Determinants of Trust in Artificial Intelligence (AI) for Health-Related Decision-Making Among Adults in Saudi Arabia: A Cross-Sectional Study

TL;DR

This study explores what influences trust in AI for healthcare decisions among adults in Saudi Arabia, finding that patient satisfaction and relationships with doctors play key roles.

Contribution

The study introduces a theory-informed mediation pathway linking patient–doctor relationships, patient satisfaction, and trust in AI in a Saudi Arabian context.

Findings

Patient satisfaction was positively associated with trust in AI for health decisions.

Stronger patient–doctor relationships were linked to lower trust in AI, possibly due to reliance on physicians.

Older adults showed less trust in AI compared to younger adults.

Abstract

Background/Objectives: Artificial intelligence (AI) is increasingly integrated into healthcare decision-making. Public trust in AI remains a critical determinant of its acceptance and effective use. Evidence on the factors shaping trust in AI within Middle Eastern contexts, particularly Saudi Arabia, remains limited. Therefore, we aimed to identify the determinants of trust in AI for health-related decision-making and to examine a theory-informed mediation pathway in which patient satisfaction mediates the association between patient–doctor relationships and trust in AI. Methods: We conducted a cross-sectional, facility-based survey of adults in Saudi Arabia, using an electronic questionnaire distributed in four primary healthcare centers. We performed multiple linear regression to assess the association of trust in AI for health-related decision-making with patient satisfaction, patient–doctor relationships, sociodemographic characteristics, and healthcare-related factors. A mediation analysis was also employed to evaluate the indirect and direct association linking patient–doctor relationships, patient satisfaction, and trust in AI. Results: Our findings showed that patient satisfaction was positively associated with trust in AI (β = 0.54, 95% CI: 0.18–0.90), while patient–doctor relationships showed an inverse association (β = −0.34, 95% CI: −0.48 to −0.20), possibly reflecting a greater reliance on physicians’ clinical judgment and a reduced perceived need for AI-supported decision-making. Trust in AI varied across age groups, with a lower trust observed in older age categories compared with younger adults. No strong associations were observed for sex, education, body mass index, or healthcare-related factors. Patient–doctor relationship quality was indirectly associated with trust in AI via patient satisfaction (ACME = 0.138, 95% CI: 0.043–0.246), alongside a direct association with trust in AI (ADE = −0.313, 95% CI: −0.456 to −0.160). This means that patient–doctor relationships influenced trust in AI both directly and indirectly through patient satisfaction, suggesting that, while interpersonal care may reduce the reliance on AI (direct effect), enhancing patient satisfaction can partially offset this effect and promote trust in AI (indirect effect). Conclusions: These findings highlight that fostering patient-centered care and satisfaction may be crucial for promoting public trust in AI, which has important implications for AI governance, ethical deployment, and the design of AI-supported healthcare systems.

Figures