Tempus AI TIME / TriHealth Cancer Institute
Verified evidenceCreator to judgeContinuous population screening can connect eligible patients to a standardized activation lane.
Healthcare screening · United States
Collections: Regulated autonomy
Executive brief
The unit is a human-confirmed match connected to a pre-negotiated activation lane.
AI value · Trial activation time
Verified37 days TIME; 80% reduction
Journal of Clinical Oncology conference supplement abstract, not a full peer-reviewed article; nonrandomized comparison and bundled operating-model change.
Before
Screens available patients against non-TIME trial criteria and confirms possible eligibility. → Negotiate trial terms and complete conventional activation.
After
Re-runs eligibility searches when patient data or trial criteria change and produces candidate matches. → Nurse reviews algorithmic matches; site staff perform confirmatory screening before consent.
Human boundary
Humans confirm eligibility and consent; AI only proposes candidates.
Why it matters
Continuous population screening can connect eligible patients to a standardized activation lane.
Before
How the work ran before the change.
TriHealth research staff
Screens available patients against non-TIME trial criteria and confirms possible eligibility.
ControlHuman clinical confirmation; the comparison reports 186-day non-TIME activation but not each queue stage.
Site contracts and IRB teams
Negotiate trial terms and complete conventional activation.
ControlSite, sponsor and IRB approvals remain required.
What changed
Continuous population screening can connect eligible patients to a standardized activation lane.
Decision rightHuman moves from creator to judge
After
How the same work runs now.
TApp
Re-runs eligibility searches when patient data or trial criteria change and produces candidate matches.
ControlNLP match only; site capability and recency filters bound the queue.
Tempus nurse and TriHealth staff
Nurse reviews algorithmic matches; site staff perform confirmatory screening before consent.
ControlPre-negotiated rate card/agreement and central IRB are bundled with AI; patient consent remains human.
Exception path
Filtered or clinically ineligible matches are discarded.
Work removed
Decision authority
Humans confirm eligibility and consent; AI only proposes candidates.
Before
TriHealth research staff
Screens available patients against non-TIME trial criteria and confirms possible eligibility.
Control: Human clinical confirmation; the comparison reports 186-day non-TIME activation but not each queue stage.
Site contracts and IRB teams
Negotiate trial terms and complete conventional activation.
Control: Site, sponsor and IRB approvals remain required.
After
TApp
Re-runs eligibility searches when patient data or trial criteria change and produces candidate matches.
Control: NLP match only; site capability and recency filters bound the queue.
Tempus nurse and TriHealth staff
Nurse reviews algorithmic matches; site staff perform confirmatory screening before consent.
Control: Pre-negotiated rate card/agreement and central IRB are bundled with AI; patient consent remains human.
Work that left the path
Human role before
Research staff screened patients against trial criteria, while site contracts and IRB teams completed activation handoffs; investigators retained eligibility and consent decisions.
Human role after
Tempus nurses and site staff confirm eligibility; patients consent; IRB and contracts govern activation.
AI roleRe-runs NLP eligibility searches when patient or protocol data changes.
186 days non-TIME→37 days TIME; 80% reduction
2022-01 to 2023-03 · 18,823 patients; 135 trials; 8.65 million searches
Journal of Clinical Oncology conference supplement abstract, not a full peer-reviewed article; nonrandomized comparison and bundled operating-model change.
Anti-pattern
Do not attribute bundled contracting gains solely to AI.
Questions
Portability conditions
Reputation risk
low
Evidence and authority
Current · updated
1 independent; publication outcomes are verified.
Bundle 1.0.0 · reviewed 2026-08-23 · stable ID 1f75e0019e001676
The Permanente Medical Group · Creator to judge
Ambient tools can draft clinical notes, while physicians remain the editors and signers.
UCLA Health · Creator to judge
Ambient tools can draft notes in practice, but vendor-specific efficiency and error results require comparison.
United Kingdom Serious Fraud Office · Creator to judge
AI can perform first-pass privilege screening at scale while lawyers retain privilege and prosecution judgments.
Google · Creator to judge
Large migrations can combine machine-found locations, LLM edits, automated checks, and engineer review.