It’s 2AM. A 39-year-old female patient lies in her hospital bed one day after an otherwise uncomplicated gallbladder surgery. Her nurse – relatively new in his career – comes into the room for routine vital signs. The patient seems more tired and slightly more confused than on previous rounds. Still, she is oriented to person, place, and location. Her heart rate is faster than it was 8 hours ago – but not so fast that it meets escalation criteria. The remaining vitals are essentially unchanged and stable. Labs look good.
In the last few months on this floor, the nurse has seen dozens of post-op cholecystectomy patients. As he chats with her, he multitasks through his routine assessment, drawing new labs, auscultating lung sounds, surveilling body and device from head to toe. He can’t shake a feeling that something just isn’t right with this one. But he can’t point to anything that couldn’t be explained away by post-op pain and narcotics.
He hesitantly pages the doctor, knowing she is likely catching up on some sleep and covering 60-plus patients that night. “Should I give the patient the scheduled ACE inhibitor?” he asks, referring to the planned blood pressure medication. The physician answers the medical question that was asked: “Sure,” she says. But she hangs up missing the point.
The nurse is worried the patient is becoming septic. He’s scared about the medication’s impact, but he doesn’t want to create a false alarm.
When the physician arrives to the unit for rounds, it’s 6:30AM. In the hall, the nurse hesitantly explains, “This patient just looks... off,” offering each subsequent observation with the upward lilt of a question more than with conviction. The physician is distracted by a crashing admission down the hall. She acknowledges the nurse’s concern but points to the reassuring vitals and labs before moving to the more acute case.
Three hours later, the next shift calls a rapid response. The patient’s blood pressure is 79/43 and her heart rate is 115. She is clearly deteriorating. The patient is transferred emergently to the ICU in shock, where she requires dialysis and prolonged intubation. She barely survives.
Physician training focuses heavily on diagnosis, treatment planning, and the science of disease. Doctors are trained to prioritize quantitative data – lab results, imaging, vital signs, and clinical exam findings, using pattern recognition honed over time. In comparison to nurses, they tend to be responsible for more patients at any given time and therefore spend fewer hours with any one patient. They are problem-focused, lasered in on answering the question: “What’s wrong and how do we fix that?”
In contrast, nurse training emphasizes patient care, coordination, and advocacy. Nurses are embedded in the rhythm of the patient’s day and far more present at the bedside, allowing them to stay anchored in the here-and-now of a few specific admissions. Their primary question often is: “What actions will best support this patient’s overall health, safety, and comfort right now?”
Both perspectives are critical to patient care and speak to the importance of the whole clinical workforce. However, the functioning of the team can be limited by communication failures. Headlines and legal cases abound with examples of blatant communication breakdowns between doctors and nurses, such as wrong site surgeries and transfusion errors, where objective data was either not communicated or communicated incorrectly, leading to tragic consequences.
But less well-recognized are the subtle failures in messaging that occur daily. In these moments, even though physicians and nurses are speaking to each other, they're doing so in different languages. Neither is truly understanding the other.
The nurse had sensed it coming. Not in the numbers, but in the small, cumulative cues he had picked up at the bedside over many rounds with the patient: subtle facial tension, constant rearranging of pillows, half-picked mash potatoes. On their own, none demanded an urgent page; together, they suggested that something was shifting. What was missing was a mechanism to transform that bedside insight into a shared, actionable signal that could be processed by both the nurse and the provider. Without such a structure, the concern was lost and the patient suffered a failure to rescue: severe harm caused not by inattention, but by the system’s inability to capture and act on early warning signs.
That evening, at the start of her new shift, the physician notices that the patient is in the ICU. She has a sinking feeling, flashing back to the conversations with the nurse. The blood pressure pill should have been held, not administered. She could have been more curious on the phone – Tell me more about why you’re asking about the ACE inhibitor. Are you worried? She might have gone into the room to evaluate the patient herself with a more critical eye. She should have trusted the nurse’s gut.
Nursing intuition, cultivated through years of hands-on care and well-studied in nursing literature, often serves as valuable early warning for clinical deterioration. However, it is often under-appreciated, particularly by novice physicians. This divide directly impacts patient outcomes. Preventable cardiac arrests on the floor or complicated ICU stays, in turn, burn hospital resources and stamina across the entire clinical team.
A shared clinical language must validate both the nurse's qualitative insight and the physician's demand for quantitative evidence. Clinical decision support tools can serve as translators, when designed thoughtfully.
With eCART, an all-cause clinical deterioration platform, we display the subtle changes that the nurses pick up on in quantitative terms, integrating trends of predictive values like respiratory rate into our model and visuals. This shows doctors hard data that legitimizes the nurses’ gut.
In turn, we have built a workflow into eCART that standardizes an assessment of nurse worry and includes it in the chart as an objective piece of data, giving it the dues that it deserves. For example, when a rising eCART score nudges a nurse to reassess the patient and document their level of concern, that subjective signal becomes visible to the physician. The physician, in turn, is prompted to conduct their own evaluation—armed with both data and documented bedside intuition.
In a single interface, nurses and doctors can see which vitals or labs are trending, which are influencing the risk score, and whether clinical concern has already been voiced. What results is not just a number, but a conversation – grounded in trust and built for action – that connects them around their shared goal of care.
Mars and Venus will always orbit on different paths, just as physicians and nurses will always bring different but complementary perspectives to patient care. If we want to achieve better real-world outcomes for our patients, we need to equip both types of clinicians with the tools they need to speak the same language: one that will improve trust, collaboration, and most importantly, save lives.
Dana Edelson, MD, MS is a hospitalist and Executive Medical Director of Rescue Care at UChicago Medicine. She is also a Co-Founder & Chief Medical Officer at AgileMD, a clinical decision support platform.
Joe Reid, MSN, RN, is a critical care nurse and former Patient Care Manager & Resuscitation Quality Specialist at UChicago Medicine. At AgileMD, he serves as Customer Success Lead, guiding eCART implementations across our health system partners.
As a nurse-physician dyad, Joe and Dana have worked side-by-side for nearly a decade. They began their partnership at UChicago Medicine, where they helped develop and evolve the rapid response function. At AgileMD, they focus on building EHR-embedded early warning analytics and workflows to identify and act on clinical deterioration, including from sepsis.
© 2011 - Present · AgileMD, Inc.
2261 Market Street #4378
San Francisco CA 94114-1600
United States
