Subscribe to ExpertReview

The latest research, meeting updates and CME, in your inbox.

AI in healthcare · Research

AI in Clinical Decision-Making: Hidden Values and the Need for Transparency

Human idea, AI creation14 Nov 2026Oncology

Drs Stijn van den Borne

Share

AI tools increasingly support clinical decision‑making but embed hidden value frameworks that shape recommendations in ways physicians and patients cannot anticipate. A new framework, Values In the Model (VIM), proposes transparent labeling of these value trade‑offs.

Illustration of the Values in the Model (VIM) framework highlighting AI hidden value trade‑offs in clinical decisions
Image: human idea, AI creation

Background

Clinical decision-making is increasingly supported by Artificial intelligence (AI) tools, yet the exact value frameworks AI utilises and which influence clinical recommendations are poorly understood by doctors and patients.

Problem

Current AI governance, including model cards alongside regulatory guidelines, describes technical specifications coupled with overarching principles but fails to reveal the clinical values that drive actual recommendations.

Proposed Solution

The RAISE symposium on Responsible AI for Social and Ethical Healthcare introduced the “Values In the Model” (VIM) framework, a clear labeling system that records how AI systems operate value‑laden clinical trade‑offs.

Implementation

instead of depending instead of using electronic health record data could reinforce institutional biases, VIM recommends creating stringent benchmarks of clinical scenarios evaluated from varied stakeholder viewpoints.

Call to Action

The authors advocating parallel pathways: public discussion of values inherent in medical AI and closely supervised pilot projects in top health‑care systems, as no existing organisation currently holds the blend of technical expertise, clinical knowledge, ethical standing, and stakeholder representation necessary to oversee such an effort comprehensively.

References

  1. Goldberg C, Balicer RD, Bhat M, Blumenthal D, Brendel RW, Brondolo E, et al. The Missing Dimension in Clinical AI: Making Hidden Values Visible. NEJM AI. 2026;3(2):AIp2501266. doi:10.1056/AIp2501266. PMID: 42932574. PubMed

This article is not medical advice. Always check the source and local prescribing information before making treatment decisions. Patients should seek a personal assessment by a licensed physician.

©2026 ExpertReview.io by [‘mediPr]

Medically reviewed by Drs Stijn van den Borne on 14 Nov 2026

Related posts

Stylized blue illustration of the small and large intestine with glowing digital pixels and light streaks on a dark backgroundResearch · OncologyMultimodal ML models using CT body composition to predict colorectal surgery complicationsA meta-analysis of eight studies (n = 7,835) found that machine learning models combining CT body composition with clinical variables showed good discrimination for short-term complications after colorectal cancer surgery, with substantial heterogeneity.10 Nov 2026Abstract dark blue cell with glowing cyan core and yellow highlights dissolving into pixels on a dark backgroundResearch · OncologyAI skin cancer diagnostics: scoping review finds little evidence of clinical or economic valueA scoping review of 33 studies found that evaluations of AI for skin cancer mostly report process and user-experience outcomes. Few studies examined clinical, prognostic, system-level or economic outcomes, and most were done in controlled settings.10 Nov 2026Stylized axial brain MRI in blue and gold, dissolving into pixels and light streaks on a dark backgroundResearch · OncologySystematic review of MRI-based machine learning for brain tumour analysisA review of 213 studies found high reported segmentation and classification performance for MRI-based machine learning in brain tumours, alongside low rates of external validation and explainability and no reporting of robustness.9 Nov 2026