Site selection

Plan the next phase footprint from peer trials in your indication: countries, active facilities, the CROs that ran recent trials, and a benchmark timeline to first patient in. Fill the plan by hand, from a live readout, or with the AI planner.

Methodology (v1)

Peer set: same primary condition as the anchor trial, filtered to benchmark phases (by default, Phase 3 when the anchor is Phase 2), plus any NCT IDs you add manually.
Data: facility and country fields from ClinicalTrials.gov location modules; enrollment from study design when reported.
Country score: weighted mix of peer trial count, distinct facilities, recruiting peers, and average enrollment per listed site across peers.
Compare A vs B: two scenario configs on the same anchor; rank delta shows how country priority shifts when weights or peer sets change.
Live readouts: industry-sponsored Phase 2 trials on ClinicalTrials.gov with primary completion from 60 days ago to 6 months ahead, at least 8 listed sites in 2 or more countries. One click fills the anchor and two scenarios (balanced vs recruiting-now).
CRO landscape: EU CTIS public records list the third parties each sponsor delegated duties to (on-site monitoring, data management, IRT, laboratory, EDC). We scan the most recent CTIS trials in the indication and phase, then add CRO names found in CT.gov collaborator fields of the peer set. CT.gov rarely names CROs, so absence is not evidence. CRO tiers are public positioning, not quality ratings.
Timeline: anchored on the registry primary completion date. Durations are published benchmarks: FDA Type B End-of-Phase 2 meetings within 70 days of request, CRO selection 8 to 12 weeks (Halloran), site selection about 3.2 months and start-up 5 to 6 months (Tufts CSDD). The parallel track starts CRO selection before topline.
AI planner: fills the same form fields you can edit, then triggers the deterministic analysis. The model does not produce scores; it summarizes page outputs. Requires sign-in.
Limitations: registry site lists are incomplete and lag protocol amendments; no patient counts or investigator master data. For competitor-aware refresh, pair with trial alerts and portfolio watches.

Full design and vendor landscape: docs/site-selection/design.md in the repo.