This is a list of some of the past research that I have worked on.
My early work in understanding information systems and tools in healthcare work systems led to an investigation of the role information plays in the design of work tools in cognitive work systems, and in keeping the work system self-organized and robust. Further, what role do the people in the system play in keeping the work system orderly? My investigations were supported by a 5-year grant (NIH equivalent of an NSF-CAREER grant funded for $768,528) from the National Library of Medicine (NLM-NIH).
Technology, Cognitive Work and Patient Safety: An Information-Oriented Model (K99/R00), 9/1/2012-8/31/2017, Funding Source: National Library of Medicine, Role: PI
My project focuses on understanding how, why and what information people create, use and share in a system to facilitate goal achievement and work performance, and how designers can effectively use that knowledge about people’s interaction with information for designing information systems. Additionally, my research highlights the nature, form, and attributes of health information, and the tools and technologies health providers use as cognitive supports for their patient care tasks.
One in three patients in the US are likely to experience a medical error. Estimates of adverse events now are ten times more than what IOM predicted ten years ago. Inadequate information support is the second major cause for medical errors and adverse drug events. A lack of timely access to critical patient information, information lost in communication, poor information sharing and transfer provisions, lack of interoperability between information sources, and poor support for provider information needs, all cause problems for safe care, and heighten the need to reevaluate how we design, operate and maintain information support systems. Managing information for safe care, especially in the new EHR environment, continues to be cognitively challenging for care providers. A recent report from an NRC Committee on Engaging the Computer Science Research in Health Care Informatics concludes that healthcare work is knowledge intensive, and that health IT systems must adequately support cognitive work of providers.
I developed an information trail model that tracks information as it is created, used and managed by people. The model captures 16 characteristics of information and models the flow of information across a system based on theories of complexity and self-organization. With this model, I have extended my work to understand information that healthcare providers create, use and manage. The model helps the medical community [both researchers and practitioners] understand how and what information health providers create and use, and how that information is integrated and shared among the different providers. I developed a prototype for representing information content and form in a novel way to best support provider needs and conducted experiments to assess usability and situation awareness.
In addition to projects in information systems and cognitive work in different healthcare areas, I have worked on other mentored projects that cuts across these areas. Two MS theses has resulted from these efforts.
The goals of the MS thesis were to understand diagnostic decision making in healthcare using a teamwork lens. Brennan’s project involved examining the structure and functioning of teams to arrive at a diagnosis. He interviewed healthcare workers in the medical intensive care unit, and used team cognitive work analysis methods and qualitative analysis to develop insights. This current and emerging topic adds to our understanding about diagnostic decision making and diagnostic errors. A manuscript is being prepared for submission to a journal.
Thomas integrated cognitive engineering concepts such as bias and data mining and information modeling methods to identify the presence of cognitive bias while making a decision about presence of sepsis in a patient. Patient and provider data about diagnosis decisions came from an alert system used by the emergency department. Thomas used data mining techniques to study these decisions for the presence of bias. His work not only has immense potential to inform how physicians make decisions using the available information, but also how human factors and data mining can be integrated to provide useful design applications to prevent medical errors.
Cognitive Work Associated with Evidence-Based Pain Management, Multi-PI: Cullen, Pennathur, 1/1/14-12/31/16, Funding Source: Adult Clinical Practice Collaborative Research Award.
Pain management is a critical need of patients managed by nurses through continuous assessment, monitoring and interventions. Pain management is a significant cognitive work activity that nurses perform, and we wanted to understand the intricacies in it. Our pilot work has since gained more importance with the recent opioid crisis, and the need for effective prescribing practices.
The IOM recently identified three imperatives for continuous learning in healthcare: managing increasing complexity, achieving greater value, and capturing opportunities provided by technology/industry. One recommended strategy is to apply systems engineering for quality improvement to remove unnecessary burden on clinicians, enhance the patient experience and, thus improve outcomes. Using cognitive task analysis methods may provide direction for innovative implementation strategies supporting adoption of EBPM.
Quality and safety standards create expectations for provision of evidence-based practice (EBP), yet EBP is not consistently provided. The complexity of EBP recommendations and current workload create a significant challenge for clinicians to integrate EBP into their workflow. Increased demands in nurses’ cognitive work such as cognitive shifts and interruptions, cognitive stacking and lack of care coordination may influence consistent adoption of EBP. System factors such as geographic layout of the facility, technology, and social and organizational aspects of their workplace can affect EBP practices. To understand the impact of cognitive work and system factors on EBP, we chose to address pain because it is frequent, complex and an important clinical condition.
Nursing pain management observations targeted post-operative total knee replacement (TKR) patients, because they often experience severe pain. Inadequate pain control is associated with poor functional recovery. Our exploratory study uses human factors engineering methods to identify cognitive and system factors affecting use of EBP. A mixed method design used cognitive task analysis and contextual inquiry techniques with qualitative and quantitative data collection including shadowing and follow-up interviews, focus group discussions, and nurses’ ratings of workload.
Pain management in Cancer and Human Factors Engineering Perspectives
The opportunity to work on understanding implications of effective pain management and cognitive work of nurses provided impetus to study this topic in related domains such as for nurses who manage pain for cancer patients. Dorota, a PhD candidate in the School of Nursing, and a practicing nurse in the oncology unit has been working with me for a couple of years (initially as an independent study) in exploring cancer pain management from a systems perspective. She has presented the work in many national and international venues.
Other related body of work is in patient safety, and application of human factors to patient safety issues. My contribution to patient safety has been on modeling cognitive and other human factors implications of environment, technologies, teams and processes found in healthcare settings such as operating rooms. I present below some of the significant projects I have carried out.
Study aim is to identify and mitigate hazards during handoffs for cardiac surgery patients. I was involved in study design, development of data collection tools such as interview guides and observation instruments, data collection in cardiac operating rooms, and analysis.
The publication I authored with several collaborators at Johns Hopkins on identification of hazards in the cardiac operating room and a cognitive framework won the Liberty Mutual Award for Best Article in Ergonomics. It was followed by the Liberty Mutual Medal with $10,000 prize for best work in Safety. The paper presented a model for how cognitive factors influenced hazards in the OR developed out of data on hazards in the operating room.
The aim of this study is to understand the impact of multi-faceted patient safety interventions in cardiac ORS, ICUs and surgical floors. In my role as a post-doctoral fellow, I was involved in experimental design, qualitative data collection and qualitative analysis during handoffs and transitions in care.
Study aim is to identify and mitigate hazards during handoffs for cardiac surgery patients. I was involved in study design, development of data collection tools such as interview guides and observation instruments, data collection in cardiac operating rooms, and analysis.
Study aim is understanding communication during multidisciplinary discharge rounds. As a postdoctoral fellow, I was involved in preliminary qualitative content analysis, creation of coded-categories, development of qualitative and behavioral themes of data recorded during multi-disciplinary discharge rounds in inpatient floors, and interviews with providers involved in the discharge process.
I worked with several colleagues in Nursing, College of Pharmacy and UI Healthcare System in patient safety and human factors applications in medication management, facility design, IT design and compliance management. We conducted preliminary work in local pharmacies to understand system factors impacting medication management. We developed a framework using the interview data.
I also worked on projects with colleagues in Internal Medicine on smartphone communication, transitions of care and EHR design. My PhD student Hamed Salehi has published conference proceedings and from this work.
I worked as a co-PI in Prof. Geb Thomas’s AHRQ grant to assist in qualitative data collection and analysis of insights about cognitive errors made when adjusting head of bed angles in an intensive care unit.
Another significant project, which grew out of my graduate course in healthcare human factors and collaboration with L. Cullen was an overwhelming success. It involved designing and implementing a dashboard displaying clinical risk indicators into the EPIC system. The dashboard was designed using human factors principles and evaluated for usability by our PhD IE alumni Mark Schall and Howard Chen under my mentoring. Several conference publications and a journal manuscript have resulted from this work in addition to the implementation and integration with Epic.
As part of my human factors in healthcare course in Spring 2018, in collaboration with L. Cullen students evaluated a smart pump for usability issues. Specifically, nurses found a design error that made the pump not dispense medication. Students evaluated the nursing process when using the pump and the design of the pump. A manuscript is in review.
Volume of Contamination and Nosocomial Infection Control, PI: Perencevich, PI for our aim: L. Herwaldt;9/30/2015-9/29/2018, Centers for Disease Control and Prevention, Role: Co-PI. The project goals were to identify factors that increase or decrease the risk of self-contamination during doffing of different personal protective equipment. We used HFE and ethnographic approaches to identify factors that affect the likelihood of doffing errors and self-contamination. For our human factors assessments, we use hierarchical task analysis and cognitive task analysis methods to identify cognitive workload, information requirements, and sensory requirements when healthcare workers doff PPE during various scenarios.
In my early work during my doctoral studies, we attempted to understand the changes in transitioning from simple manual tools such as whiteboards to electronic technologies like patient tracking systems.
Safety Implications of Computerizing ED Status Boards. (PI: Wears) 07/01/05–06/30/07, Funding Source: Emergency Medicine Foundation
The goals of the project were to evaluate the impact of the transition from a manual whiteboard to an electronic patient tracking system in an emergency room. We wanted to investigate in detail the nature of change in the information healthcare providers use, and the technical aspects of their work with changes in these cognitive tools. We analyzed the information in physical whiteboards and compared them against the content in electronic patient information systems.
This early work was timely as transitions from manual health records to electronic health records were occurring rapidly in the US. But there was limited effort in understanding the human factors aspects, particularly in documenting the cognitive aspects of healthcare work, and how work tools impacted medical errors.
We used photographs of whiteboards and text from interviews of healthcare workers to document the impact of computerization on how they worked, solved problems and made clinical decisions, and communicated with each other during the care process.
Simulating Emergency Department IT Interfaces to Enhance Patient Safety. (PI: Lin), 1U18HS01667, 09/01/06- 08/31/09, Funding Source: Agency for Healthcare Research Quality (AHRQ).
In addition to contributing to knowledge on how technology transitions impact healthcare work systems, when I was a doctoral student at the University at Buffalo, we endeavored to develop usable technology systems for healthcare, and tested the systems with feedback from health care providers.
To understand if a redesigned information system may help reduce workload and improve situation awareness for managing patient information in emergency departments.
Electronic health information systems were beginning to emerge, and there was an increased emphasis on understanding how these systems affect cognitive elements of healthcare work, so we could minimize medical errors.
We created a prototype electronic patient tracking system as a standalone application. The data for the patient tracking system was populated from volume and traffic data obtained from an ED through simulation techniques. We also created patient scripts to realistically simulate patients with different clinical conditions entering the emergency department. We also simulated phone calls and interruptions, making the prototype an immersive experimental setup for evaluating information system use.