Advancing clinical decision support using lessons from outside of healthcare: an interdisciplinary systematic review
1 RAND Corporation and Pardee RAND Graduate School, 1776 Main St, Santa Monica, CA, USA
2 David Geffen School of Medicine, University of California, Los Angeles, CA, USA
BMC Medical Informatics and Decision Making 2012, 12:90 doi:10.1186/1472-6947-12-90Published: 17 August 2012
Greater use of computerized decision support (DS) systems could address continuing safety and quality problems in healthcare, but the healthcare field has struggled to implement DS technology. This study surveys DS experience across multiple non-healthcare disciplines for new insights that are generalizable to healthcare provider decisions. In particular, it sought design principles and lessons learned from the other disciplines that could inform efforts to accelerate the adoption of clinical decision support (CDS).
Our systematic review drew broadly from non-healthcare databases in the basic sciences, social sciences, humanities, engineering, business, and defense: PsychINFO, BusinessSource Premier, Social Sciences Abstracts, Web of Science, and Defense Technical Information Center. Because our interest was in DS that could apply to clinical decisions, we selected articles that (1) provided a review, overview, discussion of lessons learned, or an evaluation of design or implementation aspects of DS within a non-healthcare discipline and (2) involved an element of human judgment at the individual level, as opposed to decisions that can be fully automated or that are made at the organizational level.
Clinical decisions share some similarities with decisions made by military commanders, business managers, and other leaders: they involve assessing new situations and choosing courses of action with major consequences, under time pressure, and with incomplete information. We identified seven high-level DS system design features from the non-healthcare literature that could be applied to CDS: providing broad, system-level perspectives; customizing interfaces to specific users and roles; making the DS reasoning transparent; presenting data effectively; generating multiple scenarios covering disparate outcomes (e.g., effective; effective with side effects; ineffective); allowing for contingent adaptations; and facilitating collaboration. The article provides examples of each feature. The DS literature also emphasizes the importance of organizational culture and training in implementation success. The literature contrasts “rational-analytic” vs. “naturalistic-intuitive” decision-making styles, but the best approach is often a balanced approach that combines both styles. It is also important for DS systems to enable exploration of multiple assumptions, and incorporation of new information in response to changing circumstances.
Complex, high-level decision-making has common features across disciplines as seemingly disparate as defense, business, and healthcare. National efforts to advance the health information technology agenda through broader CDS adoption could benefit by applying the DS principles identified in this review.