News & Events

While computer vision has gained traction in medical applications, models specifically engineered for intensive care unit (ICU) activities are limited. We present Clinical Behavioral Atlas (CBA), a...

This Viewpoint argues that standardizing and streamlining administrative tasks could decrease health care costs and have a direct positive effect on patients and clinicians.

The experience of cancer and the cost of cancer care rise to the level of a catastrophic event, and thus, patients with cancer disproportionately suffer from this fundamental failure of health care cost containment. In fact, the failure of health insurance to provide this basic financial protection in the setting of these rising costs has generated its own characterization, financial toxicity.

Check out John Wilson, MCiM ‘24, and his first published article in Harvard Public Health magazine. Currently pursuing his MPH at Harvard T.H. Chan School of Public Health, John recently wrote an insightful piece on the evolution of the U.S. marijuana industry and the pressing need for federal oversight and regulation.

Learn from Dr. Kevin Schulman, MCiM Program Director, as he explores how European governments negotiate with pharmaceutical companies to lower drug prices. This report sheds light on why Americans pay more for prescription drugs and offers insights into potential solutions.

In this paper, Dr. Kevin Schulman, introduces Precedents Thinking as a solution to the $265 billion annual administrative waste in U.S. healthcare. The authors propose standardized digital contracts and a uniform digital transaction platform as key strategies to reduce transaction costs.

Crowe, Chen, and colleagues address the ethical and practical considerations for integrating generative AI into healthcare. They present recommendations for clinicians, technologists, and healthcare organizations on using generative AI responsibly, with emphasis on clinical decision-making, health system optimization, and the patient-physician relationship.

In this cross-sectional study, Tse, Zahedivash, Anoshiravani, and colleagues investigate the application of large language models (LLMs) in handling pediatric in-basket messages. The research focuses on distinct challenges within pediatric care, particularly when processing clinical messages from adolescent patients and their proxies. The study highlights the nuances and implications of using LLMs in sensitive pediatric settings.

Explore this insightful paper by MCiM faculty, Linda Hoff and Nigam H. Shah, which examines the misalignment between regulatory frameworks, reimbursement structures, and business models for AI in healthcare. The authors highlight how these gaps can lead to rising costs without improving outcomes.

In this paper, Dr. Kevin Schulman and MCiM alumna Crystal Tin-Tin Chang, MCiM '24, explore the shortcomings of value-based care models, which continue to dominate policy discussions and private health insurance markets. The authors argue that policymakers must adopt a more holistic, market-level view to effectively address the deficiencies in the U.S. health system’s value payment program.

Exploring solutions for 'surprise' medical bills, Schulman and Richman discuss potential approaches to managing out-of-network pricing challenges in healthcare to reduce patient financial burden in their latest JAMA article.

Schulman, Narayan, and Bhushan analyze India’s rapidly advancing digital health strategy, highlighting its impact on accessibility, patient engagement, and system efficiency in the latest npj Digital Medicine article.

A recent Milbank Quarterly study, co-authored by Schulman, compares international frameworks for drug price negotiation, suggesting that the U.S. could improve Medicare outcomes by incorporating practices from other countries to negotiate more effectively on high-cost medications.

Nayak, Savage, Gallo, Rangan, and Chen address a key barrier in using large language models (LLMs) for clinical decision-making. They develop diagnostic reasoning prompts to study whether LLMs can imitate clinical reasoning while accurately forming a diagnosis.

Research study that assesses how well GPT-3.5 and GPT-4 address healthcare-related queries from a clinical informatics consultation service. Authored by Dash, Chen, Pedreira, Goh, Arnaout, Morris, Magon, Shah, and colleagues, it explores the safety and concordance of model responses to real-world healthcare questions.