Same Disease. Different Prior Therapy. Different Choice.
And now we can measure where the prescribing algorithms diverge.
The Finding
A common segmentation approach treats prescriber preference as a fixed attribute. Our analysis of 800+ cGVHD patients suggests the 3rd-line choice is pathway-dependent, dictated by which 2nd-line therapy failed. The same physician chooses differently based on prior therapy. Messaging designed for post-photopheresis patients may not translate to calcineurin inhibitor failures.
3rd-Line Choice by Prior Therapy
Same physicians. Same disease. Different prior therapy. Different decisions. N = 847 patients, 206 HCPs.
| 2L Pathway | Drug A | Drug B | Implication |
|---|---|---|---|
| Post-Photopheresis | 81.5% | 18.5% | Clear preference |
| Post-Calcineurin Inhibitor | 13% | 87% | Established pattern |
| Post-mTOR Inhibitor | 28% | 72% | Moderate opportunity |
What we measured: Patients progressing from the same 2L therapy are routed to different 3L agents based on perceived mechanism fit. Post-photopheresis patients go to Drug A at 4.4× the rate of post-calcineurin inhibitor patients.
Sensitivity Analysis
The pathway effect dominates other variables. Practice setting matters, but prior therapy matters more.
| Subgroup | Drug A Share | Drug B Share | Gap |
|---|---|---|---|
| Oral 2L patients | 22% | 78% | 56pp |
| Procedure-based 2L | 81.5% | 18.5% | 63pp |
| Academic centers | 34% | 66% | 32pp |
| Community oncology | 19% | 81% | 62pp |
Gap direction consistent (32–63pp) regardless of setting or geography.
The "2nd-Line Gauntlet"
Patients don't jump from 1L steroids to novel agents. They endure a months-long "clinical gauntlet" on legacy therapies—median 165 days on oral immunosuppressants, 124 days on photopheresis—before progressing. This creates a 4-6 month window of opportunity.
Why This Matters for Campaign Design
A single message won't work. Post-photopheresis physicians already prefer Drug A—they need reinforcement, not persuasion. Post-calcineurin physicians have an entrenched clinical algorithm—they need education that challenges their sequencing logic.
Beyond pathway, we also examined whether practice setting moderates these effects.
Academic vs. Community: A 15pp Swing
Academic centers show 34% Drug A share vs. 19% in community—a 15pp difference. But within each setting, the pathway effect still dominates. Post-photopheresis community physicians choose Drug A at higher rates than post-calcineurin academic physicians. Pathway trumps setting.
What This Means for Your Brand
This analysis demonstrates pathway-based segmentation. For clients, we can de-blind and go deeper:
Based on the 63pp pathway gap and current cGVHD incidence, a targeted campaign reaching the right HCPs at the right clinical moment could influence 120–180 prescribing decisions over 24 months. This is a directional estimate derived from observed prescribing patterns, not a clinical trial; actual capture depends on message effectiveness, access, and competitive dynamics.
Methodology
- Open medical + pharmacy claims on 330M+ lives
- Cohort: cGVHD diagnosis + systemic therapy, 2022-2025
- Line of therapy: first claim for new class
- Pathway assignment: claims sequence logic
- NPI taxonomy with manual validation
Limitations
- Observational—unmeasured confounders exist
- Pathway correlation is not causation
- Hospital outpatient may be undercounted
- Blinded for publication; de-blind in engagement
This is independent research. No manufacturer funded this analysis.
We can run this analysis for your specific brand and competitive set.
Book a Consultation →