Starting hepatitis C treatment in a single emergency-department visit — a phase 4 trial using a smart pillbox and an AI virtual agent to support adherence — a clinical trial (ClinicalTrials.gov)
A phase 4 trial of single-encounter test-and-treat: patients with hepatitis C seen in the emergency department are tested on the spot and, if eligible, given medication before they leave. A smart pillbox, an AI-powered virtual agent and community health worker support back up adherence.
Trial overview (primary data)
- StatusNot yet recruiting
- ConditionsHEPATITIS C (HCV)
- InterventionsDRUG: Epclusa (sofosbuvir and velpatasvir)
- SponsorPatrick Grace
- Target enrollment35 participants
- Period2026-12-01 〜 2028-09-01
Key points
- A phase 4 trial (target 35) of single-encounter test-and-treat: rapid testing in the emergency department and medication handed over before discharge if eligible.
- Adherence is supported by combining a smart pillbox, an AI virtual agent and community health worker outreach.
- AI here handles neither diagnosis nor treatment choice; it supports staying on the medication.
- Of the 803 AI-related clinical trials this site holds as of 2026-09-02, only 35 report a phase.
- This is a small feasibility stage; the approach has not been shown to be better than conventional referral.
1Putting treatment where the thread breaks
Hepatitis C is now curable in most cases with a short course of oral drugs. When people still fail to reach treatment, the obstacle is usually the pathway rather than the medicine. A positive test result has to be followed by booking a specialist appointment, attending it, and picking up a prescription — and contact is lost somewhere along that chain.
The emergency department is one of the few points where people without regular care touch the health system at all, and telling them a diagnosis before sending them home tends to end there. This trial tries to close that gap by delivering treatment within the same visit.
2Referring versus handing it over
What AI does here is neither diagnosis nor choosing a treatment; it sits on the side of helping people keep taking the drug. Medical AI usually brings imaging and prediction models to mind, but for a disease whose treatment is already settled, whether the medicine actually arrives and gets finished is what decides the outcome.
Combining a virtual agent, a pillbox that reveals adherence, and human outreach is notable precisely because it does not try to solve the problem with technology alone.
3One of the few trials that reports a phase
Of the 803 AI-related clinical trials this site holds as of 2026-09-02, only 35 report a phase from 1 through 4. Most are observational studies aimed at developing or validating a model, or evaluations of devices and software for which the notion of a phase does not fit. This trial is a phase 4 study — the stage that asks how an already-approved drug is used in real-world practice.
The phase label itself signals that what is being tested is the delivery design rather than the drug.
4Why it starts small
The enrollment target is 35, which is deliberately small. Single-encounter test-and-treat requires rapid testing, same-day dispensing inside an emergency department, and post-discharge follow-up to run at once, and the first question is whether that can hold together at all. What is described here is the design and objective of the study; it does not mean this approach has been shown to beat conventional referral.
Adherence and cure confirmed at follow-up will determine whether it moves to a larger scale.
Why it matters
For diseases with settled treatments, outcomes hinge on whether the drug arrives and is finished rather than on how well it works. Bundling testing, dispensing and adherence support at the emergency-department contact point is one answer to the question of public-health reach, and an implementation example that places AI in a supporting role.
FAQ
Why start treatment in an emergency department?
Does the AI decide the treatment?
What does phase 4 mean?
Sources (primary)
Source: ClinicalTrials.gov (U.S. NIH/NLM, public domain). This site does not provide medical advice. Verify the latest and exact details with the official source. This site is not endorsed or certified by the NIH/NLM.
- ClinicalTrials.gov (study record, original)
- NCT ID: NCT07793643