Lightnews — Scholar-powered news

NEJM AI @ai.nejm.org · 4h

In the latest episode of the NEJM AI Grand Rounds podcast, Dr. Jonathan Chen discusses his path from teenage programmer to Stanford physician-informatician and why machine learning has both thrilled and unnerved him. Listen to the full episode: nejm.ai/ep35

#AI #MedSky

AI GRAND ROUNDS

EPISODE 35
Medicine, Machines, and Magic: Dr. Jonathan Chen on Medical AI

1

NEJM AI @ai.nejm.org · 1d

Editorial by Robert D. Truog, MD, MA, and R. Sean Morrison, MD: Machine Learning Cannot Replace Surrogate Decision-Makers in Resuscitation Decisions for Incapacitated Patients nejm.ai/3Vysymr

#AI #MedSky #MLSky

Page 1 of the editorial "Machine Learning Cannot Replace Surrogate Decision-Makers in Resuscitation Decisions for Incapacitated Patients"

Bottom: Read full article at ai.nejm.org.

1 1

NEJM AI @ai.nejm.org · 1d

Original Article by G. Starke et al.: Machine Learning–Based Patient Preference Prediction: A Proof of Concept nejm.ai/3IAIYrD

#AI #MedSky #MLSky

Table 2. Characteristics of the 1811 Included Study Participants, Meaning All Participants for Whom the Target Variable Was Available.

Table 3. Accuracy of Predicting Preferences Toward Attempting Cardiopulmonary Resuscitation in the Case of Cardiorespiratory Arrest among Couples in the Sample.

NEJM AI @ai.nejm.org · 2d

What does a chief health AI officer do? Dr. Karandeep Singh shares on NEJM AI Grand Rounds: Keep #AI smart, safe, and ethical in health systems. Learn more: nejm.ai/ep34

#MedSky #MLSky

2 6

NEJM AI @ai.nejm.org · 5d

Editorial by Issa J. Dahabreh, MD, ScD, Robert W. Yeh, MD, MSc, MBA, and Piersilvio De Bartolomeis, MSc: Trial Emulation, Simulation, and Augmentation Using Electronic Health Records and Generative AI nejm.ai/4mqiyXj

#AI #MedSky #MLSky

Page 1 of the editorial "Trial Emulation, Simulation, and Augmentation Using Electronic Health Records and Generative AI"

Sitting at the bottom is "Read full article at ai.nejm.org."

NEJM AI @ai.nejm.org · 5d

Editorial by David Ouyang, MD, and Jeffrey M. Drazen, MD: Scoping Out Clinical Trial Emulation nejm.ai/4nnlrJY

#AI #MedSky #MLSky

Page 1 of the editorial "Scoping Out Clinical Trial Emulation"

Sitting at the bottom is "Read full article at ai.nejm.org."

NEJM AI @ai.nejm.org · 5d

Original Article by J. González et al.: TRIALSCOPE — A Framework for Clinical Trial Simulation from Real-World Data nejm.ai/4nnl1TU

#AI #MedSky #MLSky

Reposted by NEJM AI

CAUSALab @causalab.bsky.social · 9d

A new editorial in @ai.nejm.org explores how AI computational tools can transform causal analyses.

Researchers examine how AI can support #targettrial emulation, enhance observational data extraction and analysis, and improve efficiency of clinical trials.

🔗 ai.nejm.org/doi/full/10....

New publication: Trial Emulation, Simulation, and Augmentation Using Electronic Health Records and Generative AI

1 3 3

NEJM AI @ai.nejm.org · 6d

Harshe and colleagues appreciate the interest and engagement shown by Kenjiro Shiraishi regarding their Perspective and for the thoughtful comments and the opportunity to further expand on its aims and findings. nejm.ai/4mZ9Tf9

#AI #MedSky #MLSky

We hope this project can inform gaps in the reasoning of current LLMs and help us explore how future models might be better trained to reason appropriately in clinical ethics scenarios where we need practical solutions for real patient care.

“Response to ‘Before Asking AI to Make Ethical Judgments, Shouldn’t We First Ask What Ethics We Are Asking It to Follow?’” by Isha Harshe, B.S., B.A., Kenneth Goodman, Ph.D., and Gauri Agarwal, M.D.

NEJM AI @ai.nejm.org · 6d

Perspective by Isha Harshe, BS, BA, Kenneth W. Goodman, PhD, and Gauri Agarwal, MD: Can a Chatbot Be a Medical Surrogate? The Use of Large Language Models in Medical Ethics Decision-Making nejm.ai/3TajeDX

NEJM AI @ai.nejm.org · 6d

In a letter commenting on a Perspective about large language models assisting with ethical decision-making, Kenjiro Shiraishi writes that a critical dimension is missing: the structure of the question itself. Read the full letter: nejm.ai/4248rQC

#AI #MedSky #MLSky

“Ethical judgment is not just a matter of textual consistency or expert approval; it requires a coherent alignment between the question’s meaning and the underlying value system.”

Perspective
“Before Asking AI to Make Ethical Judgments, Shouldn’t We First Ask What Ethics We Are Asking It to Follow?” by Kenjiro Shiraishi

1

NEJM AI @ai.nejm.org · 7d

Original Article by M.V. Heinz et al.: Randomized Trial of a Generative AI Chatbot for Mental Health Treatment nejm.ai/43v2mOY

NEJM AI @ai.nejm.org · 7d

Perspective by Timothy G. Heckman, PhD, John C. Markowitz, MD, and Bernadette Davantes Heckman, PhD: A Generative AI Chatbot for Mental Health Treatment: A Step in the Right Direction? nejm.ai/41N193V

1

NEJM AI @ai.nejm.org · 7d

Perspective by Ilana Gratch, MS, MPhil, and Todd Essig, PhD: A Letter about “Randomized Trial of a Generative AI Chatbot for Mental Health Treatment” nejm.ai/45JDlyI

1

NEJM AI @ai.nejm.org · 7d

Heinz and colleagues appreciate the thoughtful responses provided by Heckman et al. and Gratch and Essig regarding their recent publication evaluating the generative AI chatbot, Therabot. Read the full letter: nejm.ai/3UQAyin

#AI #MedSky #MLSky

“We appreciate the opportunity to engage in this correspondence and to provide further context for our study design. We emphasize that ours is an initial trial in this space — not the last.”

Perspective
“Response to Letters about ‘Randomized Trial of a Generative AI Chatbot for Mental Health Treatment’” by M.V. Heinz et al.

1

NEJM AI @ai.nejm.org · 7d

On NEJM AI Grand Rounds, Dr. Karandeep Singh shares what his team learned from studying a sepsis prediction tool. Hear more from Dr. Singh: nejm.ai/ep34

#AI #MedSky #MLSky

1

NEJM AI @ai.nejm.org · 8d

Although ML has been shown to outperform surrogate decision-makers in predicting patient preferences for CPR, the authors of a new editorial argue that ML cannot replace surrogates in making decisions for incapacitated patients. nejm.ai/3Vysymr

#AI #MedSky #MLSky

Our greatest concern as clinicians is that the data presented in this article will be interpreted by the media and other readers as evidence that ML can replace surrogates in end-of-life decision-making.

“Machine Learning Cannot Replace Surrogate Decision-Makers in Resuscitation Decisions for Incapacitated Patients” by Robert D. Truog, M.D., M.A., and R. Sean Morrison, M.D.

1 2

NEJM AI @ai.nejm.org · 8d

A new study presents the first proof of concept for a machine learning–based patient preference predictor, which, using data from over 1800 Swiss adults, outperformed human surrogates in predicting cardiopulmonary resuscitation preferences. Learn more: nejm.ai/3IAIYrD

#AI #MedSky #MLSky

Figure 1. Confusion Matrices and Beeswarm Summary Plots.

1 5

NEJM AI @ai.nejm.org · 9d

What we say versus what we do. On NEJM AI Grand Rounds, Dr. Karandeep Singh explains why health #AI needs humility about the gap between theory and practice. Learn more: nejm.ai/ep34

#MedSky #MLSky

1

NEJM AI @ai.nejm.org · 12d

The authors of a new editorial write that although computational tools cannot address biases inherent in the process of generating observational data, observational emulations may still be valuable for informing trial design and improving trial efficiency. nejm.ai/4mqiyXj

#AI #MedSky #MLSky

“In the presence of unmeasured confounding or other biases that emulation alone cannot address, using emulation estimates as the basis for simulations is imperfect but still preferable to less data-driven approaches for trial design.”

Editorial
“Trial Emulation, Simulation, and Augmentation Using Electronic Health Records and Generative AI” by Issa J. Dahabreh, M.D., Sc.D., Robert W. Yeh, M.D., M.Sc., M.B.A., and Piersilvio De Bartolomeis, M.Sc.

1

NEJM AI @ai.nejm.org · 12d

Original Article by J. González et al.: TRIALSCOPE — A Framework for Clinical Trial Simulation from Real-World Data nejm.ai/4nnl1TU

NEJM AI @ai.nejm.org · 12d

A new editorial explores whether TRIALSCOPE, or similar applications, could generate sufficient high-quality clinical evidence such that new treatments could be released for widespread use after small, controlled trials. Full editorial: nejm.ai/4nnlrJY

“Oncology is in the unique position of having many therapeutics repurposed and reevaluated for many cancers other than the initial targets. Such a framework is ripe for clinical trial emulation....”

Editorial
“Scoping Out Clinical Trial Emulation” by David Ouyang, M.D., and Jeffrey M. Drazen, M.D.

1

NEJM AI @ai.nejm.org · 12d

When TRIALSCOPE was applied to over 1 million cancer patient records, the framework structured free text automatically, reduced confounding through causal inference techniques, and enabled the simulation of clinical trials across diverse patient cohorts. nejm.ai/4nnl1TU

#AI #MedSky #MLSky

Original Article | Sep 22, 2025
TRIALSCOPE — A Framework for Clinical Trial Simulation from Real-World Data
J. González and Others

A visual representation of TRIALSCOPE

1

Reposted by NEJM AI

NEJM.org @nejm.org · 12d

A new article in NEJM Catalyst explores AI’s growing impact on carbon emissions, electricity consumption, and costs for health systems; presents a framework for mitigation of environmental harms; and calls for a collaborative effort among health care and industry leaders. nej.md/3VN8X1Y

Article
“In this single use case, although the absolute emissions presented are not exceedingly high, health AI’s footprint can rapidly accumulate when considering hundreds of different tools for communication, clinical decision support, or predictive modeling.”

“Sustainably Advancing Health AI: A Decision Framework to Mitigate the Energy, Emissions, and Cost of AI Implementation” by Anu Ramachandran et al.

Read this article and more in the October 2025 issue of NEJM Catalyst Innovations in Health Care Delivery at catalyst.nejm.org.

2 6

NEJM AI @ai.nejm.org · 13d

Original Article by M.V. Heinz et al.: Randomized Trial of a Generative AI Chatbot for Mental Health Treatment nejm.ai/43v2mOY

Perspective by M.V. Heinz et al.: Response to Letters about “Randomized Trial of a Generative AI Chatbot for Mental Health Treatment” nejm.ai/3UQAyin