Entity

Time filter

Source Type


Dalai V.V.,University of Texas Health Science Center at Houston | Khalid S.,Center for Cognitive Studies in Medicine and Public Health The New York Academy of Medicine | Gottipati D.,University of Texas Health Science Center at Houston | Kannampallil T.,Center for Cognitive Studies in Medicine and Public Health The New York Academy of Medicine | And 4 more authors.
Artificial Intelligence in Medicine | Year: 2014

Objective: Clinicians' attention is a precious resource, which in the current healthcare practice is consumed by the cognitive demands arising from complex patient conditions, information overload, time pressure, and the need to aggregate and synthesize information from disparate sources. The ability to organize information in ways that facilitate the generation of effective diagnostic solutions is a distinguishing characteristic of expert physicians, suggesting that automated systems that organize clinical information in a similar manner may augment physicians' decision-making capabilities. In this paper, we describe the design and evaluation of a theoretically driven cognitive support system (CSS) that assists psychiatrists in their interpretation of clinical cases. The system highlights, and provides the means to navigate to, text that is organized in accordance with a set of diagnostically and therapeutically meaningful higher-level concepts. Methods and materials: To evaluate the interface, 16 psychiatry residents interpreted two clinical case scenarios, with and without the CSS. Think-aloud protocols captured during their interpretation of the cases were transcribed and analyzed qualitatively. In addition, the frequency and relative position of content related to key higher-level concepts in a verbal summary of the case were evaluated. In addition the transcripts from both groups were compared to an expert derived reference standard using latent semantic analysis (LSA). Results: Qualitative analysis showed that users of the system better attended to specific clinically important aspects of both cases when these were highlighted by the system, and revealed ways in which the system mediates hypotheses generation and evaluation. Analysis of the summary data showed differences in emphasis with and without the system. The LSA analysis suggested users of the system were more "expert-like" in their emphasis, and that cognitive support was more effective in the more complex case. Conclusions: Cognitive support impacts upon clinical comprehension. This appears to be largely helpful, but may also lead to neglect of information (such as the psychosocial history) that the system does not highlight. The results have implications for the design of CSSs for clinical narratives including the role of information organization and textual embellishments for more efficient clinical case presentation and comprehension. © 2014 Elsevier B.V. Source

Discover hidden collaborations