Clinical trial site selection, every site ranked with a cited reason
Every country and site for a new study ranked with a cited reason, and the list approved in 11 days.
See one case, screen by screen ↓The questionnaires go out to 186 sites in 14 countries. What comes back is a mix of portal answers, PDFs and email replies — and every site says it has the patients. The feasibility lead has to check each claim against what the site actually screened on the last study, pull its start-up weeks out of the CTMS, and search the registries for every same-population trial recruiting nearby.
That is done in spreadsheets, one site at a time, while country managers push for the sites they know. The rationale for each choice ends up in someone’s head or an email thread, not in the trial master file. And a famous investigator with a weak record still gets a primary slot — until the site enrols three patients against a plan of ten.
Estimated split for a Phase 3 study with about 90 usable questionnaires, by spreadsheet and with the solution.
Six specialist agents read the protocol, weigh patients, competing trials, site history and questionnaires side by side, then rank and explain; the feasibility lead signs the list.
VLT-401 is a Phase 3 study of veltrimab in Crohn’s disease, 480 participants. The questionnaires closed on Oct 2 and site list v3 is in approval. Here is how the list gets signed, screen by screen, in the working solution.
131 of 186 questionnaires back, 64 primary sites and 12 back-ups proposed from 90 scored, 12 countries plus 2 back-up. The most likely last patient in is May 2028 — 7 weeks before the June 2028 target. And 24 sites claim more than 3× their screening history.
Needs you: “Universitätsklinikum Rheinhafen — Tom Becker asks to make it primary · agents propose back-up.”
Poland leads with a score of 88: 6,900 eligible patients a year, 5 of 14 competing trials in the same population, EU CTR Part II in 76 days and a 13-week start-up. Brazil has 11,500 eligible a year but ANVISA plus national ethics takes 150 days, so it is held as back-up with three sites ready.
“Eligible patients: licensed real-world counts with the VLT-401 criteria applied · registries refreshed Oct 5.”
Each site is scored on six weighted factors: past enrollment 30 %, start-up speed 15 %, patient pool 20 %, competition 15 %, questionnaire quality 10 %, representation 10 %. Centrum Gastrologii Mazowsze in Warsaw is first at 85 — 1.44 participants a month, an 11-week start-up, and a claim consistent with its history at 0.9×.
Lena picks “Fast start-up”: start-up speed goes to 35 %, past enrollment to 25 %. Centrum Kliniczne Wielkopolska, with an 8-week start-up, climbs to #2. The live forecast follows — 16 sites open by Feb 2027 instead of 15, 4 proposals change, and the last patient moves a week later, because the faster sites enrolled less last time.
“Move a weight — agents re-propose inside each country’s allocation; decisions made by people stay.”
“Please make Rheinhafen a primary site. Prof. Albers leads the national IBD society and other sites follow his lead.” The rationale beside it ranks the site #85 of 90: in VLT-208 it enrolled 3 participants against a plan of 10 over 12 months, and start-up took 29 weeks, 13 of them on the contract.
“The questionnaire reports about 80 eligible patients a year; the site screened 14 in a year on VLT-208, so the claim is 5.7× its history. We count 28 a year.”
BRK-17 recruits at the hospital itself; LUM-402, KST-112 and ARB-310 sit 17 to 20 km away — each a public registry entry, checked Oct 5. The site’s questionnaire names only two. SOL-44, a post-operative recurrence study 39 km away, is shown but not counted.
Different-population studies never count against a site.
Lena keeps the agents’ proposal and confirms it with a reason: “Keep as back-up: 3 of 10 enrolled in VLT-208, 29-week start-up and four competing trials nearby. Opens if a German primary site falls behind after month 6.” The reason goes to Tom and into the audit trail. Moments later Tom signs Germany, the UK and France.
“Tom Becker signed — all five country managers are done.”
64 primary and 12 back-up sites in 12 countries; the country allocation already signed by Sam Patel, study lead. From 2,000 simulations: P10 Feb 2028, P50 May 2028, P90 Sep 2028. Lena signs with her password and the meaning of the signature — “I approve this site list and the rationale for each site.”
“Poland and the US carry 46 % of the plan; losing one lead Polish site moves P50 by about 3 weeks.”
Four records land in site selection, where Dana reads them: the site selection rationale for 76 sites with every sentence cited, the country allocation and site list, the P10–P90 enrollment forecast, and the approval record with signatures, meaning and time. The two changes country managers made, and Lena’s decision on Rheinhafen, are on the list with their reasons.
Amendment 1 lowers the endoscopic entry score to SES-CD ≥ 4 and allows up to four prior advanced therapies. In 3 min 12 s the agents re-read the criteria, recount the pool — Poland +17 %, the US +13 % — re-score all 90 sites and propose 3 changes. The P50 moves from May to Apr 2028. Nothing changes until Lena applies it.
The same study, seen by the four people who carry it — what site selection looked like for them, and what it looks like now.
Six specialist agents read the protocol, weigh patients, competing trials, site history and questionnaires side by side, then rank and explain; the feasibility lead signs the list.
Reads the synopsis or amendment and extracts the population, every inclusion and exclusion criterion, procedures that need site equipment, visit burden and target size.
Applies the structured criteria to licensed real-world counts per country and region, keeping the funnel from diagnosed to eligible a year.
Reads public registries for recruiting studies in the same population — per country and within 25 km of each site — matching on condition, phase, line of therapy and age.
Reads the trial management system: enrolled against plan, screened, screen failure, start-up weeks split into contract, ethics and activation, data quality and inspections.
Reads every returned questionnaire — portal, PDF or email — checks completeness, compares the eligible-patient claim with what the site screened, and drafts follow-up questions.
Scores each site on six weighted factors, writes a cited rationale, proposes primary, back-up or not selected inside each country’s allocation, and forecasts enrollment from 2,000 simulations.
Signs the list and allocation. The agents propose; a named person decides.
Ask about any country, site or forecast — or tell it what to change. Every answer points to the questionnaire, the CTMS record or the registry entry behind it.
Why is Rheinhafen only a back-up?
Three reasons. In VLT-208 it enrolled 3 of 10 planned in 12 months. The questionnaire claims 80 eligible a year — 5.7× what it screened; we count 28. Four Crohn’s trials recruit within 25 km, one at the hospital itself. Tom Becker asked to make it primary — that decision is yours.
Which sites overstate their patient numbers?
24 sites claim more than 3× the patients they screened per year on their last study. Each is counted at 2× its history until the site confirms, and follow-up questions are drafted.
Give start-up speed more weight
Done — start-up speed now counts 35 % (was 15 %), past enrollment 25 %. Inside each country’s allocation, 4 proposals change. 16 sites open by the Feb 2027 first-patient target (was 15), but the faster sites enrolled less on past studies, so the last patient moves a week later. Decisions made by people are kept.
When will we finish enrolling?
With the 64 primary sites, the last patient is most likely in May 2028 — 7 weeks before the June 2028 target. Optimistic (P10) Feb 2028, cautious (P90) Sep 2028, from 2,000 simulations of when each site opens and how fast it enrols. Poland and the US carry 46 % of the plan.
13 screens from the working solution, on its sample data. Pick one to see it large.
How sites are scored, who approves and where the evidence comes from are settings — every change recorded in the audit trail.
| Setting | Default | Choose from |
|---|---|---|
| How the ranking is weighted | Balanced | Balanced · Proven sites · Fast start-up · Representation, or your own weights |
| Count questionnaire claims at most 2× screening history | On | On · Off |
| Radius for competing trials | 25 km | 15 · 25 · 50 km |
| Ask for follow-up below this questionnaire completeness | 90 % | 80 · 90 · 100 % |
| Representation counts in the score | On | On · Off (sets its weight to 0) |
| Re-score weekly when registries change | Mondays 06:00 | On · Off — told when a site moves more than 5 places |
| Who signs the country allocation | Sam Patel · Study lead | Sam Patel · Dana Okafor |
| Who signs the final site list | Lena Ortiz · Feasibility lead | Lena Ortiz · Sam Patel |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample study data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (Tufts CSDD — about 11 % of activated sites enrol no one). People, companies and products named on this page are fictional — characters and sample data in the working solution.
Software that helps a sponsor decide which countries and sites to open for a study — combining site performance history, feasibility questionnaire answers, competing trials and patient pool estimates. The Country & Site Feasibility Analyst adds agents that read all of it, rank every site with a cited rationale and forecast enrollment, while people make the decisions.
Each site is scored 0–100 on six factors — past enrollment, start-up speed, patient pool, competition, questionnaire quality and representation — and combined with the study’s weights. Move a weight and the ranking, the proposals and the forecast follow; decisions people have made stay.
Yes. The questionnaire reader compares each site’s eligible-patient claim with what it screened on its last study and reports the ratio. By default a claim counts at most 2× the site’s screening history until the site confirms, and follow-up questions are drafted for a person to send.
A site with no past study for the sponsor gets a neutral history score, never zero, and lower confidence. Its questionnaire claim counts at 60 % until confirmed, and new answers from the site can be used to re-score it.
From 2,000 simulations of when each site opens and how fast it enrols, based on its own history and start-up. The result is a range — optimistic (P10), most likely (P50) and cautious (P90) last patient in — shown against your target date.
Yes. The agents only propose. Country managers review their own sites and can swap or add one with a written reason, the study lead signs the allocation, and the feasibility lead signs the final list with the meaning of the signature recorded.
Add the amendment and re-run. The agents re-read the criteria, recount the patient pool, re-score every site and show what changes — eligible patients, proposals and the forecast — before anything moves. Nothing changes until a person applies it.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your questionnaire, scoring weights, approval steps and CTMS history — and it goes live once every quality gate has passed. We will walk you through it on one of your own past studies first.
A trial synopsis becomes a checked first-draft protocol in days, in your own approved language.
1-min filmTMF CheckEvery trial document is read, filed in the right place and checked, so the file is inspection-ready every day.
1-min filmCSR WriterResults sections and patient narratives are drafted within days of database lock, every number linked to its table.
1-min filmWe’ll run Country & Site Feasibility on a past study of yours and compare its ranking with how the sites actually enrolled.