Learn how to better comply with the bundles, including drilling down into how to address the most common reasons for noncompliance.
The key: understand common care gaps and use change management to address them
There is certainly honest disagreement among clinicians about the value of strict compliance with the CMS 3- and 6-hour sepsis bundles. Until that disagreement is settled, however, compliance is a necessity. There are two compelling reasons. First, bundles are the closest thing to best practice that we have; compliance saves lives. Second, there are financial and reputational penalties for noncompliance.
The purpose of this post, therefore, is to give sepsis coordinators and health system quality leaders guidance on how to better comply with the bundles, including drilling down into how to address the most common reasons for noncompliance.
First, the big picture
Once you’ve identified a patient with sepsis – no easy task, but something we’ve addressed in other settings – the CMS bundles contain a number of elements. Briefly, the 3-hour bundle requires a lactate, blood cultures before antibiotics, antibiotics, and if a patient is unstable, large volume IV fluid administration (30 mL/kg). The 6-hour bundle adds a second lactate, physical exam, and patient recheck.
A few factors affect bundle compliance across the board. The first and most important is early recognition. If you can’t identify patients with sepsis early, in most cases you have little chance of full compliance and, worse, the patient is already in what is too often irreversible decline. A proven automated sepsis surveillance solution plays an invaluable role here.
A second common reason for failure is not using a sepsis-specific order set. This is actually more common than one might believe, typically because providers tend to use a disease-specific order set other than the sepsis order set. For example, pneumonia is the underlying cause for many hospitalized patients with sepsis. Because few electronic medical records facilitate the simultaneous use of two order sets, providers often turn to the pneumonia order set instead of the sepsis order set, and the pneumonia order set may be missing key sepsis orders. One solution for this common problem is to have providers utilize separate order sets back-to-back, first the pneumonia order set, then the sepsis order set. This strategy will only work if the sepsis order set is not a complete admission order set, but instead just contains additional key sepsis orders, e.g. lactate testing, blood cultures, fluids, etc. An order set committee will likely need to be involved to facilitate this effort. Change management efforts will need to be undertaken to teach physicians to use 2 order sets, because this is currently an uncommon practice. Education sessions, emails, notices – and perhaps most importantly, sharing reports of order set utilization and outcomes amongst physician staff – are key efforts to ensure improvement in order set use.
Policies also need to be established to ensure timely administration of sepsis care. All members of the care team, including physicians, nurses, pharmacists, rapid response personnel and even phlebotomists need to understand the urgency of time and their roles in the care of sepsis. Staff need to clearly understand when they should include others for help and who to call; this detail needs to be included in protocols, and should specify different workflows for initial evaluation of sepsis, treatment of sepsis with organ dysfunction and specific plans for patients that develop shock. As with physician order set improvement, this effort requires education, mailings, meetings, case review and frequent public sharing of key performance metrics, with a focus on process improvement. This last item requires analytics and reporting efforts.
In short, improvement in CMS bundle compliance requires implementing a strong automated sepsis surveillance solution to capture cases early, the design of readily accessible sepsis-specific order sets, detailed policies and procedures tailored to each clinical setting, and analysis of performance metrics that are shared with personnel.
Now let’s look at the four most common bundle compliance failures and explore what else, beyond base level sepsis management, hospitals can do to encourage full compliance.
Ordering the first and second lactates
Failure to administer the first lactate often has to do with both of the factors noted above: delays in sepsis recognition and use of an order set that doesn’t contain a lactate. Another problem, however, can be physicians asserting that they don’t need the lactate, because they don’t need lactate test results to guide their treatment, and research has not proven that lactates change outcomes. The health system response should be to acknowledge that both of those things can be true, but there is a greater good argument for ordering the lactate: doing so standardizes sepsis care – particularly important for less experienced providers – while also facilitating early retrospective analysis and apples-to-apples comparisons of sepsis care. In other words, don’t argue with this physician. Let him or her be right, but appeal to the sense of greater good, that consistency in testing all patients with the same methodology leads to consistency of care, improving overall care delivered by all clinical staff members. And it allows for more complete data collection. Lactate test results provide an objective measure that can be assessed across a population, allowing for better analytics, and, in turn, leading to additional improvements.
Failing to order the second lactate is even more common, often because of the reasons given above; thus the same arguments from the health system apply. In addition, many health systems omit the second lactate from their sepsis order set, and the provider forgets to order a second text. Regardless of the reason for the failure, the fix is to make the second lactate a “reflex” lactate; that is, the hospital policy can read: “Order a first lactate. If the result comes back greater than 2, then an automatic order for the second lactate will go into the system.” Some hospitals and labs might balk at having the EMR automatically generate the lab order, but given the costs for failing to comply, it is a worthwhile policy that, again, standardizes care.
Yet another problem with both the first and second lactates has to do with drawing them in a timely manner. At some hospitals, nurses draw the lactate at the point of care; other hospitals require that phlebotomy do the draw. Both present compliance challenges. Nurses have many distractions from different duties, and so it’s not surprising that the lactate draw occurs late or not at all. Instituting a reminder at the point of care that pings the nurse in a timely way is one solution that hospitals can easily build into their EMRs.
If phlebotomists are the reason lactate draws are late, simply making the lactate a STAT order may solve the problem. But it may require including the phlebotomy team in education, policies, analytics and outcomes assessment efforts so they understand appropriate priorities. Sometimes, though, hospitals struggle to fully staff their phlebotomy team, particularly late nights and weekends, which can also be the cause of delays. If a detailed workflow analysis reveals that the delay is, in fact, attributable to phlebotomy, hospitals may to need include nursing staff for blood draws.
Another way to speed and ease the lactates is to invest in point-of-care lactate machines in the few settings where sepsis is most common, thus averting the lengthy process of getting blood and test results to and from the lab. A key to understanding the problem with delays in lactate testing and finding an ideal solution requires classic change management analysis and improvement processes.
Administering antibiotics prior to blood cultures
CMS regulations say if there will be a significant delay before a clinician can get blood cultures drawn, he or she should administer life-saving antibiotics prior to a blood culture. Previous iterations of the Surviving Sepsis Campaign Guidelines defined a time interval for “significant delay”, but that was redacted in current guidelines because of a lack of data, so today, a “significant delay” is at the discretion of the provider. In order to be compliant with this metric, if a provider decides it’s not appropriate to wait for blood cultures before administering the antibiotics, then he or she has to document that it was a clinically driven decision to administer antibiotics because obtaining cultures would create a significant delay. Getting physicians to document appropriately will require change management! But doing so will aid abstractors in data submission and will avoid a penalty.
Delays in antibiotic administration
Sometimes providers write a timely order for antibiotics, but the process of getting the antibiotics back to the bedside is in itself time-consuming. The order must get down to the pharmacy, the pharmacist must prepare the medication and send it back, the nurse must recognize that the medication is available and administer it. In a busy hospital, where all clinicians are juggling competing priorities, it is not hard for such a process to exceed three hours.
If a hospital conducts a process flow analysis and finds that it has properly prioritized sepsis antibiotics and has staffed properly, yet is still noncompliant with this metric, another potential step is to prepare commonly used antibiotics for sepsis in advance and stocking them in the clinical locations where sepsis is most common.
Decisions on fluid administration
Fluid administration is often the most challenging and controversial of the bundle recommendations. Providers understands the need for IV fluid in cases of sepsis, but there’s disagreement about how much fluid is appropriate, depending on individual patient factors. Ultimately, CMS decided that until we have enough data, the standard is 30 mL/kg over three hours, within three hours. Nevertheless, many providers are reluctant to provide that much fluid because there are instances when it is appropriate to deliver less fluid.
The problem, however, is that some providers, unwilling to risk volume overload, incorrectly believe certain at-risk patients should not get IV fluid resuscitation and, therefore, risk noncompliance. In particular, providers worry about people who have heart failure, kidney failure and liver failure. These patients do require more careful monitoring, including frequent physical examination to detect volume overload, but in most cases of sepsis with hypotension, these patients will still need IV fluid, even if it is a reduced amount.
For hospitals to convince providers of the need for fluid in these cases, your physician champion should be Involved. The should review key studies with physician staff:
Liu, Vincent X., et al. “Multicenter implementation of a treatment bundle for patients with sepsis and intermediate lactate values.” American journal of respiratory and critical care medicine 193.11 (2016): 1264-1270.
In this study, patients with heart failure and chronic renal insufficiency had better outcomes with IV fluid administration in accordance with CMS bundle compliance. (Note that end-stage renal disease on hemodialysis was not studied.)
In another study, these patients at risk of volume overload experienced no extra harm from crystalloid fluid administration:
Leisman, Daniel E., et al. “Patterns and outcomes associated with timeliness of initial crystalloid resuscitation in a prospective sepsis and septic shock cohort.” Critical care medicine 45.10 (2017): 1596-1606.
Sepsis coordinators should have that data on hand, engage a physician champion who can help educate other providers about this evidence-based guideline, and build a process that can help providers make more informed decisions about fluid levels at the point of care.
Despite some controversy around certain elements of the CMS bundles, standardizing compliance with the bundles can help save lives. As we’ve documented here, there are solutions to many of the most frequent compliance challenges. Now is the time to implement those solutions.