Data in Healthcare with John Halamaka & Nils Reich
In a panel discussion held on June 18, 2024, at HLTH Europe in Amsterdam, experts discussed the application of data analytics and artificial intelligence (AI) in healthcare has the potential to revolutionise how care is delivered, improving efficiency, reducing costs, and leading to better patient outcomes. At the recent HealthTech conference, a panel of industry leaders from provider and payer organisations shared their perspectives on the current state and future potential of data and AI in healthcare.
The panel included John Halamka, CEO of Mayo Clinic Platform, and Nils Reich, CEO of Allianz Health. They were interviewed by Sara Siegel, Head of Health at Deloitte.
Why Hospitals Need AI the Most
Eyal Zimlichman, Chief Transformation and Innovation Officer at Sheba Medical Center, argued that hospitals have the most to gain from AI and will likely see the biggest changes. “Hospitals are probably the best place where we would see some of the major changes happen,” Zimlichman said. “Hospitals need it most — hospitals are the most inefficient, the most ineffective and unsafe institutions. 40% of the cost of countries’ healthcare spending on average goes to hospitals.”
Zimlichman believes AI will provide the most value in improving decision-making across clinical, administrative, operational, patient flow and financial areas. “If we want to focus on where we’re actually going to see most of the change, it’s on better decision making,” he said. “We need better decisions, and AI will help us get there.”
Early AI Applications Showing Promise
Panellists shared examples of AI applications already making an impact in their organisations. At Charité University Hospital in Berlin, algorithms are being used to help manage patients, such as a system that can predict acute kidney injury and prevent patients from needing dialysis. “It helps to really have up to 50 people per year not going to devices,” said Peter Gocke, Chief Digital Officer at Charité.
Sheba Medical Center has developed an AI-based avatar for mental health triage and screening in partnership with Microsoft and KPMG. After a 20–30 minute conversation, the avatar can provide a diagnosis, assess severity, and recommend treatment. “We found that it’s 94% accurate in terms of setting the right diagnosis and severity,” Zimlichman reported. “We’re going through second phase clinical trials and aim to get this regulated before the end of the year to hopefully become the first ever regulated patient-facing generative AI technology.”
Overcoming Barriers to AI Adoption
While the potential of AI is clear, panellists acknowledged significant challenges in implementing it widely. Obtaining quality data remains a hurdle. “80% of the knowledge in the healthcare data we collect is in text, but that text is hard to translate into something concrete and interpretable,” explained Halamka. “That’s so much of the work all of us are here to do.”
Proving return on investment is also difficult, especially for applications aimed at prevention. “Population health is not easy to calculate in a system that’s funded by only paying for sickness events,” said Gocke. “IT systems and AI are not funded in the regular reimbursement schemes.”
To overcome these barriers, Nadine Hachach-Haran, Founder and CEO of Proximie, advises companies to take a consultative approach to health systems. “Go in, in the early days, with more of a consultative approach and say, ‘We understand you have these challenges. We believe we have some tooling that can help you. Let’s work together to try and build the use case in the real world.’” At the same time, she recommends health systems build dedicated change management teams to work with startups on redesigning workflows.
The Path Forward
Looking ahead to 5–10 years, the panellists expressed optimism about the potential of AI, especially large language models and generative AI, to transform healthcare delivery. Hachach-Haran is excited about applications in the operating room. “This idea that you can have a surgeon co-pilot in the room with you bringing you information, interacting with you… that intersection of human-machine engagement in the OR is really on the horizon and quite exciting.”
Zimlichman believes AI will enable healthcare to be delivered anywhere, regardless of the patient’s location. “We believe in care delivery with no regard to where the patient is located — it will be on Mars in the future! We cannot do this without support by algorithms and AI.”
Halamka predicts rapid progress, estimating that data and AI utilisation will be significantly improved by 2025. Reich agrees that AI is key to making the overall healthcare system more sustainable. “Hopefully, in 10 years, we will be in a space where we can really measure outcomes, and that should enable us to change the system quite fundamentally to something that is more prevention-oriented, where we can align incentives as a payer to pay more for outcomes and results that matter to patients.”
While challenges remain, it’s clear that leading healthcare organisations are actively working to harness data analytics and AI to improve care delivery. With the right partnerships, change management, and evidence of improved outcomes and ROI, the once futuristic vision of an AI-enabled health system is steadily becoming reality. As Halamka boldly predicts, “By 2025, it will be better.”
