Related Experiment Videos
How Best to Explain Machine Learning Models to Clinicians: A User Study of Explanation Types
Bowman Brown1,2, Madeline Oguss3,2, Kyle A Carey2,4
1Department of Computer Sciences, University of Wisconsin-Madison, Madison, WI, USA.
Medrxiv : the Preprint Server for Health Sciences
|July 17, 2026
Summary
Clinicians value explanations for machine learning models, with attribution explanations being most effective. Tailoring explanations to roles like nurses and physicians can enhance clinical decision-making.
Area of Science:
- Clinical Informatics
- Machine Learning Interpretability
- Human-Computer Interaction
Background:
- Black-box machine learning models require explanations for clinical adoption.
- Various explanation types exist, but clinician preferences are unknown.
- Understanding clinician valuation of explanations is key to improving AI utility.
Purpose of the Study:
- Evaluate clinician value of different machine learning explanation methods.
- Assess impact on trust, understanding, and clinical thinking.
- Identify preferred explanation formats for clinical workflows.
Main Methods:
- User study with 39 critical care/hospital medicine clinicians (nurses and physicians).
- Comparison of attribution, counterfactual, and rule-based explanations.
- Analysis of effects on trust, understanding, importance, and preferences.
Main Results:
- Clinicians find explanations important, especially physicians post-interaction.
- All explanation types influenced clinicians; attribution explanations showed greatest positive impact.
- Nearly half of clinicians preferred viewing multiple explanation types concurrently.
Conclusions:
- Explanations are crucial for clinical machine learning implementation.
- Prioritize attribution explanations and support multiple types.
- Tailor explanation methods to specific clinical roles (nurses vs. physicians).
Related Concept Videos
01:11Blinding
Blinding is a commonly used method of not telling participants which treatment a subject is receiving. Blinding is a critical part of a randomized control trial or RCT. It reduces the bias that affects the results. In an RCT, blinding is used in the form of a placebo. A placebo effect occurs when untreated subjects falsely believe they have received the treatment and report improved symptoms. A placebo or a dummy treatment is administered to subjects to negate the bias caused by such an effect.
01:22Case Studies
There are many research methods available to psychologists in their efforts to understand, describe, and explain behavior and the cognitive and biological processes that underlie it.
01:21Mechanistic Models: Overview of Compartment Models
Mechanistic models, a category encompassing both physiological and compartmental modeling, differ from empirical models' approaches to incorporating known factors about the systems being modeled. Empirical models describe data with minimal assumptions, while mechanistic models aim to provide a robust description of available data by specifying assumptions and integrating known factors about the system. Compartmental analysis is a key example of a mechanistic model in pharmacokinetics and...