Public hesitancy for AI-based detection of neurodegenerative diseases in France

Abstract Recent advances in artificial intelligence (AI) have made it possible to detect neurodegenerative diseases (NDDs) earlier, potentially improving patient outcomes. However, AI-based detection tools remain underutilized. We studied individual valuation for early diagnosis tests for NDDs. We c...

Full description

Saved in:
Bibliographic Details
Main Authors: Ismaël Rafaï, Bérengère Davin-Casalena, Dimitri Dubois, Thierry Blayac, Bruno Ventelou
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-025-11917-8
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Recent advances in artificial intelligence (AI) have made it possible to detect neurodegenerative diseases (NDDs) earlier, potentially improving patient outcomes. However, AI-based detection tools remain underutilized. We studied individual valuation for early diagnosis tests for NDDs. We conducted a discrete choice experiment with a representative sample of the French adult population (N = 1017). Participants were asked to choose between early diagnosis tests that differed in terms of: (1) type of test (saliva vs. AI-based tests analysing electronic health records); (2) identity of the person communicating the test results; (3) sensitivity; (4) specificity; and (5) price. We calculated the weights in the decision for each attribute and examined how socio-demographic characteristics influenced them. Respondents revealed a reduced utility value when AI-based testing was involved (valuated at an average of €36.08, CI [€22.13; €50.89]) and when results were communicated by a private company (€95.15, CI [€82.01; €109.82]). We interpret these figures as the shadow price that the public attaches to medical data privacy. Beyond monetization, our representative sample of the French population appears reluctant to adopt AI-powered screening, particularly when performed on large sets of personal data. However, they would be more supportive when medical expertise is associated with the tests.
ISSN:2045-2322