Artificial Intelligence for Mental Health Diagnoses and Treatment Plans: A Protocol for Actors and Patients

Artificial intelligence (AI) is becoming prevalent in modern medicine and psychiatry. AI is based on a wide variety of computer algorithms classified under machine learning (ML).

The purpose of the present study is to evaluate the potential for mental health diagnosis using AI.

In the first part of the study, the AI will conduct an interview with standardized patients [SP] (actors) presenting a psychiatric illness. The AI will present a differential diagnosis and treatment plan. Immediately afterward, the actors will be interviewed by a board-certified psychiatrist, who will also give a differential diagnosis and a treatment plan. The results of the AI and psychiatrist will be compared.

In the second part of the study, AI will examine patients coming for consultation by a psychiatrist in the inpatient units, outpatient units, or in the emergency room (ER) at Sheba Medical center. The AI results will be compered to the psychiatrist diagnosis.

Study Overview

• Status: Recruiting
• Conditions: Mental Health Diagnosis
• Intervention / Treatment: Diagnostic test: SP patients-actors; Diagnostic test: Patients; Diagnostic test: Control group- patients

Detailed Description

Procedure:

In the firs part of the study, The AI will conduct an interview with standardized SP patients (actors) presenting a psychiatric illness. The actors will be trained in one of 5 scenarios: depression, anxiety, PTSD with cannabis abuse, psychosis, and control patients with no diagnosis. Each scenario will have 2 levels of severity. The actors will present demographic and medical backgrounds, symptoms and signs, including disorders of thought and affect. based on the interview, the AI will present a differential diagnosis and treatment plan. Immediately afterward, the actors will be interviewed by a board-certified psychiatrist, who will also give a differential diagnosis and a treatment plan. The results of the AI and psychiatrist will be compared.

In the second part of the study, AI will examine patients patients treated in the Division of Psychiatry, patients who applied for treatment in the psychiatric clinic but are not yet treated there, and patients treated in the psychiatric emergency room (ER) in Sheba Medical center. In additional, a recruitment ad will be published in the media and on social networks, to include patients who will come in specifically to participate in the study.

The patient will give informed consent and will be interviewed by AI. The interview will be done either digitally or will use voice-to-text and text-to-voice technology that will allow the patient to speak with the computer instead of typing questions into the computer. Future studies, not included in the current proposal, will use voice analysis, face recognition, and computerized emotion recognition as part of AI.

Participants:

• Actors SPs (n=10), each time employed by "MSR"- the Israeli Center for Medical Simulation, trained to simulate symptoms of psychiatric illnesses.
• Board-certified Psychiatrists from the Drora and Pinchas Zachai Division of Psychiatry in the Sheba Medical Center.
• Patients (n=150) examined in the psychiatric division and emergency room (ER) of Sheba. Or patient that recruited from the media.
• Control group (n=50) Patients who came to the ER at the hospital due to physical problems, and no psychiatric history. They will be offered while waiting for the doctor to participate in the study and be examined by the AI and a psychiatrist.

Outcome:

The primary outcome of the study will be a comparison of the interviews, differential diagnoses, and recommendations for treatment assigned by the AI with those assigned by the board-certified psychiatrist. The rates of agreement will be the outcome of interest.

• Study Type: Interventional
• Enrollment (Estimated): 220
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Mark Weiser, M.D; Phone Number: +972526666570; Email: Mark.Weiser@sheba.health.gov.il
• Study Contact Backup: Name: Hadasa Afgin, B.A; Phone Number: +972552709699; Email: Hadasa.Afgin@sheba.health.gov.il

Study Locations

• Israel
  • Ramat Gan, Israel
    • Recruiting
    • Sheba Medical Center, Psychiatry Department
    • Principal Investigator: Mark Weiser, MD
    • Contact: Mark Weiser, MD; Phone Number: +972-3-5303773; Email: mweiser@netvision.net.il
    • Contact: Hadasa Afgin, B.A; Phone Number: +972552709699; Email: Hadasa.Afgin@sheba.health.gov.il

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: Yes

Description

Inclusion Criteria:

For the SPs part:

• Actors SPs employed by "MSR"- the Israeli Center for Medical Simulation, trained to simulate symptoms of psychiatric illnesses.
• Over the age of 18
• Sign an informed consent

For the patient part:

• Patients arriving for intake at the psychiatric clinic, the various psychiatric departments and the ER at the Sheba Medical Center.
• Patients over the age of 18
• Sign an informed consent

Exclusion Criteria:

• Patients participating in another study
• Patients under hospitalization order or court order

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Diagnostic
• Allocation: Non-Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

3

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: SP patients-actors; The actors (n=10) will be trained in one of 5 scenarios: depression, anxiety, PTSD with cannabis abuse, psychosis, and control patients with no diagnosis. Each scenario will have 2 levels of severity. The actors will present demographic and medical backgrounds, symptoms and signs, including disorders of thought and affect. based on the interview, the AI (LIV) will present a differential diagnosis and treatment plan. Immediately afterward, the actors will be interviewed by a board-certified psychiatrist, who will also give a differential diagnosis and a treatment plan.	Diagnostic test: SP patients-actors; SP Actors employed by MesserMSR, the Israeli Center for Medical Simulation, trained to simulate symptoms of psychiatric illnesses.
Experimental: Patients; The patients (n=150) will be recruited for the study from patients who applied to the psychiatric clinic in Sheba, new patients, old patients and patients on the waiting list. In addition, patients from the various psychiatric departments and the psychiatric and general ER at the Sheba Medical Center - Tel Hashomer. In addition, a recruitment ad will be published on social networks.	Diagnostic test: Patients; Patients (n=150) examined in the psychiatric division and emergency room (ER) of Sheba Medical center. Or people who applied following an advertisement on social networks.
Other: Control group- patients; Control group (n=50) Patients who came to the ER at the hospital due to physical complaints, and no psychiatric history. They will be offered to participate in the study and be examined by the AI (LIV) and a psychiatrist while waiting for the ER physician. In addition, a recruitment ad will be published on social networks for patient with no psychiatric conditions.	Diagnostic test: Control group- patients; Control group (n=50) Patients who came to the ER at the hospital due to physical complaints, and no psychiatric history.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Percent agreement of the diagnosis that provide by the psychiatrist and the AI - SPs (actors) part	The accuracy of diagnosis of the AI and of the psychiatrist will be compared to the gold standard as defined by the training that the actors received. This will be quantified according to if the AI or psychiatrist diagnosis was categorized in the appropriate group: Anxiety disorders, Depressive disorders, Schizophrenia spectrum and other psychotic disorders and Trauma and stressor related disorders	3 month
Percent agreement of the diagnosis that provide by the psychiatrist and the AI- Patients part	The psychiatrist and the AI will both interview the patient and will present a diagnosis. The psychiatrist will be considered as the gold standard, and the AI diagnosis will be compared to those given by the psychiatrist. This will be quantified according to if the AI defines the diagnosis in its appropriate group: Anxiety disorders, Depressive disorders, Schizophrenia spectrum and other psychotic disorders and Trauma and stressor related disorders.	5 month

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Percent agreement of the recommendations provided by the AI and the psychiatrists- SPs (actors) part	For biological treatment divided into groups: anti-depressive medication, Electroconvulsive therapy [ECT]/ketamine, anti-psychotics, mood stabilizers and anxiolytics.; Psychological treatments: Cognitive behavioral therapy [CBT], dynamic and supportive psychotherapy. The accuracy of the recommended course of treatment given by the AI and by the psychiatrist will be compared to the gold standard as defined by the training that the SPs received.	3 month
Percent agreement of the recommendations provided by the AI and by the psychiatrists - Patients part	For biological treatment divided into groups: anti-depressive medication, ECT/ketamine, anti-psychotics, mood stabilizers and anxiolytics.; Psychological treatments: CBT, dynamic and supportive psychotherapy. The psychiatrist will be considered as the gold standard, and the AI recommendations of treatment will be compared to those given by the psychiatrist.	5 month

Collaborators and Investigators

• Sponsor: Sheba Medical Center
• Investigators: Principal Investigator: Mark Weiser, M.D, Sheba Medical Center

Study record dates

Study Major Dates

• Study Start (Actual): December 12, 2023
• Primary Completion (Estimated): December 31, 2024
• Study Completion (Estimated): December 31, 2024

Study Registration Dates

• First Submitted: April 10, 2024
• First Submitted That Met QC Criteria: May 14, 2024
• First Posted (Actual): May 16, 2024

Study Record Updates

• Last Update Posted (Actual): May 16, 2024
• Last Update Submitted That Met QC Criteria: May 14, 2024
• Last Verified: November 1, 2023

More Information

Terms related to this study

• Keywords: Artificial intelligence; Medical technology; Mental health diagnosis
• Additional Relevant MeSH Terms: Pathologic Processes; Disease
• Other Study ID Numbers: SHEBA-23-0356-MW-CTIL

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No