Editor’s Pick: Q&A with Kelly Gleason, PhD, RN, FAAN, on the CONCERN study
Real-time surveillance system for patient deterioration: a pragmatic cluster-randomized controlled trial
Sarah C. Rossetti, Patricia C. Dykes, Chris Knaplund, Sandy Cho, Jennifer Withall, Graham Lowenthal, David Albers, Rachel Y. Lee, Haomiao Jia, Suzanne Bakken, Min-Jeoung Kang, Frank Y. Chang, Li Zhou, David W. Bates, Temiloluwa Daramola, Fang Liu, Jessica Schwartz-Dillard, Mai Tran, Syed Mohtashim Abbas Bokhari, Jennifer Thate & Kenrick D. Cato
Nature Medicine, April 2, 2025
https://doi.org/10.1038/s41591-025-03609-7
__
In this Q&A, Anjana Sharma, MD, MAS, CODEX Learning Hub faculty lead, spoke with Kelly Gleason, PhD, RN, a nurse researcher and former bedside nurse who now focuses on improving communication in healthcare settings. Kelly shared her expert insights on the recently published CONCERN study, a landmark trial published in Nature Medicine that demonstrated a significant reduction in hospital mortality through a novel approach to early warning systems.
Dig into the full conversation about what the researchers found and the potential of nursing insights to transform patient care—and how CONCERN is helping surface it.
Q&A Highlights
What do you think is the most important takeaway from the CONCERN study?
It's hard to pick just one, because a huge part of it is that early warning scores have rapidly proliferated throughout the years, but we don't really have strong evidence showing the extent to which they work, especially not in big, randomized trials like this.
What makes this study particularly exciting is that it's not drawing off typical physiological parameters, but markers of concern: Are you taking more vital signs than usual? Are you taking them at times that are not typical, like the middle of the night when care teams minimize interruptions to patients? Are you giving more medications that are as-needed and not scheduled? These are informal ways that people show concern.
The findings are enormous—it shows that you're a third less likely to die, which is significant. And honestly, even getting something into the electronic health record that's used is huge. The way they implemented it was uninterrupted, built into the chart in a way that clinicians can click into for further information, rather than forcing them to stop and acknowledge alerts and interrupt the clinical workflow.
Are there important limitations we should consider before widely implementing this approach?
I wouldn't call it a limitation so much as considerations. It's significant to consider how much charting practices vary not just hospital to hospital, but unit to unit, and even [among nurses] within the same unit.
Sometimes, depending on how the vital signs are documented and how where you’re documenting them [within the system], nurses or patient care techs might take all their vital signs. And then, the busier their shift gets, which can happen with higher acuity patients, the longer they might delay documenting those vital signs.
So, even though they were doing the things that were marked, [I'd consider:] When were thos [vitals signs] getting into the computer? Were they documented at the time they were taken? Was it a hectic shift, so the time wasn’t precise? Not that this would be a bad thing or flaw of the study at all. It could mean nurses are being encouraged to show their worry in this way and it’s more effective.
How does this study relate to diagnostic excellence?
I actually first got into the diagnostic quality world because of a failure-to-rescue inpatient case involving an abdominal aortic aneurysm. When I read this paper, I think of it as being an assistant to faster diagnosis of these conditions. Diagnosing inpatient deterioration is unambiguously part of the diagnostic process that we care about and helping pick up on these concerns that we can't exactly name is such a huge part of the information gathering that happens in the diagnostic process.
Does this study reveal anything about interdisciplinary collaboration in healthcare?
I have extreme nursing pride toward this paper. I do want to acknowledge that many different team members can lead to that increase in documentation that the system picks up on. It could be the family member rushing out to get the nurse saying they're worried. It can be the patient care tech taking vitals out of their own recognition. It can be a physician or medical student walking into the room asking for more information.
So, while the nurse is the most likely person in some cases to do this, it's really showing the whole team's concern in a different way than we typically look at it. Very, very neat findings.
What makes these results particularly significant?
What’s powerful about the fact that this was rigorously as a randomized trial is that they were able to show that moving the patient beforehand [/intervening early] ended up significantly reducing mortality and sepsis to such an extent. What I have seen happen, and this might be more anecdotal, is that when you catch something so soon that you’re able to act on it, then it never happens. Don’t you want rescuing before you need a rescue? I can also leave you with a question of, well, was that rescue ever going to be needed?
The study was done in such a strong that it showed… a third of [those patients] would have passed away had we not done that, which is significant.
__
About the CONCERN Study
The COmmunicating Narrative Concerns Entered by RNs (CONCERN) early warning system is a machine learning model that detects clinical deterioration in hospitalized patients based on how nurses document care. Recently, a multisite randomized trial tested its real-time use in 74 clinical units across two health systems. The study found statistically significant reductions in mortality, sepsis, and length of stay in units using CONCERN.
Notably, the study captures that the warning system is not just about the objective vital sign readings—it captures when healthcare workers are sensing something concerning before they can even put their finger on exactly what's wrong.
About Our Q&A Guest
Kelly Gleason, PhD, RN, FAAN
Associate Professor, Johns Hopkins School of Nursing
Co-Lead, Armstrong Institute’s Center for Diagnostic Excellence’s Team Core
Kelly’s research focuses on quality improvement and patient safety through: 1) Designing and incorporating informatics content in the pre-licensure and doctoral nursing curriculum, 2) Improving the diagnostic process through engaging nurses and patients on the diagnostic team, and 3) Leveraging electronic medical record data to better understand determinants of patient outcomes and improve communication between clinicians and patients.
About Editor's Picks
Curated by the UCSF CODEX team, each Editor’s Pick features a standout study or article that moves the conversation on diagnostic excellence forward. These pieces offer meaningful, patient-centered insights, use innovative approaches, and speak to the needs of patients, clinicians, researchers, and decision-makers alike. All are selected from respected journals or outlets for their rigor and real-world relevance.