Jonathan H Chen MD PhD
@jonc101.bsky.social
Physician Data Scientist - Stanford Center for Biomedical Informatics Research + Division of Hospital Medicine + Clinical Excellence Research Center + Biomedical Data Science
and hospital managers, not just model developers, choose decision thresholds for generating alerts that balances predictive accuracy, alert burden and lead time to clinician actions."
November 11, 2025 at 6:06 PM
and hospital managers, not just model developers, choose decision thresholds for generating alerts that balances predictive accuracy, alert burden and lead time to clinician actions."
some unique results: 1) Sub-models separately trained on ED and ward patients perform better in these groups than a base model trained and applied to all patients combined; 2) Sepsis-3 criteria performs poorly as the reference label; 3) A methodology has been developed that can assist clinicians
November 11, 2025 at 6:06 PM
some unique results: 1) Sub-models separately trained on ED and ward patients perform better in these groups than a base model trained and applied to all patients combined; 2) Sepsis-3 criteria performs poorly as the reference label; 3) A methodology has been developed that can assist clinicians
hospitals in Queensland, Australia. Steps are now underway in designing and evaluating systems for presenting QSA predictions to treating clinicians in a simulation environment prior to a live data clinical pilot in an emergency department (ED) and a general medical ward. Our research has generated
November 11, 2025 at 6:06 PM
hospitals in Queensland, Australia. Steps are now underway in designing and evaluating systems for presenting QSA predictions to treating clinicians in a simulation environment prior to a live data clinical pilot in an emergency department (ED) and a general medical ward. Our research has generated
Abstract: "Sepsis remains a potentially lethal but highly treatable condition that affects about up to 1 in 10 hospital admissions. We have developed and validated, in a prospective silent trial, an AI-enabled sepsis risk prediction algorithm trained on more than 700,000 admissions to 7 digital
November 11, 2025 at 6:06 PM
Abstract: "Sepsis remains a potentially lethal but highly treatable condition that affects about up to 1 in 10 hospital admissions. We have developed and validated, in a prospective silent trial, an AI-enabled sepsis risk prediction algorithm trained on more than 700,000 admissions to 7 digital
interests in clinical informatics, evidence-based medicine, clinical reasoning and quality and safety improvement.
November 11, 2025 at 6:06 PM
interests in clinical informatics, evidence-based medicine, clinical reasoning and quality and safety improvement.
has collaborations with colleagues within the UQ Digital Health Centre, the Centre for Health Informatics at Macquarie University, the CRC in Digital Health at Queensland University of Technology and the Clinical and Business Intelligence Unit of eHealth Queensland. He has longstanding research
November 11, 2025 at 6:06 PM
has collaborations with colleagues within the UQ Digital Health Centre, the Centre for Health Informatics at Macquarie University, the CRC in Digital Health at Queensland University of Technology and the Clinical and Business Intelligence Unit of eHealth Queensland. He has longstanding research
chairs the Metro South Clinical AI Working Group and Queensland Health Sepsis AI Working Group, and is Professor in Clinical Decision-making at the University of Queensland (UQ). He has co-authored multiple papers on the use of AI in healthcare, is principal investigator for several AI trials and
November 11, 2025 at 6:06 PM
chairs the Metro South Clinical AI Working Group and Queensland Health Sepsis AI Working Group, and is Professor in Clinical Decision-making at the University of Queensland (UQ). He has co-authored multiple papers on the use of AI in healthcare, is principal investigator for several AI trials and
Bio Ian: Dr Ian Scott is consultant general physician and former Director of Internal Medicine and Clinical Epidemiology at Princess Alexandra Hospital in Brisbane. He is currently Clinical Consultant in AI at the Digital Health and Informatics Division of Metro South Hospital and Health Service,
November 11, 2025 at 6:06 PM
Bio Ian: Dr Ian Scott is consultant general physician and former Director of Internal Medicine and Clinical Epidemiology at Princess Alexandra Hospital in Brisbane. He is currently Clinical Consultant in AI at the Digital Health and Informatics Division of Metro South Hospital and Health Service,
at Queensland health to develop and prospectively trial AI methods to predict sepsis and clinical deterioration.
November 11, 2025 at 6:06 PM
at Queensland health to develop and prospectively trial AI methods to predict sepsis and clinical deterioration.
implementing sophisticated software programs. Recently Anton architected and managed two projects within Queensland Health to support AI experimentation with health data, including the development of CLARA, a Clinical AI research Accelerator data lab. Currently Anton is collaborating with clinicians
November 11, 2025 at 6:06 PM
implementing sophisticated software programs. Recently Anton architected and managed two projects within Queensland Health to support AI experimentation with health data, including the development of CLARA, a Clinical AI research Accelerator data lab. Currently Anton is collaborating with clinicians
Bio Anton: Dr Anton Van Der Vegt is an Advanced QLD Industry Research Fellow with the Centre for Health Services Research at UQ Faculty of Medicine. Originally trained as a Mechanical Engineer and Computer Scientist, Anton has worked across Australia, Europe, the US designing, developing and
November 11, 2025 at 6:06 PM
Bio Anton: Dr Anton Van Der Vegt is an Advanced QLD Industry Research Fellow with the Centre for Health Services Research at UQ Faculty of Medicine. Originally trained as a Mechanical Engineer and Computer Scientist, Anton has worked across Australia, Europe, the US designing, developing and
Webinar ID: 978 8759 6012
Webinar Passcode: 420642
Webinar Passcode: 420642
November 11, 2025 at 6:06 PM
Webinar ID: 978 8759 6012
Webinar Passcode: 420642
Webinar Passcode: 420642
I will conclude my talk by showing the limited ability of current artificial intelligence technology to fill that role."
November 4, 2025 at 1:28 PM
I will conclude my talk by showing the limited ability of current artificial intelligence technology to fill that role."
in our understanding of cervical cancer affects his conclusions. The analysis I will describe demonstrates how important medical questions can be answered. Despite the current explosion in clinical data availability, the role for mathematical analysis in finding those answers still remains.
November 4, 2025 at 1:28 PM
in our understanding of cervical cancer affects his conclusions. The analysis I will describe demonstrates how important medical questions can be answered. Despite the current explosion in clinical data availability, the role for mathematical analysis in finding those answers still remains.
cervical cancer. Eddy performed his analysis before the role of the human papilloma virus in cervical cancer was fully understood. In addition, studies conducted after his work have documented the natural history of cervical cancer. My talk will update Eddy’s analysis by showing how these changes
November 4, 2025 at 1:28 PM
cervical cancer. Eddy performed his analysis before the role of the human papilloma virus in cervical cancer was fully understood. In addition, studies conducted after his work have documented the natural history of cervical cancer. My talk will update Eddy’s analysis by showing how these changes
annual testing, had little objective basis. Eddy showed that increasing the screening interval to 3 years would have minimal effect on a woman’s life expectancy. Despite the initial rejection by the medical community, Eddy’s findings are now reflected in the screening protocol recommended for
November 4, 2025 at 1:28 PM
annual testing, had little objective basis. Eddy showed that increasing the screening interval to 3 years would have minimal effect on a woman’s life expectancy. Despite the initial rejection by the medical community, Eddy’s findings are now reflected in the screening protocol recommended for
My talk will update Eddy’s analysis of cervical cancer screening. This disease once was a major cause of death in women. Introduction of the Pap smear greatly reduced the likelihood a woman would die from cervical cancer. However, the initial cervical cancer screening protocols, based on
November 4, 2025 at 1:28 PM
My talk will update Eddy’s analysis of cervical cancer screening. This disease once was a major cause of death in women. Introduction of the Pap smear greatly reduced the likelihood a woman would die from cervical cancer. However, the initial cervical cancer screening protocols, based on
to observe the effects of cancer screening protocols on patient outcomes would require follow-up studies lasting longer than the useful lifetime of the screening technology. This meant Eddy’s work addressed an important question that could not be answered by other means.
November 4, 2025 at 1:28 PM
to observe the effects of cancer screening protocols on patient outcomes would require follow-up studies lasting longer than the useful lifetime of the screening technology. This meant Eddy’s work addressed an important question that could not be answered by other means.
Abstract: "In the late 1970s David Eddy demonstrated how mathematics could predict the effects changes to cancer screening protocols would have on patient outcomes. Ideally the questions Eddy addressed would have been answered using randomized clinical trials. However, the time scale needed
November 4, 2025 at 1:28 PM
Abstract: "In the late 1970s David Eddy demonstrated how mathematics could predict the effects changes to cancer screening protocols would have on patient outcomes. Ideally the questions Eddy addressed would have been answered using randomized clinical trials. However, the time scale needed
occasional consulting and writing.
November 4, 2025 at 1:28 PM
occasional consulting and writing.
of a medical device and development of information systems for the disease management industry. Eventually his daughter established funding for her education. This let him retire so that he could do something for a living other than work. That “something” includes the teaching he now do plus some
November 4, 2025 at 1:28 PM
of a medical device and development of information systems for the disease management industry. Eventually his daughter established funding for her education. This let him retire so that he could do something for a living other than work. That “something” includes the teaching he now do plus some
He spent five years helping that company develop information management products for healthcare. Oracle was not a fun place to work but he did learn where products come from. That knowledge made him useful at the startups he joined after leaving Oracle. His work at startups included the design
November 4, 2025 at 1:28 PM
He spent five years helping that company develop information management products for healthcare. Oracle was not a fun place to work but he did learn where products come from. That knowledge made him useful at the startups he joined after leaving Oracle. His work at startups included the design
clinical decisions to reflect the cost of care. After completing his dissertation he left academia and spent his career in industry. This included 16 years in the medical department of Hewlett Packard’s central research lab. He left when HP sold its medical business to Phillips and moved to Oracle.
November 4, 2025 at 1:28 PM
clinical decisions to reflect the cost of care. After completing his dissertation he left academia and spent his career in industry. This included 16 years in the medical department of Hewlett Packard’s central research lab. He left when HP sold its medical business to Phillips and moved to Oracle.
about the time the field was born in the 1970s. As a graduate student at Stanford he joined a group developing what we now call predictive models. Our initial focus was the clinical presentation of chest pain caused my coronary artery disease. His research concerned the optimal adjustment of
November 4, 2025 at 1:28 PM
about the time the field was born in the 1970s. As a graduate student at Stanford he joined a group developing what we now call predictive models. Our initial focus was the clinical presentation of chest pain caused my coronary artery disease. His research concerned the optimal adjustment of