Sepsis initiative takes on elusive foe

New systems yield success; efforts continue to speed response time
July 20th, 2016

The dangerous condition known as sepsis is an elusive opponent that can escape detection until it’s too late to save its victims. Providers at UCHealth continue to hone systems to pinpoint sepsis as early as possible and also head off the progression of other serious conditions.

It’s been nearly two years since UCHealth launched a system-wide offensive against sepsis, a systemic response to infection. There are more than a million cases of severe sepsis – defined as those with sepsis and indications of organ failure – in the United States each year, and between one-quarter and one-half of them die, according to the National Center for Health Statistics.

The basic goal of the initiative was and is to save lives; the Centers for Medicare and Medicaid Services also provided a spur to action by making early recognition of sepsis and implementation of evidence-based processes (“bundles”) to reduce mortality a core measure of quality care (see box).

The UCHealth work led to an early warning system (EWS) that allows bedside providers in medical/surgical units to assess the risk of sepsis by identifying infection and altered mental status and enter vital signs, such as pulse and respiratory rates, and body temperature, into the Epic electronic health record. If the information indicates the patient is at risk for sepsis, Epic triggers a “best practice advisory” (BPA) and nurses can order a lactate and complete blood count (CBC) test to confirm or rule it out. If the patient is septic, providers are to order broad-spectrum antibiotics, fluids and, if necessary, vasopressors to raise blood pressure.

The EWS went into effect at UCHealth hospitals in northern Colorado last fall and in the Medicine Specialties Unit at University of Colorado Hospital in December. The Surgical Specialties Unit at UCH followed earlier this year. The goal is to implement it in all medical/surgical units by the end of 2016, said Nicole Huntley, RN, CNS, sepsis coordinator for UCH.

By the numbers

As of the first week of June, nurses in the Medicine Specialties and Surgical Specialties units at UCH had used the EWS to screen a total of more than 2,000 patients, Huntley said. They identified 120 high-risk patients, 55 of whom were diagnosed as septic.

The system has also led to some notable saves of individuals with problems other than sepsis, she added, including a retroperitoneal bleed, two massive pulmonary embolisms and an abdominal wound rupture after an elective surgery.

“The tool has given nurses autonomy and a greater voice in the patient’s care,” Huntley said. With the backing of specific criteria that point to sepsis, she said, nurses can speak objectively about their concerns rather than simply saying they have a hunch. If it turns out the patient isn’t septic, the effort isn’t wasted, she added. Providers can turn their attention earlier to identifying what the problem really is.

Kristen Selke, RN, a direct-care nurse on the Surgical Specialties Unit at UCH, noted in a clinical narrative on infection that the EWS has been helpful in “cluing the nurse in to minor changes in the patient’s condition.” She acknowledged that screening a patient’s vital signs every four hours adds to nurses’ already heavy charting time. But the benefits of having an algorithm that flags “a higher risk patient that I might have skimmed right over,” have far outweighed the burdens, Selke wrote.

The approach to identifying potential sepsis cases in UCHealth emergency departments is slightly different, but the aim is the same. The need for a “high suspicion” of sepsis is crucial, as some 70 percent of sepsis patients enter the hospital through emergency room doors, said Robin Scott, RN, CNS, clinical nurse specialist for the ED at UCH. Providers in the ED ask patients screening questions to determine a possible source of infection and altered mental status and take vital signs to assess for sepsis. If the risk is high, Epic fires the BPA for a CBC and lactate test. Providers use the lactate results to order fluids and cultures, if necessary, Scott said.

Signs of success

The series of measures have produced some success, according to an April 26 presentation of the UCHealth Sepsis Executive Committee. For example, in the baseline year of July 2013 to June 2014, the mortality index for all UCHealth sepsis patients was 1.16 (meaning the actual number of deaths from sepsis exceeded the expected number. The goal is to achieve a mortality index less than 1.0). From April 2015 to March 2016, the mortality index for all UCHealth sepsis patients fell to 0.96. That decrease translated to a potential 137 lives saved.

The index also fell for the same period for patients with severe sepsis from 1.24 to 1.07. The decrease meant a potential 85 lives saved.

Meanwhile, the mortality rate – the percentage of patients with a condition who died from it – for sepsis patients at UCHealth stood at 9.2 percent in March. Huntley said that the initiative aims to lower the mortality rate by at least 1 percentage point every year.

Need for speed

The key to driving down sepsis mortality is rapid response, said Christopher McStay, MD, chief of clinical operations for the CU School of Medicine’s Department of Emergency Medicine.

“The sooner we recognize sepsis and treat it, the better the outcomes,” McStay said. “If we don’t recognize it in a timely fashion, patients will go to the ICU or they will die.”

The Executive Committee report supports McStay’s point. Most notably, the mortality index and mortality rate each decreased significantly when sepsis patients throughout UCHealth received antibiotics within one hour of diagnosis. The mortality index also fell when hospital teams adhered to the three- and six-hour bundles of care for sepsis patients.

Hiding in plain sight

Given the well-established threat that sepsis poses, why has the condition proven so difficult to fight? One problem is that recognizing sepsis isn’t as clear-cut as, say, stroke or heart attack.

“There is still a lack of understanding of the definitions of sepsis,” McStay said. The challenges of diagnosis gained prominence with the case of Rory Staunton, a 12-year-old New Yorker who died in 2011 after a bacterial infection that entered his bloodstream through a seemingly minor cut on his arm escaped attention for a time, in part because his symptoms mimicked other possible conditions.

The Staunton case, chronicled in the New York Times, led New York hospitals to institute more stringent measures for assessing and treating sepsis. That effort led to the CMS core measures, which McStay says are driving hospitals to develop rapid response operations like the sepsis alert teams at UCHealth Northern Colorado and “mini rapid response” teams in UCH’s ED.

But a rapid response still relies on the clinical judgment of providers, and sepsis can be “a hard condition to grab by the reins,” Huntley said. McStay noted, for example, that a young patient admitted to the ED and diagnosed with strep pharyngitis might meet the criteria for a sepsis diagnosis, but providers wouldn’t – and shouldn’t – order the major medical workup administered to a sepsis patient. He’d be given antibiotics and discharged.

The tools for identifying sepsis are also fluid. A “Sepsis-3” group introduced “qSOFA” (quick Sequential Organ Failure Assessment) last February as a new method for identifying infected patients with a poor prognosis. The qSOFA screen would replace the SIRS (Systemic Inflammatory Response Syndrome) criteria currently in use at UCHealth and elsewhere.

Debates over screening tools and scoring systems aside, CMS will continue to expect hospitals to meet its core measures for treating sepsis, McStay said.

“The CMS metrics haven’t changed,” he said.

Scott believes that providers’ acceptance of the bundles will increase as data supporting the effectiveness of standardized assessment and treatment piles up. Ultimately, she said, the goal is to save lives and prevent sick patients from getting sicker.

“We’re trying to flag the signs that tell us there is something seriously wrong,” Scott said. “Then we can do the detective work to find out what is really going on with a patient. It’s about catching the very acutely ill.”


Getting to the core

The “bundles” CMS considers essential to reducing sepsis mortality:

The 3-Hour Bundle

  1. Measure lactate level
  2. Obtain blood cultures prior to administration of antibiotics
  3. Administer broad spectrum antibiotics

 The 6-Hour Bundle

  1. Administer 30ml/kg fluids for hypotension or lactate ≥ 4 mmol/L
  2. Apply vasopressors for persistent hypotension after fluid challenge
  3. Re-measure lactate if initial lactate elevated

About the author

Tyler Smith has been a health care writer, with a focus on hospitals, since 1996. He served as a writer and editor for the Marketing and Communications team at University of Colorado Hospital and UCHealth from 2007 to 2017. More recently, he has reported for and contributed stories to the University of Colorado School of Medicine, the Colorado School of Public Health and the Colorado Bioscience Association.