Diagnosis is complicated, but researchers are beginning to understand the factors that help and hinder the process in teaching hospitals.
A recent study documented approaches to diagnosis on four teams at two academic medical centers. Observers recorded field notes during trainees' activities before rounds, morning rounds with the entire team, and afternoon work that included senior residents, interns, and students. They collected information on both the diagnostic process itself (e.g., gathering data, integration and interpretation, creating a working diagnosis, treatment delivery, outcomes) and the work system (e.g., team members, organization, physical environment, technology and tools, tasks).
Researchers identified four key themes of diagnosis: 1) It is a social phenomenon, 2) The data necessary to make diagnoses are fragmented, 3) Distractions and interruptions consistently interfere with the process, and 4) Time pressures impede decision making. While the work system facilitated social interactions, it contributed to data fragmentation and distractions that led several clinicians to use headphones to drown out noise. In addition, trainees often had to skip learning sessions to make time for diagnosis.
Given these challenges, teaching hospitals need to adopt targeted strategies to improve diagnosis and reduce errors, according to lead study author Vineet Chopra, MD, MSc, FACP, associate professor of medicine and chief of the division of hospital medicine at Michigan Medicine in Ann Arbor. He recently spoke with ACP Hospitalist about the implications of the findings, which were published online in April 2018 by the Journal of Hospital Medicine.
Q: What led you to study this issue?
A: Diagnosis is perhaps the most important task any hospitalist does—yet we know little about the process, especially as it pertains to those of us that work in academic medical centers. We therefore wanted to better understand the ways in which diagnoses are made and some of the barriers to making them.
Q: What surprised you the most about your findings?
A: Perhaps the two most surprising aspects were just how fragmented and disparate the information sources needed for diagnosis were, as well as how commonly distractions from diagnosis occurred. While I think we intuitively knew these would be relevant problems, it was surprising to see learners and faculty at different stages of their career all zoom into the same issues.
Q: What makes diagnosis a social phenomenon, and what are the implications of that?
A: This was a very interesting finding during our study. Diagnosis is not made by a single person or at a single point in time. It's a constant iteration of facts, findings, and new information that occurs in the context of many people thinking about the same patient. Rounding in the morning is a quintessential information exchange where diagnoses are made, shared, and reflected on by many people on the team. In fact, this is a strength in some way of resident teams where many people look at the same fact and make judgments regarding what fits, what doesn't, and why. You can't help but wonder when those layers are removed—the social fabric in which diagnoses are discussed—and you have hospitalists flying solo for direct care, whether things might worsen. In effect, there are no safeguards for wrong diagnoses or flawed interpretations, and more errors may happen.
Q: What are your suggestions for mitigating distractions and time pressures that interfere with the diagnostic process?
A: We think the best way to focus on diagnosis is to create carved-out time for the process. This is no different than how we think about surgical timeouts before major procedures, so there is precedence for it. One approach might be a team huddle at the end of the day where every member of the team hears about all their patients, updated labs, and provides their input on what might be going on [and] whether the workup or treatment appears appropriate (the so-called collective wisdom approach). Another approach might be a more structured format, such as a checklist, that forces certain cognitive tasks within the diagnostic process carved out in defined space and time. We think both of these have merit and are testing these in an ongoing study.
Q: Can you provide more detail about the study?
A: We are starting a pilot study that embeds a checklist for diagnosis within the workflow of admissions for hospitalists. This checklist forces a pause within admissions, including questions such as considering differential diagnoses, actions such as personally reviewing EKGs and X-rays, as well as reaching out to consultants should there be concerns regarding unexplained findings. We will also be measuring biological variables from hospitalists within the study using a [wearable] device...[that] measures respirations, anxiety, calm and focused states, and we think the checklist will encourage those versus [in participants] that do not have the same exposure.
Q: How might teaching hospitals improve diagnosis overall?
A: I think teaching hospitals have a unique advantage in improving diagnosis in that they can study it in ways others cannot. If you think about “teaching” diagnosis, there is often little formal guidance given to most clinicians. Developing a curriculum around diagnosis therefore might be one approach where teaching facilities can pioneer the way. The other approach is to innovate around elements such as diagnostic timeouts, checklists, as well as shared decision making with others. Both of these have promise and can be well addressed in a teaching hospital environment. After all, if we can't figure this out, study it, and teach it—who else?