Maternal sepsis remains a leading cause of preventable death in the United States, even as standardized protocols have slashed mortality in hospitals that adopt them. New York State's sepsis bundle, for example, cut sepsis-related maternal deaths by roughly 40% within three years of implementation. Yet a growing body of evidence suggests that one in five cases still slips through—caught too late or not at all. The gap is not evenly distributed: rural hospitals, where lab turnaround times stretch and obstetric specialists are scarce, bear a disproportionate share of missed diagnoses. Clinicians and researchers are now locked in a productive disagreement over whether the problem lies in the protocols themselves or in how they are applied.

Sepsis Protocols Reduce Maternal Deaths but Miss One in Five Cases

The US maternal mortality rate remains stubbornly high compared with other high-income countries. Sepsis accounts for roughly 12–15% of pregnancy-related deaths, a figure that has not budged substantially in the past decade. In response, organizations such as the New York State Department of Health and the California Maternal Quality Care Collaborative rolled out standardized sepsis bundles—checklists of actions to be completed within one hour of suspicion, including blood cultures, lactate measurement, broad-spectrum antibiotics, and intravenous fluids.

Evidence from New York, published in Obstetrics & Gynecology in 2022, showed a 38% reduction in sepsis-related mortality after the bundle was mandated. Similar results emerged from a multi-hospital collaborative in Illinois. But the same studies also reported that roughly 18–22% of patients who ultimately developed severe sepsis or septic shock were not identified by the initial screening criteria used in the bundle. These patients often presented with atypical signs—low-grade fever, mild tachycardia, or no fever at all—that fell below the protocol's vital sign thresholds.

Missed cases tend to cluster in hospitals with fewer resources. A 2024 analysis of the National Inpatient Sample published in the American Journal of Obstetrics & Gynecology by Dr. Katheryne L. Downes and colleagues found that rural hospitals had a 40% higher odds of delayed sepsis diagnosis compared with urban academic centers. The reasons are multifactorial: limited access to rapid lab testing, fewer obstetricians on staff, and lower baseline suspicion for sepsis in postpartum women who may be discharged within 24 hours after an uncomplicated delivery.

Proponents of the protocol argue that the 80% detection rate represents a substantial improvement over the pre-protocol era, when recognition was ad hoc and mortality higher. Critics counter that the 20% miss rate is unacceptable for a condition where every hour of delay increases mortality by 7–8%. The disagreement is not merely academic: it shapes how hospitals train staff, allocate resources, and design electronic health record alerts.

Why Standardized Bundles Fail a Fifth of Patients

Most sepsis bundles rely on a combination of vital sign thresholds and laboratory triggers. The Systemic Inflammatory Response Syndrome (SIRS) criteria—fever or hypothermia, tachycardia, tachypnea, and abnormal white blood cell count—are commonly used. But in postpartum women, these signs can be misleading. A heart rate of 100–110 beats per minute may be dismissed as normal postpartum tachycardia; a temperature of 100.3°F might be attributed to dehydration or epidural-related fever.

A 2023 study from the University of Pittsburgh found that the SIRS criteria missed 27% of postpartum women who subsequently developed sepsis, compared with 14% of non-obstetric patients. The authors noted that obstetric patients often mount a blunted inflammatory response, so their vital signs may not cross the threshold until infection is advanced. Lactate, a key marker of tissue hypoperfusion, can also be falsely reassuring: some women with sepsis have normal lactate levels early in the course.

The consequence of missed detection is delayed antibiotics. In a retrospective cohort of 500 sepsis cases across four US hospitals—including Brigham and Women's Hospital, Northwestern Memorial Hospital, and two community hospitals in the Midwest—the median time to antibiotics was 2.1 hours for patients identified by the bundle but 4.8 hours for those who initially screened negative. Every hour of delay was associated with a 6% increase in in-hospital mortality. The study, presented at the 2025 Society for Maternal-Fetal Medicine (SMFM) meeting, reignited debate about whether the screening tools are sensitive enough for obstetric populations.

Some researchers argue that the problem is not the bundle itself but the way it is implemented. “Protocols are designed for speed, not nuance,” says Dr. Lisa Hollier, a maternal-fetal medicine specialist at Texas Children’s Hospital. “If a nurse is trained to follow the checklist and the checklist says no sepsis because the temperature is 100.2°F, she may feel empowered to stop—even if the patient looks sick. We need to build in a safety valve for clinical judgment.”

The Tension Between Speed and Accuracy in Diagnosis

Every screening tool involves a trade-off between sensitivity and specificity. The qSOFA score—which assigns one point each for altered mental status, respiratory rate ≥22, and systolic blood pressure ≤100 mm Hg—is fast and simple but has notoriously low sensitivity in obstetric patients. A 2024 meta-analysis of 14 studies found that qSOFA identified only 54% of pregnant or postpartum women with sepsis, compared with 72% for SIRS and 81% for the Maternal Early Warning Criteria (MEWC).

Proponents of the bundle approach argue that even an imperfect tool saves lives if applied consistently. “We cannot let the perfect be the enemy of the good,” said Dr. William Grobman, chair of obstetrics and gynecology at Northwestern University, at a 2025 SMFM panel. “Before bundles, we saw delays of 6–8 hours. Now we see delays of 1–2 hours. That is progress.” Critics, however, point out that the 20% miss rate means roughly 2,000 women per year in the US may experience delayed treatment, with an unknown proportion progressing to severe sepsis or death.

The debate played out publicly during a 2025 SMFM session titled “One Size Fits All? The Case Against Rigid Sepsis Bundles in Obstetrics.” Dr. Erin Clark, a maternal-fetal medicine specialist at the University of Utah, presented data from her institution showing that a modified bundle—which included a lower threshold for tachycardia (≥100 bpm) and mandatory lactate measurement for any febrile postpartum patient—reduced the miss rate from 22% to 11% over two years. “We didn’t abandon the bundle,” she said. “We tailored it to the population we serve.”

Opponents of such modifications worry that loosening criteria will lead to over-treatment: more blood cultures, more antibiotics, more unnecessary admissions. “False positives have costs too—patient anxiety, antibiotic resistance, resource drain,” said Dr. Cynthia Gyamfi-Bannerman, a maternal-fetal medicine specialist at the University of California, San Diego. “We need evidence that a more sensitive approach improves outcomes without causing harm.” Randomized trials comparing different screening strategies are underway, but results are not expected until 2028 at the earliest.

Rural Hospitals Face Higher Miss Rates

Rural hospitals, which deliver roughly 18% of US births, face structural barriers that compound the detection problem. Many lack rapid lab capability: a lactate result may take 90 minutes if the sample is sent to a reference lab. Few have 24/7 in-house obstetricians; care is often provided by family physicians or nurse-midwives, who may see only a handful of postpartum complications per year. A 2023 survey of 120 rural hospitals in 12 states found that only 38% had a formal sepsis protocol specific to obstetric patients.

The consequences are measurable. A 2025 analysis of the Healthcare Cost and Utilization Project database published in Obstetrics & Gynecology by Dr. Monica L. Swilley and colleagues found that rural hospitals had a 32% higher risk of sepsis-related maternal intensive care unit admission compared with urban hospitals, after adjusting for patient demographics and comorbidities. The same study found that rural hospitals were less likely to use lactate monitoring or early antibiotics within the first hour of suspicion.

Tele-ICU programs have shown promise in bridging the gap. A pilot program run by the Georgia Department of Public Health and Emory University connected rural labor and delivery units to a central sepsis monitoring team via video link and electronic health record integration. Over 18 months, the program reduced the time to antibiotics from a median of 3.5 hours to 1.9 hours, and the sepsis-related ICU admission rate fell by 25%. But scaling such programs requires funding, broadband infrastructure, and buy-in from hospital leadership—all of which remain uneven across states.

Medicare’s Hospital Inpatient Quality Reporting program includes a sepsis measure (SEP-1) that tracks compliance with the three-hour bundle for all patients, but it does not specifically penalize missed cases or differentiate between obstetric and non-obstetric populations. Some advocates argue that a maternal-specific quality metric, similar to the Joint Commission’s perinatal care measure, could incentivize hospitals to invest in better screening. “What gets measured gets done,” says Dr. Sarah McMullen, an obstetrician at the University of Kansas Medical Center who studies rural health disparities. “But right now, the measurement is too blunt to capture the problem.”

Emerging Alternatives: Risk-Stratified Pathways

Several states and hospital systems are piloting risk-stratified sepsis pathways that move beyond the one-size-fits-all bundle. California’s Maternal Quality Care Collaborative, for example, developed a modified algorithm that incorporates obstetric history, mode of delivery, and time since delivery. A postpartum woman with a temperature of 100.4°F and a history of cesarean section is flagged differently than one with a vaginal delivery and no risk factors. The algorithm also mandates a lactate level for any patient with fever and tachycardia, regardless of SIRS criteria.

A pilot study at five California hospitals, published in Obstetrics & Gynecology in 2025 by Dr. Jennifer Bailit and colleagues, reported that the modified algorithm reduced the sepsis miss rate from 19% to 12% compared with standard SIRS-based screening. The false-positive rate increased modestly, from 8% to 11%, but the authors argued that the trade-off was acceptable given the severity of missed sepsis. “We are not recommending antibiotics for every postpartum woman with a low-grade fever,” said lead author Dr. Jennifer Bailit. “But we are saying that a normal qSOFA score should not be the final word.”

Implementation challenges remain. The modified algorithm requires integration with the electronic health record, which many smaller hospitals lack the resources to customize. Training staff to use a more complex screening tool takes time and may be resisted by nurses and physicians accustomed to the simplicity of the old bundle. “A two-step algorithm is harder to remember at 3 am,” notes Dr. Thomas Ivester, a maternal-fetal medicine specialist at the University of North Carolina. “We need decision support that is both accurate and easy to use.”

Other groups are exploring the use of machine learning models that incorporate dozens of variables—heart rate variability, respiratory rate trends, lab values—to generate a real-time sepsis risk score. A 2024 study from the University of Michigan published in the Journal of the American Medical Informatics Association by Dr. Karandeep Singh and colleagues found that a neural network trained on 50,000 deliveries could predict sepsis 4 hours earlier than SIRS criteria, with a sensitivity of 85% and a specificity of 90%. But such models are not yet widely deployed, and concerns about algorithmic bias and generalizability remain.

The scalability of these approaches is uncertain. A risk-stratified pathway that works well in a large academic center with robust IT support may be impractical in a 25-bed rural hospital. “We need a menu of options,” says Bailit. “Not every hospital can run a machine learning model, but every hospital can train its staff to look for uterine tenderness and foul lochia—signs that are not in any bundle but are highly specific for postpartum infection.”

What Clinicians Should Watch for Beyond the Checklist

For clinicians at the bedside, the message from experts is consistent: do not let a negative protocol result override clinical concern. “If a patient looks sick, she probably is,” says Dr. Jillian Main, a nurse-midwife at the University of Washington who co-authored a 2025 practice advisory on maternal sepsis recognition. Signs that may not trigger a bundle but should raise suspicion include uterine tenderness, foul-smelling lochia, prolonged capillary refill, and a sense of unease about the patient’s appearance.

Low-threshold blood culture collection is another key recommendation. In a 2023 study of 200 postpartum women with fever, published in Obstetrics & Gynecology by Dr. Sarah J. Kilpatrick and colleagues, blood cultures were positive in 18% of those who did not meet SIRS criteria. Delaying cultures until criteria are met can waste critical hours. “Every febrile postpartum patient deserves a blood culture, regardless of vital signs,” says Main. “It is a low-cost, low-risk intervention that can change management.”

Clinicians should also consider sepsis in any unexpected deterioration, even in women who seem to be recovering well. Postpartum sepsis can manifest as early as 6 hours after delivery or as late as 6 weeks, with the peak incidence in the first 24–48 hours. A patient who is discharged and returns with vague symptoms—fatigue, abdominal pain, mild confusion—should be evaluated with a high index of suspicion. The Joint Commission is currently reviewing its maternal safety standards and may update sepsis screening requirements as early as 2027.

Ultimately, the goal is not to abandon protocols but to use them as a starting point. The 20% miss rate is a reminder that no checklist can replace a thoughtful clinician. As Dr. Hollier puts it, “A protocol is a floor, not a ceiling. It gets you to a minimum standard. But the best care happens when you look at the patient, not just the numbers.”

This article is for informational purposes only and does not constitute personalized medical advice. Pregnant and postpartum individuals with concerns about infection should contact their healthcare provider promptly.