Presentation
Data Aquisition In The Emergency Room, It Is Harder Than You Think!
SessionPoster Session 2
DescriptionBackground/Introduction:
Diagnostic failure is an area of patient safety in which much of the literature is either retrospective or based on expert opinion. The literature primarily cites faulty cognitive processes including shortcomings of physicians with information gathering and synthesis as major contributors to diagnostic breakdowns, however these conclusions have been based on discussions closer to the realm of work as imagined (WAI) with lack of understanding of work as done (WAD). To close the gap between WAI and WAD, we sought to further understanding of how physicians gather information by completing a prospective observational study of physicians in the pediatric emergency department, a clinical arena chosen for its diagnostically dense work. Through a data collection tool created and inspired by the Systems Engineering Initiative for Patient Safety (SEIPS) model, we recorded and timed tasks physicians were doing along with who was present and what tools were utilized with a goal of illustrating the complexity of diagnostic work being completed. The aim of this study was to observe how physicians acquire information about patients to gain understanding and insights into diagnostic work not captured with retrospective methods.
Methods:
In the pediatric emergency department, four-hour blocks of direct observations of the attending physicians were completed in the summer of 2023. A novel tool was created in REDCap and designed using the lens of SEIPS to capture information including tasks completed, timing of those tasks, interactions with people, tools/technology used, and location where it took place. This data was captured in context of arbitrarily selected patients, labeled as the “index patients” from the moment the physician heard the patient’s story through the first 50 minutes of diagnostic work. Any work observed for patients other than index patient was labeled as “other patients.” The tasks recorded were categorized into data acquisition as a known component of diagnostic work, with the other tasks being categorized as other or therapeutic.
Results:
We observed 26 attending physicians caring for 66 index patients. We captured 2,836 tasks. Physicians took histories for only a subset (65%, 43/66) of the 43 patients (M = 1.5 min, Median = 1.3 in duration). About 66% of physician time was spent completing information acquisition necessary for diagnostic work, and 33% doing therapeutic work. We documented and characterized a multitude of information sources including electronic health record (EHR) (and embedded tools/apps), non-EHR related technology, health care personnel, and the patient/family. The most common tasks used to gather information were performing a physical exam, staffing a patient, and taking a history, and less commonly recorded than expected was the task, ‘reading about the patient’ in the EHR. Instead, we observed with the EHR used primarily while the physicians were hearing about the patient from the resident or advanced practice provider (APP). We represented findings of in a bubble chart illustrated a wide distribution of sources collected between EHR related sources, other personnel, and the family.
Conclusion:
Diagnosis is a process that is not a straightforward “data in, data out’ process. Our results show the complexity of what is required to acquire diagnostic information, and it is different from what you expect. In fact, we found that to perform diagnostic work a history taken from the patients was not necessarily required. What we did find was an impressive distribution of sources between people, EHR -related technologies, and non-EHR related technology.
Diagnostic failure is an area of patient safety in which much of the literature is either retrospective or based on expert opinion. The literature primarily cites faulty cognitive processes including shortcomings of physicians with information gathering and synthesis as major contributors to diagnostic breakdowns, however these conclusions have been based on discussions closer to the realm of work as imagined (WAI) with lack of understanding of work as done (WAD). To close the gap between WAI and WAD, we sought to further understanding of how physicians gather information by completing a prospective observational study of physicians in the pediatric emergency department, a clinical arena chosen for its diagnostically dense work. Through a data collection tool created and inspired by the Systems Engineering Initiative for Patient Safety (SEIPS) model, we recorded and timed tasks physicians were doing along with who was present and what tools were utilized with a goal of illustrating the complexity of diagnostic work being completed. The aim of this study was to observe how physicians acquire information about patients to gain understanding and insights into diagnostic work not captured with retrospective methods.
Methods:
In the pediatric emergency department, four-hour blocks of direct observations of the attending physicians were completed in the summer of 2023. A novel tool was created in REDCap and designed using the lens of SEIPS to capture information including tasks completed, timing of those tasks, interactions with people, tools/technology used, and location where it took place. This data was captured in context of arbitrarily selected patients, labeled as the “index patients” from the moment the physician heard the patient’s story through the first 50 minutes of diagnostic work. Any work observed for patients other than index patient was labeled as “other patients.” The tasks recorded were categorized into data acquisition as a known component of diagnostic work, with the other tasks being categorized as other or therapeutic.
Results:
We observed 26 attending physicians caring for 66 index patients. We captured 2,836 tasks. Physicians took histories for only a subset (65%, 43/66) of the 43 patients (M = 1.5 min, Median = 1.3 in duration). About 66% of physician time was spent completing information acquisition necessary for diagnostic work, and 33% doing therapeutic work. We documented and characterized a multitude of information sources including electronic health record (EHR) (and embedded tools/apps), non-EHR related technology, health care personnel, and the patient/family. The most common tasks used to gather information were performing a physical exam, staffing a patient, and taking a history, and less commonly recorded than expected was the task, ‘reading about the patient’ in the EHR. Instead, we observed with the EHR used primarily while the physicians were hearing about the patient from the resident or advanced practice provider (APP). We represented findings of in a bubble chart illustrated a wide distribution of sources collected between EHR related sources, other personnel, and the family.
Conclusion:
Diagnosis is a process that is not a straightforward “data in, data out’ process. Our results show the complexity of what is required to acquire diagnostic information, and it is different from what you expect. In fact, we found that to perform diagnostic work a history taken from the patients was not necessarily required. What we did find was an impressive distribution of sources between people, EHR -related technologies, and non-EHR related technology.
Event Type
Poster Presentation
TimeTuesday, March 244:45pm - 6:15pm EDT
LocationRhinelander Gallery
Patient Safety Research and Initiatives