How We Build An Enrollment Number
Why do prevalence-based numbers mislead?
Most inflated feasibility comes from two mistakes. First, starting from disease prevalence instead of patients a site actually treats. Second, applying eligibility criteria simultaneously instead of sequentially: a panel where 40% meet criterion A and 50% meet criterion B does not contain 45% eligible patients, because criteria interact. Honest funnels apply every filter in order, each one operating on the survivors of the last.
What goes into a Stryde projection?
| Step | What we use | Why it matters |
|---|---|---|
| 1Active patient count | Tagged counts from the practice EMR, with severity markers | Real patients the investigator treats, not census arithmetic |
| 2Sequential protocol filters | Each inclusion and exclusion criterion applied in order | Criteria interact; sequence keeps the number honest |
| 3Screen-failure rate by study type | Roughly 25 to 30% for oral therapies, 35 to 45% for injectables, 45 to 55% for imaging-confirmed indications | Estimates that skip this are fiction with a confidence interval |
| 4Coordinator capacity ceiling | 12 to 15 screenings per month per coordinator, applied as a hard limit | An eligible pool of 400 means nothing at a site that can process 15 candidates a month; this is the constraint most feasibility ignores |
| 5Competition adjustment | Roughly 15% reduction per competing study for an overlapping population | Your study is not the only one asking for these patients |
A worked example
A moderate-to-severe plaque psoriasis protocol at one of our dermatology sites. The practice sees 1,200+ psoriasis patients annually. Chart-confirmed filters for BSA of 10% or more, PASI of 12 or more, and IGA of 3 or more reduce that to roughly 145. Applying reachability, therapy-history exclusions, an injectable-class screen-failure rate, and one coordinator’s screening capacity produces a commitment of roughly 45 to 50 enrollments over twelve months at that site. That is the number that goes in the questionnaire, and the site is expected to beat it.
What we will not claim
- Guaranteed enrollment.
- Numbers built from prevalence.
- Capacity we have not staffed.
Where a protocol does not fit our sites, the feasibility response says so within the same 48 business hours.