EUVIMEDEUVIMED entdecken
QuerschnittstudieOpen Access · CC BY 4.0

Patient Perspectives on Artificial Intelligence in Diagnosis and Care: A Cross-Sectional Survey in a Romanian Hospital with Reference to Romanian and European Union Legal Provisions

Cristina Modiga, Sorin Dan Chiriac, Ștefania Dinu, Cristina Ioana Talpoș Niculescu, Adriana Pădure, Bogdan Hoinoiu, Alina Doina Tănase

Nursing Reports · 17.11.2026 · 16 (11) · 416

DOI: 10.3390/nursrep16110416 ↗

Kurz gefasst

Querschnittstudie in der Zeitschrift Nursing Reports, online erschienen am 17.11.2026. Verfasst von Cristina Modiga und 6 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background and Objectives: Artificial intelligence (AI) is increasingly used to support diagnosis and clinical care, but Romanian patients’ expectations regarding disclosure and consent remain insufficiently characterised. We assessed AI awareness, trust, disclosure preferences and legal-rights knowledge among hospital outpatients to inform responsible clinical integration. Materials and Methods: This cross-sectional questionnaire survey included 86 adults, comparing 34 respondents with prior personal use of AI tools for health questions with 52 without prior use. Outcomes included AI trust, legal-rights knowledge, disclosure preferences, acceptance of AI-assisted care and comfort with AI-only interpretation. Group comparisons were performed, and pre-specified exploratory adjusted analyses, including reduced and penalised logistic models, examined factors associated with wanting disclosure.

Results

Overall, 86.0% wanted disclosure of AI use, 58.1% wanted explicit written consent and 7.0% were comfortable with AI-only interpretation. Prior users reported higher trust scores (65.2 versus 53.1; difference 12.1, 95% confidence interval 2.0–22.2; p = 0.020) and legal-rights knowledge (median 4 versus 2; p = 0.011), while disclosure preferences were similar (85.3% versus 86.5%). Legal-rights knowledge was associated with wanting disclosure in the exploratory adjusted model (odds ratio 1.79 per correct item, 95% confidence interval 1.13–2.83), an association that persisted in reduced and penalised models, although the small number of negative responses (n = 12) limits the precision of these estimates. The adjusted association between prior use and trust was not statistically significant.

Conclusions

Most respondents expected disclosure and continued physician responsibility. Familiarity with AI did not remove this expectation. Responsible integration of AI into diagnosis and care should include accessible explanations within patient communication and consent procedures; the findings apply to Romanian hospital outpatients, and larger multicentre studies are needed to assess generalisability. Nurses, as the usual first point of contact, are well placed to deliver these explanations.

Zitieren

APA

Modiga, C., Chiriac, S. D., Dinu, Ș., Niculescu, C. I. T., Pădure, A., Hoinoiu, B., & Tănase, A. D. (2026). Patient Perspectives on Artificial Intelligence in Diagnosis and Care: A Cross-Sectional Survey in a Romanian Hospital with Reference to Romanian and European Union Legal Provisions. Nursing Reports, 16(11), 416. https://doi.org/10.3390/nursrep16110416

Vancouver

Modiga C, Chiriac SD, Dinu Ș, Niculescu CIT, Pădure A, Hoinoiu B, et al. Patient Perspectives on Artificial Intelligence in Diagnosis and Care: A Cross-Sectional Survey in a Romanian Hospital with Reference to Romanian and European Union Legal Provisions. Nursing Reports. 2026;16(11):416. doi:10.3390/nursrep16110416

BibTeX
@article{Modiga2026,
  author = {Modiga, Cristina and Chiriac, Sorin Dan and Dinu, Ștefania and Niculescu, Cristina Ioana Talpoș and Pădure, Adriana and Hoinoiu, Bogdan and Tănase, Alina Doina},
  title = {Patient Perspectives on Artificial Intelligence in Diagnosis and Care: A Cross-Sectional Survey in a Romanian Hospital with Reference to Romanian and European Union Legal Provisions},
  journal = {Nursing Reports},
  year = {2026},
  volume = {16},
  number = {11},
  pages = {416},
  doi = {10.3390/nursrep16110416}
}
Kostenlos starten →