Research Methodology

Tidera Health is an educational clinical trial discovery platform. We help patients learn about recruiting clinical trials that may be relevant to their condition — and understand what those trials are actually looking for, in plain English. We are not a medical provider, an eligibility determination service, or a patient recruitment company. We exist solely to reduce the information asymmetry between patients and the research ecosystem.

All trial data comes directly from ClinicalTrials.gov — the official U.S. registry of federally and privately supported clinical studies, maintained by the National Library of Medicine (NLM) at NIH. We use the ClinicalTrials.gov v2 REST API (ctg-api.clinicaltrials.gov) to fetch live trial records in real time. We are not affiliated with NLM or NIH. Trial records are fetched fresh on every search — we do not serve stale cached data older than the current request.

When you submit a search, Tidera Health issues a real-time query to the ClinicalTrials.gov API using your condition and location. Queries are filtered to: • Status: RECRUITING only (we exclude completed, terminated, and not-yet-recruiting trials) • Condition match: keyword-based condition query against the official trial registry • Location proximity: trials are filtered or de-prioritized by distance from your location No results are pre-cached for your specific profile. Every search produces a live, current snapshot of what ClinicalTrials.gov knows at that moment.

Before AI explanations are generated, a deterministic pre-ranking engine scores each trial across several dimensions: 1. Recruitment status weight — actively recruiting trials scored higher 2. Trial phase — later-phase (Phase 3, Phase 4) trials receive a modest boost 3. Location proximity — trials at closer sites score higher when location is provided 4. Condition specificity — trials with a direct condition match rank above broader studies 5. Prior therapy signals — when medication history is provided, trials with relevant inclusion/exclusion criteria (e.g., "must have failed prior biologic therapy") are surfaced more prominently This pre-ranking step is rule-based and fully deterministic. It runs before the AI layer to ensure the most relevant subset of trials is prioritized for explanation.

After pre-ranking, Tidera Health uses Claude (Anthropic) to generate plain-English summaries of each trial's eligibility criteria. The AI layer produces: • A plain-English summary of what the trial is studying • A fit score (1–10) estimating alignment between your profile and the trial description • Key inclusion/exclusion factors explained in accessible language • Questions you might ask a research coordinator • Medication considerations where relevant Important: AI-generated fit scores and explanations are educational estimates only. They are NOT eligibility determinations. The AI reads trial text — it does not adjudicate whether you qualify. Only the trial's research team can determine eligibility.

To be explicit about our limitations and boundaries: • We do NOT confirm eligibility for any trial • We do NOT provide medical advice of any kind • We do NOT store personally identifiable health information (PHI) • We do NOT facilitate enrollment — all enrollment happens directly with research teams • We do NOT rank based on commercial relationships with sponsors • We do NOT share your health information with third parties

Clinical trial data can lag real-world status. A trial may appear as "Recruiting" in the registry even if enrollment has temporarily paused. Trial eligibility criteria are often written in dense clinical language that can be ambiguous — our AI interpretations may not capture every nuance. Always verify trial status, eligibility, and site availability directly with the research coordinator before making any decisions. Use Tidera Health as a starting point for discovery, not as a definitive guide to your options.