem that IV biologics could not. The switch reason in the chart is “inadequate response to prior biologic” — the actual driver was route of administration.
• P5: Physician evaluated both IL-23 inhibitors. Insurance authorization was the deciding factor: subcutaneous induction cleared approval more readily than IV. Skyrizi’s induction requirement created a payer friction point that Tremfya did not.
• P2: JAK inhibitors and anti-TNFs excluded on malignancy risk. Within-class selection ran on physician familiarity and clinical experience, not comparative efficacy. The chart records a Skyrizi initiation following infliximab antibody development. The safety-driven class elimination that preceded it does not appear anywhere in structured data.
• P5: Prior Humira inadequacy drove the switch. Physician moved directly to IL-23 consideration — the anti-TNF class was not re-evaluated. Intraclass, insurance authorization and patient preference for SC administration determined the final drug.
• P6: Remicade inadequate. The class-level exclusion of anti-TNFs was implicit, not documented. Within IL-23s, the physician assessed Tremfya and Skyrizi as clinically comparable and selected on familiarity.
• P7: Remicade inadequate. Post-TNF class shift followed. Insurance authorization and patient preference for subcutaneous administration determined the final selection within the IL-23 class.
• P6: Physician assessed Tremfya and Skyrizi as “honestly quite similar” on clinical grounds. Skyrizi selected on accumulated prescribing experience alone. A checkbox PCA records a Skyrizi prescription; it does not record that the two were considered equivalent and that experience, not efficacy, made the call.
• P2: Physician preference for Tremfya existed at the time of prescribing but familiarity with Skyrizi was greater. The drug that was “mildly preferred” was not selected. The drug that was more familiar was.

“I prefer Tremfya because of the subcutaneous induction and maintenance therapy.”
— Gastroenterologist, UC patient audit (P5) — Tremfya over Skyrizi

COMMERCIAL IMPLICATION 
Three Skyrizi near-prescriptions in seven conversations — all blocked by IV induction architecture or the payer friction it creates. A traditional PCA shows Tremfya preferred; the natural commercial conclusion is a messaging or awareness problem. The actual driver is induction format and access. Those require different commercial responses.

2. Safety Profile Forces the Class

The same dataset that shows three Skyrizi near-prescriptions lost to induction architecture contains one case that runs in the opposite direction. A patient with a documented history of recurrent Basal Cell Carcinoma. JAK inhibitors and anti-TNF agents were not deprioritized — they were categorically excluded on malignancy risk before the drug conversation began. That left the IL-23 class. Within it, Skyrizi was selected on accumulated clinical experience. The commercial argument here is not induction architecture. It is the only defensible safety profile for a specific and identifiable patient type.

“I wanted to avoid anti-TNF therapy and JAK inhibitors due to her history of malignancy.”
— Gastroenterologist, UC patient audit (P2) — Skyrizi as first clinical choice

COMMERCIAL IMPLICATION 
UC tells two Skyrizi stories. When Skyrizi loses, the reason is induction architecture and payer friction. When Skyrizi wins, it is the only defensible class for a patient where JAK inhibitors and anti-TNFs carry documented malignancy risk. Both stories require different commercial responses. Neither surfaces in checkbox output.

3. Class Before Drug

For patients with prior anti-TNF failure, the prescribing conversation in this dataset did not begin with a drug. It began with a class. Physicians who had already seen anti-TNF inadequacy were not re-evaluating whether to use another anti-TNF — that decision was made before the appointment. The drug conversation happened within the IL-23 class only. A traditional PCA records the switch from Remicade to Skyrizi or Tremfya as an “inadequate response” event. The class-level elimination that preceded that switch is absent.

“Patients already previously tried Remicade, thus it did not seem wise to try another TNF.”
— Gastroenterologist, UC patient audit (P6) — post-TNF class shift

COMMERCIAL IMPLICATION 
For post-anti-TNF patients, the class decision is made before the drug conversation. What looks like intra-class prescribing data in a traditional PCA is actually the tail end of a two-stage decision process — only the second stage is visible in the chart.

4. Familiarity Breaks the Tie

When physicians in this dataset assessed Tremfya and Skyrizi as clinically equivalent, the tiebreaker was not a clinical argument. It was accumulated prescribing experience. This dynamic appeared in two cases and ran in both directions — familiarity with Skyrizi broke in favor of Skyrizi; familiarity with the outcome of a prior switch broke against Tremfya when it was newer to market. The clinical profile of the drugs was not the differentiator. The physician’s exposure to them was.

“Tremfya was very newly on the market at the time, but it would have been absolutely a consideration if I had been more comfortable with it when I made the therapy change.”
— Gastroenterologist, UC patient audit (P2) — Skyrizi selected over preferred Tremfya

COMMERCIAL IMPLICATION 
Familiarity, not efficacy, is the tiebreaker when physicians consider the two IL-23 inhibitors clinically equivalent. The commercial lever is peer-to-peer case sharing and real-world experience data — not additional head-to-head comparisons. Physicians who have used either drug are already convinced on clinical grounds. The gap is exposure.

The Cases Behind the Patterns

Three patient groups. Seven cases. Each illustrates a prescribing dynamic that a traditional PCA would have recorded incompletely — or not at all.

Group 1 — Three Skyrizi Near-Prescriptions Blocked by Induction Architecture

P1 — 21M | COMMERCIAL | LEFT-SIDED, MODERATE | BIOLOGIC-NAIVE → TREMFYA

21-year-old male. Commercial insurance. Moderate left-sided UC. No prior advanced therapy.

Biologic-naive and motivated to start treatment. Skyrizi was in the physician’s consideration set — not ruled out on clinical grounds. The patient wanted at-home induction, and Skyrizi does not currently offer that as an option. Tremfya was initiated. The chart records a Tremfya first-line start. It does not record that Skyrizi was evaluated and excluded on induction format — or that a different formulation would likely have changed the outcome.

P3 — 38M | COMMERCIAL | MODERATE-SEVERE | POST-HUMIRA (ANTIBODY DEV.) → RINVOQ

38-year-old male. Commercial insurance. Moderate-severe UC. Humira antibody development documented.

Humira antibody development drove the switch. The prescribing record stops there. What it does not capture: the patient’s travel schedule made infusion center commitments impractical. An oral JAK inhibitor solved the administration problem that biologic agents could not. Rinvoq was selected. The switch reason will appear as “inadequate response”. The actual driver was route of administration and scheduling reality.

P5 — 40F | COMMERCIAL | PANCOLITIS, MODERATE | POST-HUMIRA → TREMFYA

40-year-old female. Commercial insurance. Moderate pancolitis. Humira inadequate response.

Post-Humira, the physician evaluated both Tremfya and Skyrizi. Both were clinically appropriate. Insurance authorization was the deciding factor: subcutaneous induction clears approval more readily than intravenous. Patient preference for at-home administration reinforced the call. Tremfya was selected. The prescribing data will show a switch from Humira to Tremfya. It will not show that Skyrizi was considered and that its IV induction requirement was the reason it was not prescribed.

Group 2 — When the Safety Profile Makes Skyrizi the Only Answer

P2 — 42F | UMR | PANCOLITIS, MODERATE | POST-INFLIXIMAB · BCC HISTORY → SKYRIZI

42-year-old female. UMR insurance. Moderate pancolitis. Infliximab antibody development. History of recurrent Basal Cell Carcinoma.

The prescribing logic in this case ran through the safety profile before it reached the drug. JAK inhibitors and anti-TNF agents were categorically excluded on malignancy risk — not deprioritized, not down-tiered. Excluded. That narrowed the consideration set to the IL-23 class before a drug name was raised. Within the class, the physician volunteered that “Tremfya is actually mildly preferred because it has a quicker onset of action and a similar safety profile” — but selected Skyrizi on accumulated clinical experience at the time of the switch. The chart records a Skyrizi initiation following infliximab failure. The two-stage decision — safety-driven class elimination, then familiarity-driven within-class selection — appears nowhere in structured data.

Group 3 — After TNF Failure, Class Is Decided — Within-Class Runs on Familiarity

P6 — 32F | DPO | EXTENSIVE PANCOLITIS, SEVERE | POST-REMICADE → SKYRIZI

32-year-old female. DPO insurance. Severe extensive pancolitis. Remicade inadequate response.

Remicade failed. A second TNF agent was not considered. The class decision was made before the drug conversation started. Within IL-23s, the physician assessed Tremfya and Skyrizi as clinically comparable: “These two medications honestly seem quite similar. However, I’m just more experienced with prescribing Skyrizi.” Skyrizi was selected. The prescribing record shows a switch from Remicade to Skyrizi. The class-level exclusion of anti-TNFs is invisible. The familiarity-driven within-class tiebreaker is invisible. What looks like a Skyrizi preference is actually a two-step decision that the checkbox format compresses into one.

P7 — 55M | COMMERCIAL | PANCOLITIS, MODERATE | POST-REMICADE → TREMFYA

55-year-old male. Commercial insurance. Moderate pancolitis. Remicade inadequate response.

Remicade inadequate. Post-TNF class shift followed — same clinical logic as P6. Within the IL-23 class, insurance authorization and patient preference for subcutaneous administration determined the final selection. Tremfya was prescribed. The chart records a switch from Remicade to Tremfya under “inadequate efficacy.” The class-level decision, the drug-level consideration set, and the nonclinical factors that produced the final outcome are all absent.

The Gap Between the Chart and the Decision

In Ulcerative Colitis, the prescribing decisions that determine market share are being made in physician reasoning that traditional PCAs are not designed to capture. Three Skyrizi near-prescriptions in seven conversations — none of which appear in any chart as a Skyrizi exclusion. One case where Skyrizi won not on efficacy but because it was the only defensible class for a patient with a documented malignancy history. And a consistent within-class pattern where familiarity, not clinical preference, breaks the tie between two drugs physicians consider equivalent. 

A brand team reading a traditional PCA in this indication sees Tremfya preferred and Skyrizi passed over — and draws the obvious commercial conclusion: messaging or awareness. The actual barrier is induction architecture and the payer friction it creates. The commercial response is patient services and access support, not awareness spend. Those are not the same investment, and the wrong one does not move the metric. 

The gap between what physicians decided and what the chart recorded is where commercial strategy gets made on incomplete information. Conversational AI closes that gap — not by replacing researcher judgment, but by giving it better raw material. 

See What Patient Scribe Would Uncover for Your Brand 

ZoomRx is offering a zero-cost pilot wave for teams who want to see what Patient Scribe surfaces for their brand. Same rigor as a full PCA study, no commitment required.