HugoScore hugoscore.org

Full review

TrialScreen CAIHL draft report

Evidence-linked HugoScore draft report for a health AI tool that affects patients.

Agency checklist · October 30, 2026

Mixed 6 of 12 yes

Document-based draft. Independent confirmation is incomplete. Hugo has not personally verified this service.

Assessed function: TrialScreen public clinical trial search and automated eligibility prescreening for patients; sponsor-managed recruitment (REFER) and site enrollment decisions assessed only where they touch the patient

A no can mean missing documentation or institution-dependent control; it does not by itself establish harmful behavior. These categories are not clinical safety or product-quality ratings.

  • Public documents do not establish actual product behavior.
  • The independent review screened rules and selected evidence; it did not replicate all twelve answers for this tool.
See all twelve answers and their evidence
QuestionAnswerEvidence / reviewSource and reasoning
1. Told it is AINonot-documented
Not independently replicated
Quotation not recorded; evidence repair pending. Source

Privacy says 'the eligibility checker is an automated process and the outcome is based on the information you provide and the eligibility criteria of the trial'. No public page says AI or machine learning is used. It may be rule-based; if so this question may be not applicable.

2. Operator and model maker namedNonot-documented
Not independently replicated
Quotation not recorded; evidence repair pending. Source

TrialScreen Pty Ltd and TrialScreen Inc. are named in the privacy policy contacts. No AI model or maker is named, and none is disclosed.

3. Patient chooses to use itYesdocumented
Not independently replicated
check your own fit before you share a single contact detail Source

Patient searches and chooses whether to run a check and whether to share contact details. Sponsors page: 'Sharing details isn't a requirement, it's a choice.'

4. Patient sets the taskNodocumented
Not independently replicated
TrialScreen is a clinical trial awareness and lead generation company. Source

Judgment call. Patient chooses what to search, but the operator's stated business is sponsor lead generation, paid per contactable lead, and the flow routes patients to study teams. Operator goal shapes the function.

5. Patient can decline or stopYesdocumented
Not independently replicated
check your own fit before you share a single contact detail Source

Self-started rule applied (cites question 3 source).

6. Patient sees the AI outputYesdocumented
Not independently replicated
based on your answers, you may not meet the criteria for any of your saved studies right now Source

Eligibility result is shown to the patient in the site's own UI strings (public JS asset). Also 'You may be eligible - contact the study team'. Quote is from the English string in the bundle.

7. Patient can correct or challengeNonot-documented
Not independently replicated
Quotation not recorded; evidence repair pending. Source

Privacy offers access and correction of personal information held. No documented route to contest the automated eligibility result.

8. Patient can get a copyYesdocumented
Not independently replicated
You may request access to the personal information we hold about you Source

Privacy also lists a right to data portability, depending on location.

9. Patient can deleteYesdocumented
Not independently replicated
you can change your mind and ask us to remove your details Source

Applies where consent is the basis. If the patient was connected to a trial, the request is forwarded to the study team.

10. Training or sale does not override patient choiceYesdocumented
Not independently replicated
We do not sell your personal information. Source

No model training is disclosed (no AI documented). Sharing with sponsors or sites is 'only with your consent or where reasonably expected'. Pay-per-lead model is covered in question 12.

11. Patient can run it or take it elsewhereNonot-documented
Not independently replicated
Quotation not recorded; evidence repair pending. Source

Vendor-hosted only.

12. No advertising or data businessNodocumented
Not independently replicated
You only pay per contactable lead, we take on the design and advertising risk. Source

Sponsors pay per lead of consenting patients. Applying the rule that data business counts even with consent. About page: 'clinical trial awareness and lead generation company'.

Initial coding: Claude Sonnet helpers, compiled by Claude Opus 5.5 (Anthropic). Selective review and corrections: GPT-6 Astra (OpenAI). Editor: Hugo Campos. Acceptance of corrections is not product verification.

Checklist rules · Independent review and limitations · Later evidence follow-up

Earlier report

TrialScreen: CAIHL reassessment

September 8, 2026. Published AI-assisted draft.

TrialScreen offers trial search and automated prescreening. A match is not a final enrollment decision. Current policy permits some disclosures beyond express consent alone, and removal of information already sent to study teams depends on those teams.

Scope and agency posture

Public trial search and automated prescreening, distinct from sponsor-managed recruitment services and ultimate study enrollment decisions.

Posture: Mixed, open patient search funded by sponsor recruitment.

Axis: 55/100, retained provisionally. No numerical recalibration was performed.

Sources and documented findings

  • The researcher page describes managed-trial criteria, AI-assisted registry criteria, deterministic matching and recruitment lead generation. These document recruitment objectives but do not establish paid ranking of search results. https://info.trialscreen.org/Researchers
  • The privacy policy calls eligibility checking automated, provides access/correction requests and forwards post-connection removal requests to study teams. It permits sponsor/site disclosure with consent or where reasonably expected and health-data processing with explicit consent or other lawful permission. A privacy-email placeholder remains. https://info.trialscreen.org/Privacy

Patient authority

Inference from the documented workflow: Historical no-account search gives users room to explore. Screening may narrow perceived options, so correction of input and access to researchers matter more than free access alone.

Critical capacity

Editorial assessment: Trial discovery supports action. The reviewed public pages do not establish an explanation or appeal path for false-negative screening, so do not equate a match result with final eligibility.

Informed control

Editorial assessment: The baseline's consent-only sharing language is stronger than the current policy, which contains additional permitted bases. Downstream removal depends on study teams.

Assessment and limits

The Mixed interpretation remains reasonable but needs precise privacy wording. Recruitment funding alone is not a finding that patient interests are subordinated.

Confidence: Moderate for published screening and data terms.

Remaining uncertainty: Search ranking, sponsored labels, explanation of exclusions and patient screening-data retention remain unverified

Published documentation is evidence of stated conditions, not proof of actual implementation. Unknowns did not receive automatic negative points. Funding, sponsorship, and public code do not determine agency by themselves.

Review provenance

  • Reviewer/model: OpenAI Codex / GPT-6.
  • Method: Focused public-source reassessment using CAIHL: patient authority, critical capacity, and informed control. Existing evidence plus one focused primary-source pass and at most one targeted follow-up. No live product testing. Numerical scores remain provisional editorial placements, not a new calculation.
  • Human review: Hugo Campos authorized publication of these AI-assisted draft reassessments on September 8, 2026. This does not claim comprehensive human verification of every finding.
  • Earlier review: 2026-06-10. Historical assessment. Earlier claims are not automatically reverified by this publication.