Sepsis Rules, Real Patients: When Quality Metrics Collide With Clinical Judgment

Picture the scene unfolding in a busy emergency department: an internal medicine resident presents a patient with suspected urosepsis, reporting that two liters of intravenous saline have resolved the patient’s hypotension. The attending physician does rapid mental arithmetic based on the patient’s listed weight, calculates that the protocol calls for 2,250 milliliters of fluid, glances at the clock, and realizes there are seven minutes left before the three-hour window closes. The attending interrupts and asks the resident to enter an order for a 250-milliliter fluid bolus right now. The resident complies, but is left wondering what medical significance such a tiny amount of saline could possibly carry, and whether the moment represents a genuine teaching opportunity or simply the execution of a checklist obligation. That tension, between what the protocol demands and what the patient actually needs, sits at the heart of a new perspective published in the Journal of General Internal Medicine by Drs. Josué Zapata and Gurpreet Dhaliwal of the San Francisco VA Medical Center and the University of California, San Francisco.

The target of their critique and their pedagogical advice is SEP-1, the Severe Sepsis and Septic Shock Early Management Bundle, a federally mandated quality measure built around a prespecified set of interventions. The bundle requires blood cultures, broad-spectrum antibiotics, lactate measurements, and defined intravenous fluid dosages, all enacted under a strict timeline after severe sepsis or septic shock has been established. Hospitals must adhere to every element of the protocol to receive credit, although there are allowances for documented exclusions, such as a fluid goal not met because of decompensated heart failure. Introduced by the Centers for Medicare and Medicaid Services in 2015, SEP-1 has taken on new weight as it transitioned from a pay-for-reporting measure to a pay-for-performance measure in fiscal year 2026. More than 3,000 hospitals participating in the CMS Hospital Value-Based Purchasing Program now face financial rewards and penalties tied to their SEP-1 performance, and for larger institutions those quality scores can swing millions of dollars in Medicare payments.

The stakes of that reimbursement link ripple directly into the training wards, because clinicians and health systems often apply the full SEP-1 protocol to any patient in whom sepsis is even a consideration rather than risk a compliance failure. Distinguishing sepsis from severe sepsis and septic shock in real time is notoriously difficult and depends on technical definitions, so the safest institutional strategy is broad application of the bundle. The premise of value-based purchasing is that standardizing care will reduce morbidity and mortality, but researchers have questioned the asserted benefit of SEP-1, citing real-world studies that challenge its impact on mortality in undifferentiated sepsis presentations. A 2025 systematic review in Annals of Internal Medicine examined the effect of SEP-1 compliance and implementation on sepsis mortality, and other analyses have found that compliance is associated with improved outcomes for septic shock but not for severe sepsis. Nonetheless, as Zapata and Dhaliwal emphasize, the compliance-reimbursement link has been forged, and it is now the daily reality for hospitals, and therefore for the residents and teachers who work inside them.

What makes SEP-1 different from the financial considerations educators have long folded into their teaching? Documentation requirements to support billing are a well-established part of residency education, a tacit acknowledgment of insurance reimbursement realities, and teaching high-value care similarly asks educators to weigh costs when discussing tests and treatments with trainees. SEP-1 differs by rigidly coupling time-sensitive clinical actions to financial and reputational outcomes for the institution. It is more complex than predecessor quality metrics, and its measurement is substantially more burdensome. It compels extra work such as explaining counterintuitive clinical actions, like the additional 250-milliliter fluid bolus in the opening vignette, or documenting the rationale for mundane omissions, such as why a second lactate level was not measured. Compounding the problem, the very subject of the measurement, sepsis, is a heterogeneous, complex, and contested syndrome whose definitions and best practices continue to evolve, which means the bundle’s fixed 2015-era requirements can conflict with current evidence, including modern recommendations to avoid blood cultures in non-severe pneumonia.

To show how deep those conflicts run, the authors construct a detailed teaching case: a 92-year-old man arrives with lethargy, cough, and dyspnea, a fever of 38.4 degrees Celsius, a heart rate of 108, and a blood pressure of 88 over 58. He weighs 63 kilograms and carries a history of heart failure with an ejection fraction of 25 percent, stage 3 chronic kidney disease, chronic urinary obstruction managed with an indwelling catheter, a prior Pseudomonas urinary tract infection, and recent Clostridioides difficile colitis. His lactate is 2.2 millimoles per liter, his urinalysis shows pyuria, and a respiratory viral panel comes back positive for influenza A. The electronic medical record fires a sepsis alert, establishing time zero, based on suspected infection, two systemic inflammatory response syndrome criteria, and organ dysfunction. After one liter of normal saline and acetaminophen, his temperature, heart rate, blood pressure, and mental status all improve within 90 minutes, and no antibiotics have been given.

Walking through the bundle element by element, the attending in the case demonstrates what the authors call threading the needle. Blood cultures before antibiotics align completely with standard practice, so no extra justification is needed. Broad-spectrum antibiotics within three hours only partially align with clinical judgment: the rapid improvement without antibiotics supports influenza as the likely diagnosis, the pyuria may be a chronic finding rather than acute infection, and reflexive antibiotics carry the risk of recurrent C. difficile colitis, so the attending openly weighs watching the patient for a few hours against the protocol’s three-hour window. The initial lactate has limited prognostic value in this context but fulfills the metric. The repeat lactate required within six hours because the first value exceeded 2.0 directly conflicts with the evidence, which does not support using lactate changes to guide therapy, and the attending asks the resident to call the admitting team, explain that the test is needed for metrics purposes, and document the order. Meanwhile, the ward nurse or inpatient physician must explain to a sleeping, improving patient and his family why an extra blood draw is needed at 2 a.m.

The fluid requirement generates similar friction. The bundle stipulates 30 milliliters per kilogram of crystalloid for patients with hypotension, which for this patient would mean roughly 1.9 liters, yet a single liter has already rendered him euvolemic, and more fluid poses a real risk of hypervolemia given his severe heart failure and advanced kidney disease. The attending instructs the resident to document the rationale for deviating from the stipulated amount so the team is not penalized on sepsis auditing, and to explain the decision to the bedside nurse, who is herself well versed in the sepsis protocol. Each deviation demands timestamping, documentation, and cross-team communication, the invisible administrative labor that SEP-1 generates even when clinicians act entirely appropriately. The authors argue that these moments should not be hidden from trainees but narrated aloud, so that learners can see precisely where clinical reasoning, evidence-based care, and protocol requirements overlap, diverge, or collide.

Beyond real-time narration, the authors describe how post hoc reflection consolidates these lessons and connects protocol-driven actions to broader professional responsibilities. One approach ties bundle-aligned actions to a larger goal that transcends a single patient: an educator might explain that the intensive care unit’s funding last year was contingent on performance on a few CMS metrics, so the patient in front of them may not need the ICU now, but another patient will in six months. Another reframes metrics as instruments for quality improvement rather than compliance, as when an attending describes how benchmarking revealed that the hospital’s heart failure readmissions stemmed from missed goal-directed therapy before discharge, a shortcoming the attending also recognized in their own practice. Faculty can also name a metric’s flaws while clarifying its purpose, sharing stories in which a sepsis alert prompted blood cultures that seemed excessive but later returned positive and changed care, or cases in which weight-based fluid resuscitation, a contested practice with limited supporting evidence, caused harm. Inviting discussion of where SEP-1 fails and where it helps encourages trade-off thinking rather than absolutism, a dialectical habit that transfers to other clinical contexts.

The perspective lands at a moment when pay-for-performance sepsis measurement has become a flashpoint in American medicine, with a joint position paper from major infectious disease and emergency medicine societies already calling for reform of electronic quality measures. Zapata and Dhaliwal’s contribution is neither a defense of SEP-1 nor a call for open defiance, but a practical pedagogy for living inside an imperfect system. When trainees perceive pressure to prioritize institutional objectives over patient care, skepticism and confusion arise, and educators who stay silent risk teaching cynicism by omission. The authors’ proposed stance is candid acknowledgment of a metric’s limitations alongside honest articulation of its utility to patients, clinicians, and institutions. The objective, they write, is not uncritical adherence but the thoughtful demonstration of how to maintain clinical judgment and professional integrity while working within mandates that will not wait for the evidence to settle. Medical educators, on this view, must guide learners through today’s system while preparing them to shape tomorrow’s.

Subject of Research: Teaching clinical judgment alongside the CMS SEP-1 sepsis quality bundle in medical education

Article Title: Teaching to the Test: Threading the Needle on Sepsis Quality Metrics

Article References: Zapata, J., & Dhaliwal, G. (2026). Teaching to the Test: Threading the Needle on Sepsis Quality Metrics. Journal of General Internal Medicine. https://doi.org/10.1007/s11606-026-10710-1

Image Credits: AI Generated

DOI: 10.1007/s11606-026-10710-1

Keywords: sepsis, SEP-1, quality metrics, medical education, pay-for-performance, CMS, clinical judgment, value-based purchasing, residency training, health systems science, evidence-based medicine, sepsis bundle