Several broad areas of recent and current research interest. Links will be provided soon for specific projects
Medicine is a rich and complex domain, with scientific evidence of disease at its core. The process of care faces the formidable task of making an accurate diagnosis and planning the optimal treatment in an uncertain and dynamic environment. Frequently, problems are ill-structured, with shifting, poorly defined, or competing goals. Critical data may not be available at the time when an important decision has to be made, may be incomplete or be just mere estimates. Time stress and high stakes often apply additional load on clinicians working in such a demanding setting.
Medical professionals are assisted in their decision making and planning by clinical information systems that aggregate information from many sources, process it, and present it to the user either as a comprehensive analysis of the problem in question, or offer specific suggestions at the immediate point of decision. Interaction with dynamic clinical systems may change the way clinicians obtain, organize and reason with medical knowledge. Investigations in the field of medical cognition attempt to characterize and explain the cognitive processes central to medical reasoning and decision making
Our ongoing research project is directed towards assessing the effects of computer-based medical ordering on the cognitive behavior of physicians and other professionals involved in care. One of our goals is to investigate whether training physicians (interns, residents) who work primarily or exclusively with physician order entry, will develop different reasoning strategies from those writing mostly paper orders. We also strive to develop a cognitive model of an ordering task that would be system-independent, and identify areas and tasks that either contribute to or alleviate cognitive load of users. Attention and cognitive resources should be focused to problem states, outcomes and strategy planning (e.g., a treatment plan), while peripheral or routine tasks should be highly automated.
This research encompasses the areas of clinical computing, human-computer interaction, cognitive science, medical error and patient safety. It is intended to build a framework of concepts, vocabulary and design principles necessary for scientific discourse among developers of clinical software. As a result, a new set of design considerations may be proposed, specific to medical environments, that will inform the development of systems that support clinicians in completing their tasks. Specifically, using theoretical foundations from cognitive science, I investigate the interaction of clinicians with a complex provider order entry system (CPOE).
The process of computer-based clinical ordering is frequently made excessively cognitively demanding by poorly designed interfaces. We are developing methodology to characterize the sources of unnecessary cognitive complexity of the interface. Specifically, we investigate the interaction of clinicians with a complex provider order entry system (CPOE) using theoretical foundations from cognitive science and an explanatory framework based on the theory of distributed cogntion. The goal is to characterize the nature of cognitive demands imposed by the interaction and its effects on usability, user performance and medical error. Overly complex or inconvenient interfaces that are difficult to navigate claim a disproportionate share of human limited working memory resources. In effect, they divert focus away from the main clinical task and delay its completion. Well designed interfaces allow users to focus primarily on higher-order cognitive activity, such as clinical reasoning and treatment planning. Also, making the ordering process fit better into established workflow routines may help overcoming the frequently strong initial opposition of clinicians to CPOE and smooth the progress of large-scale implementation of this safety-enhancing technology in US hospitals