Published in the December 2016 issue of Today’s Hospitalist
FOR HOSPITALISTS, caring for patients with severe sepsis on the floor has always been a challenge. But now, the Centers for Medicare and Medicaid Services (CMS) has upped that ante.
Last year, the agency adopted a bundle of sepsis interventions as a core measure in its inpatient quality reporting program. As a result, hospitalists say, the challenges of treating sepsis have become much tougher.
“It’s definitely the hardest core measure I’ve ever tried to meet.”
Why? According to Trina Dorrah, MD, a hospitalist and regional quality director at Baylor Scott & White Medical Center-Round Rock in central Texas, a major challenge with compliance is time.
“It’s definitely the hardest core measure I’ve ever tried to meet because it has a lot of components, and you have only three or six hours to complete them,” Dr. Dorrah says. “As physicians, we often don’t realize that ‘Time Zero’ has even been met because we’re so busy trying to take care of the patient, not watch the clock.”
Around the country, hospitalists report similar problems. Some report trouble with specific steps in the bundle process, like ordering and administering antibiotics within three hours.
Others are trying to figure out how to diagnose sepsis faster to know when to start the clock. And doctors say a big part of the compliance burden has to do with documentation, not patient care.
All or nothing
The CMS began requiring hospitals to submit sepsis-related data in October 2015. Eventually, those data may be publicly reported and affect hospital reimbursement.
The CMS bundle, which reflects Surviving Sepsis Campaign recommendations, requires hospitals to complete several interventions within three hours of a patient presenting with sepsis, then several more within six hours. (See “What’s in the CMS sepsis bundle?”) To get credit for meeting the measure, hospitals must achieve 100% compliance with all bundle elements, an “all-or-nothing” requirement that many hospitals struggle with.
“We are now being graded by the CMS on how well we do something that we haven’t been able to do very well” before now, says Andrew J. Odden, MD, a hospitalist at Barnes-Jewish Hospital in St. Louis, who has worked on many sepsis management initiatives. “The reason we haven’t is because it’s really difficult to do.”
According to Dr. Odden, there are many reasons behind poor compliance. “Some of it is fear, some is reluctance, some is knowledge and some is structural. Ignoring any piece of that oversimplifies the challenges of delivering sepsis care.” That’s particularly the case on medical and surgical floors where, unlike in the ICU or emergency department, providers don’t have low patient-to-provider ratios that make it easier to meet time-sensitive requirements.
Even hospitals that have put improving sepsis care front and center may be “functioning at only the 40%-50% level” in terms of meeting the CMS measure, says R. Phillip Dellinger, MD, a pulmonology/critical care physician who is chief and chair of medicine at Cooper Medical School of Rowan University in Camden, N.J. A national expert on sepsis, Dr. Dellinger was lead author of the Surviving Sepsis Campaign’s 2012 guidelines.
Unfortunately, he points out, hospitals that have significantly improved their guideline-congruent care may still look bad because of the all-or-nothing scoring.
The antibiotics component of the bundle has turned out to be a “sticking point.”
“You get no more credit for achieving six out of seven measures than for two out of seven,” Dr. Dellinger says. “You don’t get credit unless you are 100% on the measures in an evaluable patient, so it is often frustrating.”
Delivery and documentation
For Aroop Pal, MD, a hospitalist at the University of Kansas Medical Center and a member of a Surviving Sepsis Campaign collaborative, a major barrier to compliance is an elemental one: how hard it is to determine when patients have severe sepsis. That’s particularly challenging with hospitalized patients with chronic diseases—when it may not be clear whether acute end-organ damage may be present.
“Because we don’t get information all at once,” says Dr. Pal, “it’s incumbent on the physician to put the whole picture together” and to begin or continue administering the bundle correctly. “Getting all that information from different places at different times to the right provider at the right time does stretch the abilities of clinical decision-support.”
According to Muhammad Mullick, MD, hospitalist program director for the Apogee Physicians group at RWJ Barnabas Health’s Newark Beth Israel Medical Center in Newark, N.J., having to respond to rapid response-type Code SMART (Sepsis Management Alert Response Team) calls, deliver all the bundle elements and then document them properly in the required amount of time takes continuous re-education, reminders and careful attention to detail.
As far as Dr. Mullick is concerned, “Without hospitalists, compliance rates would be worse.” He can’t help but notice that his Code SMART calls— “maybe two or three a week”—are for patients of primary care physicians who aren’t in the hospital and can’t come close to providing rapid-fire care.
But even the most efficient hospitalists face compliance barriers. At the University of Kansas Medical Center, for instance, 60% of sepsis patients come through the ED, while 15% are transferred in from other places and 25% occur in the hospital, either post-operatively or on the floor. That breakdown is part of what makes compliance so difficult.
“Unlike in an emergency department or an ICU, where you see a lot of severe sepsis, you don’t see it frequently on the med-surg units,” Cooper’s Dr. Dellinger points out. “On the floors, you have to train a lot of people and keep them poised for something that doesn’t happen frequently.”
In any given week that he works the hospitalist service, Dr. Odden in St. Louis may see no sepsis or severe sepsis cases, or he may see five or six.
“You are seeing it potentially regularly, but not every day,” he notes. “That makes recognition a challenge.”
Hospitals investing resources in improving sepsis care have focused on several areas. Those include educating providers, nurses as much as doctors; reworking EHRs to include bundle-compliant order sets, progress note templates, countdown clocks, flow sheets, best practice alerts and early warning systems; launching sepsis rapid response teams; creating nurse-driven regular screening programs; and giving nurses protocols that allow them to order blood cultures or lactates, even before doctors intervene.
It’s too early to tell if any of these interventions work better than any other, in part because success depends on individual hospital culture, resources and personnel. There is also a problem with many physicians and nurses not keeping up-to-date with the best practices developed by the Surviving Sepsis Campaign. That educational piece has been further muddled by the publication this year of controversial new definitions of sepsis and septic shock. (See “The big definition debate.”)
Some clinicians also either disagree with or are leery of some bundle elements, such as the call for large fluid boluses for nearly all patients with severe sepsis, even those with heart failure or end-stage renal disease. And even when people stay current, they may not be able to meet the measure, given the complicated institutions they work in.
“I can order the right things at the right time,” Dr. Odden explains. “But an order for the right broad-spectrum antibiotics doesn’t count toward bundle compliance unless the antibiotics are administered within the required amount of time.”
Jennifer A. LaRosa, MD, a pulmonary/critical care physician who directs the ICU at Newark Beth Israel Medical Center, spearheaded that center’s Code SMART. As Dr. LaRosa puts it, meeting the measure “is not just a doctor or a nurse thing. It really takes a lot of different disciplines to get this right: lab, transport, runners, techs, nurses—a lot of people.”
Steven Q. Simpson, MD, a pulmonology/critical care physician who is the head of the medical ICU at the University of Kansas Medical Center and a leading sepsis expert, says that the antibiotics component of the bundle has turned out to be a “sticking point” at his hospital.
“Our average antibiotics time was four hours the last time we looked,” Dr. Simpson says. That’s due to delays in recognizing sepsis, in ordering tests and drugs, in pharmacy procedures, in transporting the medications from the pharmacy to the floor, in nurses realizing the meds are available, and, finally, in nurses administering the dose to the patient.
Hospitals hit similar snags with other bundle items, including rechecking lactates and obtaining blood cultures.
As Dr. Dorrah notes, fluid management is one component that is difficult for hospitals to meet. In her health system, a system-wide multidisciplinary “SepsisPalooza” brought participants together for a day and a half last March. The problem, that initiative revealed, wasn’t that doctors weren’t ordering fluids, but that the ordered fluids did not always match Medicare’s specifications in terms of amount, speed of administration or time frame.
LAC+USC Medical Center in Los Angeles studied the barriers to complying with sepsis bundle elements. While researchers hypothesized that there would be less compliance at night and on weekends, they found the exact opposite: Overall compliance suffered during the day vs. during nights and weekends (18.9% vs. 38.0%).
As for why, the study noted that clinicians have fewer “distractions” at night and on weekends, allowing them to “focus on meeting the parameters necessary.” (A study abstract was presented at SHM’s 2016 annual meeting.) Interestingly, the study also found that an increased rate of bundle compliance didn’t boost patients’ rate of survival to discharge.
Problems to tackle
According to Andrew Young, DO, the service chief of medicine at LAC+USC Medical Center and a co-author of that study, more recent preliminary data from his hospital show a 23% compliance rate with the CMS sepsis bundle. “The most challenging bundle elements to accomplish seem to be the repeat lactate, documenting the fluid infusion rate and having a stop time for the infusion.”
As for problems with documentation, “we are used to writing ‘2 liters IV bolus,’ but our abstractors are saying that doesn’t meet the core measure,” Dr. Young explains. “Either the nurse needs to document when the bolus is done or we need to put in the infusion rate. We need to figure out how to tackle that.”
For Dr. LaRosa in Newark, consistent bundle documentation can be frustrating. That’s because the required focused exam within six hours of the Code SMART call is highly specific as to what elements must be contained in the note—and, of course, in the patient evaluation—to meet the standard.
She reviews every Code SMART called in her hospital. When doctors miss any bundle element because they did not use the order set, “I discuss it with them,” says Dr. LaRosa, “and reeducate them.”
At Baylor Scott & White in Round Rock, this year’s SepsisPalooza led to the formation of various workgroups to address specific bundle items.
Dr. Dorrah is co-chair of the sepsis floor workgroup, which was charged with crafting one process that all seven hospitals in her system can follow to treat patients who develop sepsis on the floor. Rolled out this fall, the initiative includes a combination of nursing and provider education, nursing protocols, alerts in the Epic EHR and standard template documentation.
She notes that the development of pop-up Epic alerts—letting providers know when patients meet criteria for sepsis or severe sepsis—has been particularly helpful; clinicians can then open the sepsis order set directly from the pop-up alert. The order sets, says Dr. Dorrah, will go a long way to help meet sepsis bundle requirements.
Venkataraman Palabindala, MD, lead hospitalist at the University of Mississippi Medical Center in Jackson, Miss., says his first priority right now is to hire more hospitalists. When he’s fully staffed, he wants to consider implementing some of the innovations put in place at his former hospital in Dothan, Ala., that improved sepsis care and bundle compliance.
At his former facility, one hospitalist—designated as the hospital’s sepsis champion—received fewer patients per shift. That allowed the champion, says Dr. Palabindala, to “watch any diagnosis of sepsis” as it unfolded, help attendings hit all the right marks and document the entire process correctly, and “give real-time feedback.”
The champion also, he adds, helped colleagues earn their potential bonus. That’s because the group decided to tie a portion of the hospitalists’ individual incentive to their success in meeting the measure.
Beyond systems and process barriers, Dr. Simpson in Kansas sees hospitals commonly “go off the rails” when “nursing techs have no idea what the vital signs mean that they are taking.” Too often, those techs don’t share vitals with the nurse for two hours, and the nurse may not look at them right away. “When the nurse does—but only if he or she has been trained to think through is there an infection, is there SIRS, is there evidence of organ dysfunction—and finally calls the doctor, the doctors very frequently don’t respond or react fast enough.”
Part of Dr. Simpson’s solution is outreach throughout Kansas, educating clinicians in hospitals state-wide about sepsis to help them get a jump on the diagnosis.
“I’ve been training hospitals over and over again, multiple times,” he says. “It takes multiple exposures to training because there is no simple diagnostic test. Instead, it takes integrating several clinical findings and understanding what they mean.”
His own medical center has eliminated that string of delays on the floor. “The mantra is that we screen every patient every shift every day,” Dr. Simpson says, a practice that Dr. Dorrah’s hospital has adopted as well. Nurses answer three questions: Does the patient have a new or worsening infection, does the patient meet at least two SIRS criteria, and is there evidence of organ dysfunction?
“There are a lot of advantages to nurse-based screening because it’s a moment when someone is looking at all the information, even if it is very quickly,” says Dr. Odden in St. Louis. The aim of screening is in part to “prompt a conversation with the physician. The secret ingredient is engaging both the nurses and physicians to try to put this together.”
Deborah Gesensway is a freelance writer who covers U.S. health care from Toronto.
COMPLICATING MATTERS for hospitalists trying to deliver the new sepsis treatment bundle is the fact that groups of sepsis experts are publicly fighting over how to define sepsis.
In its Feb. 23, 2016, issue, just four months after hospitals began reporting compliance with the sepsis bundle to the CMS, the Journal of the American Medical Association published an international consensus statement redefining sepsis and septic shock.
The new definitions, known as Sepsis-3, concluded that the term “severe sepsis” was redundant. It also stated that it is better to diagnose the infection-related condition by calculating the clinical Sequential [sepsis-related] Organ Failure Assessment (SOFA) or a modified Quick SOFA score, rather than assessing systemic inflammatory response syndrome (SIRS) criteria, which physicians have used since the 1990s.
The authors of the new definition argue that relying on SIRS is “misleading” because sepsis-severe sepsis-septic shock is not a continuum. They also characterized SIRS criteria as “inadequate” in terms of “specificity and sensitivity.” The new definition, they argue, will “facilitate earlier recognition and more timely management.”
But an editorial published in the May 2016 issue of Chest contended that “widespread application of this new definition could cost patient lives.”
“Given that use of the current definitions results in saving lives, it seems unwise to change course in midstream by shifting the definition,” wrote editorialist Steven Q. Simpson, MD, a University of Kansas critical care physician. “A change in definition and diagnostic criteria could set back decades of work persuading providers at all levels to recognize sepsis early and to intervene aggressively.”
In the meantime, Dr. Simpson told Today’s Hospitalist that physicians don’t need to quibble about how to define the condition—particularly if that debate “is an excuse for inertia.” Instead, hospitalists should devote themselves to meeting the CMS bundle to prove they are treating the condition aggressively.
“The sooner you name it and treat it like the killer it is,” says Dr. Simpson, “the more likely you are to successfully intervene.”
In addition, points out R. Phillip Dellinger, MD, senior critical care attending at Cooper University Health Care in Camden, N.J., Sepsis-3 is merely “proposed.” Although it’s been endorsed by many leading organizations, it “absolutely cannot be operationalized” for at least several years, he notes, while neither the CMS nor ICD-10 has adopted Sepsis-3.
“I think the confusion is that there are pockets of people who read the proposed new definitions and then try to figure out how to integrate them into the CMS measures,” says Dr. Dellinger, past president of the Society of Critical Care Medicine and lead author of the current sepsis guidelines. “You can’t.”
Trina Dorrah, MD, hospitalist and regional quality director at Baylor Scott & White Medical Center–Round Rock, Texas, says that colleagues often mention that the consensus definitions conflict with CMS’s sepsis definitions. Her response is to point out that the JAMA paper focuses on definitions and identification, not on treatment.
“What I say is that we still have to provide good medical treatment for patients with sepsis,” Dr. Dorrah points out. “Right now, we are going to treat them while being cognizant of Medicare guidelines because that is what we are being measured on.”
IN OCTOBER 2015, hospitals began reporting their compliance with a bundle of sepsis interventions put forth by the Centers for Medicare and Medicaid Services (CMS). That bundle requires hospitals to complete the following four elements within three hours of a patient presenting with sepsis, the not-always-obvious “Time Zero”:
- measure a lactate;
- obtain blood cultures prior to administering antibiotics;
- administer broad-spectrum antibiotics; and
- administer 30 ml/kg crystalloid fluid for hypotension or for a lactate level of 4 mmol/L or more.
Within six hours, providers need to re-measure an initially elevated lactate and apply vasopressors if warranted. And if the patient has persistent hypotension or the initial lactate was 4 or more, doctors must document—in a particular, spelled-out manner—that they reassessed volume status and tissue perfusion through a focused exam that includes vitals, a cardiopulmonary exam, an evaluation of capillary refill and peripheral pulse, and a skin exam.
And as part of that reassessment, they need to include two of the following: central venous pressure measurement, central venous oxygen saturation, bedside cardiovascular ultrasound, and a passive leg raise or fluid challenge exam.