TITLE:
Exploring Romanian Healthcare Worker Perceptions of AI Integration in Resource-Variable Clinical Settings: A Cross-Sectional Survey Study
AUTHORS:
Anne Patton, Barry Kolar, Eustace Ohuakanwa, Lexy Farrington, Radu Ioan Malancea, Archie Heddings
KEYWORDS:
Artificial Intelligence, Healthcare Professionals, Technology Acceptance, Digital Health, Romania
JOURNAL NAME:
Open Journal of Applied Sciences,
Vol.16 No.11,
November
3,
2026
ABSTRACT: Background: Artificial intelligence (AI) is rapidly transforming healthcare through applications that improve diagnostic support, workflow efficiency, and clinical documentation. However, most AI development and implementation have occurred in high-income countries, and little is known about healthcare workers’ perceptions of AI in resource-variable clinical settings. This study evaluated healthcare workers’ exposure to and perceptions of AI at a regional hospital in Northeastern Romania. Methods: A cross-sectional survey study was conducted in June 2025 at Mavromati Emergency County Hospital to explore healthcare worker perceptions of Artificial Intelligence into clinical care. Paper surveys with English and Romanian Language translations were distributed to physicians, nurses, residents, and practicants across multiple hospital departments. The questions included multiple-choice, Likert-style, and select-all-that-apply responses. Participation was voluntary and anonymous. Completed surveys were manually entered into Microsoft Excel, and descriptive statistical analysis was performed using frequencies and percentages. Results: One hundred healthcare professionals completed the survey, including 62 nurses and 32 physicians. Most respondents (61%) reported no prior exposure to AI in clinical practice, while attitudes toward AI were predominantly neutral (58%). Documentation (33%), administrative tasks (23%), and patient follow-up (20%) were identified as the most valuable potential applications of AI. Nurses were significantly more likely than physicians to identify patient safety as a concern (56% vs. 31%, p = 0.020), whereas physicians more frequently reported concerns regarding misdiagnosis (59% vs. 31%, p = 0.007). Hands-on demonstrations (51%), additional training (48%), and institutional policies (39%) were the most identified resources needed to support AI implementation. Conclusion: Healthcare workers in this Romanian clinical setting demonstrated cautious optimism toward AI despite limited prior exposure. Respondents were most receptive to AI applications that improve workflow efficiency while preserving clinician oversight, with concerns centered on patient safety, diagnostic reliability, and maintaining the clinician-patient relationship. Successful AI implementation in resource-variable healthcare settings should prioritize practical training, clear governance, and gradual integration of workflow-support applications before broader clinical deployment. Study Design: Cross-sectional survey study.