Narrative Review of Decision-Making Processes in Critical Care

Abstract

Several theories describing the decision-making process in the intensive care unit (ICU) have been formulated. However, none of them appreciate the complexities of the process in an eclectic way by unifying several miscellaneous variables in 1 comprehensive theory. The purpose of this review is to highlight the key intricacies associated with the decision-making process in the ICU, to describe the theoretical frameworks with a special emphasis on gaps of knowledge, and to offer some avenues for improvement. The application of theoretical framework helps us to understand and to modify the structure of the process. Expected utility theory, regret theory, prospect theory, fuzzy-trace theory, construal level theory, and quantum probability theory were formulated over the years to appreciate an increased complexity of the decision-making process in the ICU. However the decision makers engage, these models may affect patient care because each of these frameworks has several benefits and downsides. There are gaps of knowledge in understanding how physicians match the different theoretical frameworks of the decision-making process with the potentially high ICU variability and load, especially when the "best outcome" is often nondiscrete and multidimensional. Furthermore, it is unclear when the preferential application of reflexive, habitual, variable, and biased-prone processes results in patient and staff detriment. We suggest better matching of theoretical frameworks with strengths of the human decision-making process and balanced application computer aids, artificial intelligence, and organizational modifications. The key component of this integration is work to increase the self-awareness of decision-making processes among residents, fellows, and attending physicians.

Future Medical Artificial Intelligence Application Requirements and Expectations of Physicians in German University Hospitals: Web-Based Survey

“…While a balance between the increasing amount of documented data on the one hand and the demographic change and aging populations on the other hand challenges our health care systems, big data and artificial intelligence (AI) in medicine offer a huge potential to relieve physicians from the increasing complexity of today's health care and information overload when treating patients [1,2]. Over the last decade, research on AI in medicine and biomedicine and the number of publications in these fields have substantially increased [3,4].…”

Examination of Electrolyte Replacements in the ICU Utilizing MIMIC-III Dataset Demonstrates Redundant Replacement Patterns

“…Even when standardized protocols are implemented to improve patient outcomes, adherence is scattered [10,20,[25][26][27][28][29]. Patterns of decision making overlay this complex network of factors, affecting electrolytes in healthcare in general [7,15,20,30]. The intensive care unit environment exaggerates all these factors while adding the time pressure of a highly dynamic and cognitively challenging environment [7,9,15,[30][31][32][33].…”

“…Patterns of decision making overlay this complex network of factors, affecting electrolytes in healthcare in general [7,15,20,30]. The intensive care unit environment exaggerates all these factors while adding the time pressure of a highly dynamic and cognitively challenging environment [7,9,15,[30][31][32][33]. Simultaneously, medical training struggles to adequately equip medical professionals with the skills to execute decisions rooted in evidence and based on heuristics [9,[34][35][36].…”

“…Current patterns and the available literature suggest superfluous electrolyte repletion patterns in the ICU [30]. There is also a gap in knowledge in understanding how electrolyte lab values are analyzed and considered when electrolytes are replaced or not.…”

“…There is also a gap in knowledge in understanding how electrolyte lab values are analyzed and considered when electrolytes are replaced or not. Furthermore, a prior study did not explore the nature of electrolyte replacement thresholds (discrete or continuous), did not look into scenarios where electrolytes were not repleted, and was subjected to bias, as the study came from one institution [30]. Extending the research project to MIMIC-III allows for cross-validation of the results of the prior data, especially in terms of efficiency of the electrolyte replacements [30].…”