News for the field of mental health: As part of a research operation of the Sheba Medical Center based on Microsoft technology in collaboration with KPMG, a system based on generative artificial intelligence (Generative AI) called “Lib” was developed, which aims to help psychiatrists perform the initial assessment of patients dealing with various mental conditions, including post-trauma, anxiety, depression and emotional distress.
The results of the first phase of the research conducted by the developers indicate that the experimental artificial intelligence system succeeds in identifying with a very high level of accuracy (94%) patients with various mental states, including post-trauma, anxiety, mania and depression. The “Lib” system is built in the configuration of an interactive conversation tool and allows the patient to have a voice or written conversation in Hebrew, and create a personalized character that will talk to him.
The field of mental health has been in great demand in recent years, and following the events of October 7 and the war, this demand is only increasing and is expected to increase even more. And this, even though this is a field that for years has been suffering from dwindling resources and a lack of standards and therapists, including psychologists, psychiatrists and emotional therapists. At the Sheba Medical Center, they recognized at an early stage that the news brought by Microsoft’s Generative AI technology could help the medical teams who are busy with the identification and initial assessment of people who need mental health care and may suffer from post-traumatic stress disorder, and began researching this field.
The purpose of the research, conducted at ARC – Sheba’s innovation center, is to produce the first artificial intelligence platform of its kind in the world, which will be a decision support tool for psychiatrists and will streamline the process of diagnosing and sorting patients from the field of mental health. This is how “Lib” was created, which is built in the configuration of an interactive conversational tool and allows the patient to have a conversation in Hebrew in a natural language.
The patient can choose to have a voice or written conversation, and also, create a personalized character to talk to him. Thanks to the innovative technology, the system knows how to express empathy towards the patient and conduct a relaxed, honest and in-depth conversation with him, at the end of which a summary is received that includes a diagnosis, and recommendations for personalized, honest, broad and in-depth treatment. At the end of the conversation, the system produces a summary for the therapists that includes a diagnosis, differential diagnosis and recommendations for personalized treatment.
As part of the research, which is currently taking place under the leadership of Prof. Mark Weiserdirector of the Department of Psychiatry at Sheba and Dr Assaf Caspidirector of the Department of Psychiatry A at Sheba, the researchers compared the recommendations of “Liv” to those of psychiatrists given the same data and information about patients, and a high agreement of over 94% was found between Liv and the doctors.
On top of that, the study participants reported that the interaction with the artificial intelligence was empathetic and attentive, sometimes even more than that done with human therapists. The statistical results were analyzed by the Gartner Institute for Epidemiology and Health Policy Research. The promising interim results of the study will be presented in March 2024 as part of the HIMSS conference – the largest digital medicine conference in the world.
Prof. Eyal Tsimlichman, Vice President of Innovation and Transformation at the Sheba Medical Center: “The high demand in the field of mental health worldwide, and in Israel in particular in these challenging times, requires solutions that will succeed in bridging the existing gaps in manpower. Artificial intelligence and Generative AI in particular have the ability to help therapists respond to more patients, provide them with more effective care and improve the availability of quality care.
The work with Microsoft Israel and KPMG illustrates part of the potential of artificial intelligence to produce a move that we believe will change the face of the healthcare field. We anticipate that within a few months Liv will be able to be put into use and bring significant value to our patients and other patients around the world.”
“Lib” is based on the artificial intelligence capabilities of Microsoft Azure which relies on the developments of OpenAI in a secure way and maintains the privacy of the user and the organization. Microsoft has implemented Generative AI capabilities in all of its technology platforms. The system was developed thanks to the breakthrough achieved by Open AI about a year ago and was implemented in Microsoft’s cloud.
How It Works?
In the first step, the system asks the user to fill out a questionnaire that includes personal details, including the use of medications and their dosage. After completing the questionnaire, “Lib” will conduct the conversation with the user according to the gender, age and other personal details that the patient provided in the questionnaire. The conversation will be conducted in Hebrew and with a demonstration of empathy, similar to a conversation with a human being. Later, after receiving the patient’s consent, the system will be able to connect to his medical file at the hospital, in order to obtain additional medical background that will enable a more accurate diagnosis.
In cases where the system assesses that the medical condition is suitable for a softer intervention, it will offer ways of emotional treatment such as mindfulness, relaxation exercises, training videos and more. On the other hand, in cases identified as more serious, the system will alert the medical staff in real time that an urgent and immediate response must be given to the patient, and will also alert the patient that it is important that he receive immediate treatment.
At the end of the conversation, the system produces a summary containing the background on the patient and treatment recommendations, which is forwarded to the human therapist for review. The therapist will verify the recommendations given by the system in a personal conversation with the patient and determine the best treatment for him. The system is expected to save therapists many hours and according to estimates will double the amount of patients a therapist can receive, compared to the current situation.
“The high demand in the field of mental health throughout the world, and in Israel in particular in these challenging times, requires solutions that will succeed in bridging the existing gaps in manpower,” said Prof. Eyal Tsimlichman, VP of Innovation and Transformation at the Sheba Medical Center, “Artificial intelligence has the ability to help therapists respond To more patients, to provide them with more effective treatment and to improve the availability of quality treatment. We anticipate that within a few months Liv will be able to be used and bring significant value to our patients and other patients around the world.”
The system is expected to save therapists many hours and according to estimates will double the amount of patients a therapist can receive, compared to the current situation. One of the challenges of the mental health system also lies in the need to distinguish between patients whose condition requires immediate intervention and patients who are in a simpler situation in order to know how to prioritize the resources available to the public health system in a situation of peak demand and limited supply.
Therefore, in cases where the system assesses that the medical condition is suitable for a softer intervention, it will offer ways of emotional treatment such as mindfulness, relaxation exercises, training videos and more. On the other hand, in cases identified as more serious, the system will alert the medical staff in real time that an urgent and immediate response must be given to the patient, and will also alert the patient that it is important that he receive immediate treatment.