An observational study in History Taking and Artificial Intelligence (AI), sponsored by Beth Israel Deaconess Medical Center. Enrolling by invitation at 1 site in United States. Open to participants aged 18 Years and older. Per ClinicalTrials.gov, last updated 2026-10-16.
Sponsored by Beth Israel Deaconess Medical Center · Observational
This study evaluates whether an artificial intelligence (AI) conversational agent called MINERVA (Medical Interface for Navigating the Emergency Room with Voice Assistance) can safely and feasibly gather medical histories from emergency department (ED) patients and produce accurate summaries. The main questions it aims to answer are:
Can an AI conversational agent gather and summarize patient histories in the emergency department?
Are the resulting summaries accurate? Do they capture essential clinical information?
Can an AI agent overseeing these conversations independently identify red-flag symptoms that may need time-sensitive clinical interventions?
Adult ED patients interact with MINERVA via either voice chat or typing on a keyboard. MINERVA asks them questions about their symptoms and generates a summary which the patients review for accuracy. MINERVA reviews the conversation for clinical red flags that suggest urgent evaluation is required. Participants fill out a survey providing feedback on their experience. The summary is not saved in the medical record and clinical care is not affected.
Physician reviewers will evaluate the safety, accuracy, and clinical value of the MINERVA interview transcripts and summaries using standardized rubrics, including whether MINERVA was correct in identifying red-flag symptoms that may need time-sensitive clinical interventions.
This is a single-arm prospective observational cohort study evaluating safety, feasibility, and acceptability of the MINERVA conversational agent for gathering and summarizing patient histories in the emergency department. ED patients who have already been assessed by a triage nurse will be identified by research assistants (RAs) and verbally consented. MINERVA will ask the patient questions about their symptoms, medical history, and risk factors, as well as clarifying questions to highlight pertinent positives and negatives or red flag symptoms. MINERVA will generate a summary at the end of the conversation which patients may edit. Patients will complete a survey after finishing the interaction.
After the interaction, MINERVA will determine whether there are any red flags (called Safety Diagnoses below) that would require urgent evaluation. These red flags will be compared to the assessments of expert physicians. Additionally, MINERVA will generate a differential diagnosis from the interview for clinician review. Descriptive statistics will be calculated for survey responses. MINERVA transcripts and summaries will be evaluated by physician reviews for accuracy, completeness, clinical utility, and safety according to standardized rubrics
1,382 studies on the registry are indexed under Disease; 612 are open to participants now.
This study's planned enrollment of 250 is above the median of 209 across 598 observational studies indexed under Disease.
Browse Disease studies →Beth Israel Deaconess Medical Center is the lead sponsor of 565 studies on the registry; 81 are open to participants now.
Of its 75 completed or terminated interventional studies of FDA-regulated products, 62 (83%) have results posted.
Counted across the registry records on this site, refreshed daily.
Adult emergency department patients
Go through the criteria one by one — yes, no or not sure — and take the list to your doctor or the study team. Kept in this browser only.
Exclusion Criteria:
All participants are in this arm, in which participants provide history to the MINERVA AI Chatbot
Other: MINERVA AI Chatbot
The MINERVA AI chatbot is AI software that allows patients to provide medical history through text or voice.
Safety Diagnoses
Emergency physicians will review each patient-AI interview transcript using a standardized rubric to determine whether evaluation for possible myocardial infarction, acute stroke, or sepsis is warranted. MINERVA's assessments will be compared with these physician assessments.
Time frame: 1 Day
Secondary Safety Diagnoses
Emergency physicians will review each interview transcript using a standardized rubric to determine whether evaluation for diagnoses including major traumatic injury, aortic syndromes, ovarian/testicular torsion, anaphylaxis, behavioral emergency, major bleeding, pre-eclampsia, other pregnancy emergencies, spinal cord compression, limb ischemia, toxic ingestion, or sickle cell crisis are warranted. MINERVA's assessments will be compared with these physician assessments.
Time frame: 1 Day
Conversation completion
Whether interview was completed, terminated early, or patient declined to participate.
Time frame: 1 Day
Conversation length
Length of patient interaction with AI conversational agent, measured in number of turns.
Time frame: 1 Day
Turn Length
Length of turns for patient and AI conversational agent, measured in seconds and words.
Time frame: 1 Day
Patient Acceptability of AI Conversational Agent
Patients will complete an ED Conversational Agent Patient Impressions (ED-CAPI) questionnaire before and after interacting with the chatbot, including components of the following questionnaires: * General Medical Counsel Patient Questionnaire * Practical Assessment of Clinical Examination Skills * Articulate Medical Intelligence Explorer post-provider survey * General Attitudes towards Patient-Facing AI Scale Further questions include: * Ease of use * Perceived value and effects on healthcare system * Willingness to share information with the AI * Accuracy of the AI summary of the interview Responses are aggregated with percent of maximum possible (POMP) methodology. All survey questions are rescaled to a 0-100 scale with 0 the worst possible impression and 100 the best. The final outcome is the 0-100 average of all responses.
Time frame: 1 Day
Physician Acceptability of AI Conversational Agent
The panel of ED physicians will review interview transcripts and summaries generated by the AI conversational agent. They will complete an ED Conversational Agent MD Acceptability (ED-CAMA) survey to measure perceived utility of the tool in their clinical workflow.\[
Time frame: 3 Months
Primary Diagnosis
The primary diagnosis managed during the patient's emergency department encounter will be assessed via review of emergency department notes as well as admission history and physical and discharge summary when available. Bond score will measure AI diagnostic accuracy relative to the ground truth diagnosis.
Time frame: 1 Day
Plan to share: No — The primary data collected by the study is free text interview transcripts which contain private health information.
Start is estimated for Oct 2026; primary completion for Jan 2027, 3 months later.
Of 282 non-industry observational studies under Disease, half reached primary completion within 53 months of starting (the median), counting those still running; a quarter within 26. This study's planned 3 months is below that.
Counting every sponsor, half of 307 studies reached it within 51 months; a quarter within 26.
Disease covers many different conditions, not one, so the studies compared here vary widely. Read these figures as a rough guide.
How these figures are worked out, and which studies are left out of them.
From the registry record's own update history. This site started tracking changes on Sep 25, 2026; for anything earlier, see the record history on ClinicalTrials.gov ↗
Eligibility is decided by the study team. Share this record with your doctor or contact the team directly.
No contact was published for this record. The registry link below has the sponsor’s details.
Get an email when the registry record changes — status, dates, results — or when someone posts here.
Sign in to followQuestions and observations about this study, from anyone following it. Not medical advice, and not a channel to the study team — their contact details are on the registry record.
Sign in to join the discussion. Reading takes no account; posting does. You choose a display name, and a pseudonym is the default.
Nothing here yet. If you are running this trial, taking part in it, or weighing whether to, this is the place to say so.
Beth Israel Deaconess Medical Center →