An Empirical Study on Trust, Bias, and Decision Quality to Assess the Effect of AI-Powered Decision Support Systems on Managerial Decision-Making in Healthcare

There is increased usage of Artificial Intelligence-based Decision Support System (AI-DSS) in healthcare industry, but no empirical data to indicate their influence on managerial decisions. To analyze the entire impact of AI-DSS on managerial decision-making within healthcare firms, this research paper examines the relationships between trust in AI, minimization of biases, and the quality of decisions.
These were mixed-methods underlying empirical data and combined survey-based measures of perceived trust, mental bias, reliance behavior, and decision accuracy with experimental studies of decision-making control. Quantitative analysis involves the application of regression and structural equation modeling to evaluate the impact of trust and decision outcomes and the features of AI-DSS design, such as explainability, transparency, and the quality of feedback, determine how much managers are vulnerable to bias.
Through analyzing the opinions of the managers about autonomy, dependability and system constraints of AI, qualitative insights can be used to complement these findings. The results indicate that over- and under-reliance are harmful to the results, but balanced trust in AI-DSS significantly enhances the quality of decisions.
The study contributes to the body of literature by presenting an evidence-based paradigm, which demonstrates the dynamics between bias and trust and the level of their interaction in influencing the management decision performance in the AI-supported setting. Suggestions on how to improve the design of an AI-DSS interface, reduce bias, and train management to achieve success in human-AI cooperation in healthcare facilities are just some of those practical implications.
Keywords— Artificial Intelligence, Decision Support Systems, Trust, Bias, Decision Quality, Managerial Decision-Making, Healthcare Administration, Human–AI Interaction, Algorithmic Transparency.