ASSESSING THE IMPACT OF ARTIFICIAL INTELLIGENCE ON CLINICAL DECISION-MAKING FOR OLDER ADULT HEALTH AND AGING CARE IN PRIMARY HEALTH CARE: A CROSS-SECTIONAL SURVEY OF HEALTHCARE PROFESSIONALS

Najlaa Saadi Sheet AL-Safar*

ABSTRACT

Background: Artificial intelligence (AI) is increasingly used to support clinical decision-making, but primary health care (PHC) implementation for older adults requires evidence on workforce readiness, perceived benefit, trust, professional oversight, and governance. Objective: To assess healthcare professionals‘ familiarity with and use of AI, perceived decision-support benefit, perceived benefit for older-adult care, trust and human oversight, governance concerns, and implementation conditions in PHC. Methods: A cross-sectional electronic survey dataset containing 575 completed response rows was analyzed as supplied. The current workbook contains no unique respondent identifier or timestamp field that would justify record-level deduplication. Thirty-four attitudinal items were scored on a 5-point Likert scale. Descriptive statistics, Cronbach alpha, Spearman correlations, and HC3-robust multivariable linear regression were applied. Results: Perceived decision-support benefit averaged 3.55/5 (SD 0.76), and perceived benefit for older-adult care averaged 3.52/5 (SD 0.68). The model explained 62.7% of the variance in perceived older-adult-care benefit (adjusted R²=0.624). Trust and human oversight was the strongest independent correlate (standardized β=0.67, 95% CI 0.61 to 0.74, P<.001); readiness (β=0.13, P<.001) and AI use (β=0.07, P=0.037) were also independently associated, whereas familiarity and governance concerns were not statistically significant after adjustment. Only 21.7% agreed that they had received adequate AI training. Conclusions: Healthcare professionals perceived meaningful potential for AI-supported decision-making and older-adult care, but implementation readiness remained uneven. Safe adoption should prioritize calibrated trust, human professional oversight, competency-based training, explainability, privacy, fairness, accountability, and prospective evaluation in real PHC workflows.

Keywords: artificial intelligence; primary health care; clinical decision support; older adults; geriatrics; trust; explainable AI; human oversight; survey.