AI Applications in Healthcare: Sustaining Trust and Responsibility - Insights from Our Symposium

Modern AI in Healthcare at the Cedars-Sinai Conference

AI Applications in Healthcare: Sustaining Trust and Responsibility - Insights from Our Symposium

Sharing is Caring! Share this post on Facebook Share this post on Twitter Email this Post

The recent AI in healthcare conference, jointly organized by Cedars-Sinai and our website, took place on May 16 in Los Angeles. The event invited leading experts to discuss the emerging ethical, societal, and legal issues surrounding AI in medicine and biomedical research. Keynote speakers at the conference included Vardit Ravitsky, President of our website, and Nancy Berlinger, senior research scholar at The Hastings Center. Let's dive into some highlights from their discussions!

What was the core focus of your talk or panel, and what stood out as the most significant aspect of it?

Vardit: I had a captivating discussion with David Rhew, Chief Medical Officer at Microsoft, under the tagline "Stepping Towards Responsible AI in the Medical Field." We touched upon the thrilling prospects and limitations of AI today and, notably, the groundbreaking use of AI in analyzing patients' eye data to early detect chronic diseases and enhance affordable healthcare access. The most striking part of this discussion was the potential for AI to address health disparities by simplifying and improving the accessibility of screening procedures for rural and underserved communities.

Nancy: I moderated a panel discussing the use of AI in the inpatient care of gravely ill patients. The panelists, consisting of an ICU physician, an AI ethicist, a spiritual care director, and a clinical ethicist, explored AI models designed to advise physicians on appropriate treatment selections when confronted with uncertain benefits, burdens, or risks. The panelists agreed that AI could serve as an assistant but must not dictate the final decision, leaving room for clinical judgment and hands-on expertise.

The session also highlighted that Cedars-Sinai is testing a mental health chatbot for use by spiritual care services. The spiritual care director explained that chatbots and other AI technology can provide supplementary support to patients but are unable to replicate human presence. Besides, they invited us to ponder new possibilities for AI, such as leveraging virtual reality to offer transcendent experiences for seriously ill individuals.

What were the main takeaways from the afternoon workshops?

Vardit: I led a workshop focusing on AI-powered research that emphasized the ability of voice analysis to serve as a vital biomarker in diagnosing diseases, detect changes in voice, track disease progression, and provide valued public health support. Sharing recorded voice samples with clinicians via the cloud could make this technology beneficial even for patients who cannot meet with a voice specialist in person, thereby potentially reducing health disparities. We addressed the ethical and social implications of voice as a biomarker, including concerns related to deepfakes that could exploit this information for identity theft.

Nancy: I co-led a workshop with Charles Binkley, focusing on the expectations for healthcare institutions concerning transparency with patients and caregivers regarding AI models. For instance, AI-driven "ambient scribes," which record and summarize interactions between doctors and patients or family caregivers, may soon become standard. The question then arises: how should this technology be disclosed and explained to patients? Additionally, talks on hospital policy cycles highlighted that they struggle to keep pace with AI evolution, leading some institutions to create specialized AI committees to speedily draft and update AI-related policies.

Upon reflecting on the conference, what one aspect left the most lasting impression?

Vardit: I was astounded by the swift implementation of certain AI tools into healthcare at present. However, I was equally inspired by the thoughtfulness demonstrated by clinicians and healthcare systems along their journey towards ethical and accountable AI adoption. Everybody understood the importance of maintaining public trust in AI, even if it calls for slowing down to ensure diligent practice. I left the conference feeling hopeful and optimistic.

Nancy: Michael Nurok, who holds training in anthropology in addition to medicine, reminded us that every nascent technology initially elicits uncertainty and apprehension along with enthusiasm. As with AI, we should anticipate similar reactions and adapt accordingly. He also encouraged us to remember that caring for gravely ill patients and their families involves a higher degree of uncertainty. AI, for all its potential, will not resolve these everyday challenges. In the next section, we delve deeper into AI-driven healthcare innovations.

[PHOTO: Michael Nurok, Nancy Berlinger, Charles Binkley]

Share this post on Facebook Share this post on Twitter Email this Post

In the shadows of the conference, a revolution in healthcare is underway, led by AI applications utilizing uncommon biomarkers such as data from patients' eyes and voices.

Eye-Based Biomarkers

• Retinal Analysis and Early Disease Detection: AI technologies are increasingly employed to gauge retinal images for discerning disease signs like diabetes, high blood pressure, and neurological disorders, predicting risks before physical symptoms arise.
• Oculomotor and Pupillometry Examination: Researchers are investigating the potential of AI to scrutinize eye movement, pupil dilation, and blink patterns as biomarkers for cognitive decline, mental health issues, and neurological disorders. Although currently experimental, this approach holds promise for unobtrusive monitoring and early intervention.
• Combining with Other Data Streams: The sophistication of AI models is rapidly evolving, enabling smarter amalgamation of retinal imaging data with genetic, demographic, and clinical information for more comprehensive risk assessments and personalized treatment plans.

Voice-Based Biomarkers

• Speech Analysis for Neurological and Psychiatric Disorders: AI-backed speech analysis tools are being developed to recognize speech patterns indicative of conditions like Parkinson’s disease, Alzheimer’s disease, depression, and autism. Analysis of key features like pitch, rhythm, tone, and fluency can detect deviations from regular speech patterns.
• Remote and Early Detection: Voice-based AI solutions facilitate remote patient monitoring, flourishing in the management of chronic diseases, and assisting patients in rural or underserved areas.
• Integration with Healthcare Platforms: AI-driven voice analysis is seamlessly integrated into telemedicine platforms, digital health records, and patient engagement tools, providing continuous health monitoring and support.

The broader healthcare AI landscape is experiencing momentous change as well:

• Medical Imaging: AI-fortified MRI and CT scanners furnish more precise and early diagnoses for diseases like cancer, with spectacular advancements being accomplished in radiology—almost 400 FDA-approved AI algorithms are available to date.
• Predictive Analytics and Clinical Decision Support: AI helps analyze patient data to foresee disease progression, readmission risks, and treatment responses, enhancing treatment outcomes and efficiency.
• Robotics and Automation: Robotics within surgery and AI-enhanced automation for administrative tasks precipitate workforce shortages and augment procedural accuracy.

Key Trends and Advancements

• AI in Remote and Rural Healthcare: Organizations are deploying portable X-ray units and AI kits for tuberculosis screening and other diagnostics in underserved areas, showcasing AI-driven healthcare's scalability and accessibility.
• Precision Medicine and Genomics: AI is transforming genomic analysis, enabling more accurate identification of genetic variants and custom-tailored treatments, specifically in oncology.
• Human-Centered AI Integration: The emphasis is gradually shifting towards AI systems that collaborate with clinicians, fostering transparency, accountability, and trust.

1. The use of AI in analyzing patients' eye data, such as retinal images, is a promising development in healthcare, potentially enabling early detection of chronic diseases like diabetes, high blood pressure, and neurological disorders.
2. AI-backed speech analysis tools are being developed to recognize speech patterns indicative of conditions like Parkinson’s disease, Alzheimer’s disease, depression, and autism, providing remote and early detection capabilities.
3. The intersection of AI, technology, and artifical intelligence is transforming various aspects of healthcare, including medical imaging, predictive analytics, clinical decision support, robotics, and automation, aiming to improve diagnosis accuracy, treatment efficiency, and patient outcomes.

Read also: