Information shapes decisions: The access and use of decision support

The decision to access and use Decision Support (DS) is complex, influenced not only by the system’s properties, such as cost and accuracy, but also by users’ evolving perceptions of its utility. While earlier research has explored the decision to use DS, few studies have investigated how users’ past interactions with DS influence their future access decisions. This study addresses this gap by examining the dynamic relationship between DS use and subsequent access behavior. We employed Signal Detection Theory (SDT) to model human binary classification decisions under uncertainty, including the theoretically optimal strategy, defined as the normative behavior that maximizes expected utility, for accessing decision support (DS). In an experiment involving 140 participants, each participant decided whether to access DS to obtain binary information about a simulated patient’s condition and determined whether the patient required an intervention. Results showed that using DS improved decision efficiency by reducing classification time and enhancing decision accuracy. However, participants did not always utilize DS optimally, achieving lower performance than predicted by the normative model. Participants’ access decisions were influenced by their prior experiences with the DS. Greater performance improvement was associated with a higher likelihood of repurchasing the DS. Thus, the discretionary use of DS is shaped not only by normative reasoning but also by experience-driven decision strategies. Our study demonstrates how decisions to access algorithmic decision support evolve, underscoring the need to consider these dynamics when designing and deploying decision aids in real-world contexts.