At Issue: The Sepsis 72-hour Admission

The Centers for Medicare & Medicaid Services (CMS) and third-party auditors are going after sepsis admissions for which patients are discharged within 72 hours. It may seem preposterous, from the hospital’s perspective, that they are being allegedly unfairly targeted for payment denials for sepsis admissions. But is this actually the case?

Before Sepsis 1.0, there were around 200,000 cases recorded annually in the U.S. After, it jumped more than fourfold, to around 900,000! Were we so inefficient in calling out sepsis before? Are we today seeing a lot more patients who are septic? Are we making our patients sicker?

It’s not that there has been a seismic change in sepsis. Patients alternately survived and died of sepsis even before the medical world knew what it was. What has changed is that through the years, we have arrived at a better understanding of sepsis and how to manage it. This then led to better outcomes.

The crux of the matter is that providers have not been as accurate in calling it when it really is sepsis (and conversely, calling it sepsis when it is not).

Yes, there have been changes in its definition and parameters, but this is because we have a better understanding of the pathophysiology of sepsis. It does not mean that sepsis changed in its iterations. What has been consistently changing is when and how providers use the term “sepsis.” And at times, it is not even dependent on how sick the patient is!

For the longest time, many physicians documented sepsis after blood cultures turned up positive, with a feasible pathogenic organism. Not that all positive blood cultures are pathogenic; indeed, there are colonizers and contaminants that grow in many blood cultures. Alternately, negative blood cultures do not really rule out sepsis. Only about 40-60 percent of patients who are truly septic may grow positive blood cultures.

In my travels across the country helping establish clinical documentation improvement (CDI) programs, I’ve seen all the discrepancies in the documentation of sepsis by providers. The infectious disease specialists themselves are beset with different spins on the subject. Some do not even want to use the word, and instead insist on “bacteremia.”

Somehow, “sepsis” was too plagued with inconsistencies, so bacteremia became the more adequate diagnosis to some folks. Bacteremia, septicemia, and sepsis have become synonymous to many providers (Please refer to previous articles I have written on the matter for the differentiation).

The other big issue has to do with time frames. Providers in general are used to looking at documentation in reference to isolated daily episodes. That’s how providers bill their E/M services (professional fees). On the day of admission, the patient presents with symptoms/manifestations, and laboratory and ancillary studies are performed, which may produce results consistent with “sepsis.” In such cases, the patient meets the criteria for sepsis and is placed on sepsis protocol.

The follow-up documentation in the subsequent progress notes will probably be consistent with what was documented in the initial impression. It is going to be more convenient to leave it as is. Changing it later may give the impression of inefficient diagnostic and medical management skills. It may also seem as though they embellished and overbilled at the outset. When asked if sepsis was the definitive diagnosis, the provider would respond that it met criteria and patient was placed on sepsis protocol – and he or she is standing by it. But, could there be other dynamics that could explain the patient’s presentation – after study?

One important aspect of medicine is that clinical parameters are not infallible. Other factors can account for certain symptoms a patient manifests. Unique patient circumstances (including age, activity, and comorbidities) can change what one expects for certain diseases. In the extremes of ages (i.e., the very young and the very old), sepsis will not present with the classic picture of infirmity. In the elderly, a change in mental status may be the only clue. On the other hand, for many elderly patients at nursing homes who refuse to eat and are not getting enough fluids, hypovolemia alone can result in the same mental status changes. Other considerations can also come into play: not only hypovolemia, but medications and other comorbidities can account for mental as well as other changes that mimic sepsis (e.g., other organ dysfunctions such as hypotension, pre-renal azotemia, etc.).

This leads me to another crucial documentation deficiency: the course of illness. The condition’s course and response to therapy need to be taken into account. In the preceding case, if the elderly patient’s mental status changes, hypotension, and pre-renal azotemia resolve within 4-6 hours of IV fluid replacement, then the patient was hypovolemic, not septic. Barring any other issues, the patient will be well enough to transfer back to a skilled nursing facility (SNF) after 48-72 hours. If the patient were truly septic, the course of illness and the patient’s response to sepsis therapy would not be as dramatic!

And this is why CMS and third-party auditors are focusing on sepsis admissions that are discharged after fewer than 72 hours. It’s all about probabilities. What do you think is the incidence of true sepsis (and severe sepsis, i.e., with an organ failure or worse, multi-organ failure) in patients discharged alive within 72 hours? It does not necessarily mean some of these patients were not truly septic. The clinical picture must support the diagnosis of sepsis. It is crucial to have documentation reflecting symptoms/manifestations/lab findings that show a clear delineation between sepsis as opposed to a regular infection (e.g., pneumonia, UTI, cellulitis, bacterial endocarditis, ascending cholangitis, acute pyelonephritis, etc.) If a patient is truly septic, then Sepsis 3 and qSOFA criteria will be met in most cases. Otherwise, there should be an explanation why criteria were not met.

It all comes down to the clinical picture and not the word “sepsis” or even “severe sepsis!” Even if you have sepsis/severe sepsis written consistently and ad nauseam, if the clinical picture and hospital course do not paint a true picture of sepsis, it’s not going to pass scrutiny. The words by themselves may not be enough, but the confluence of the words that build the story is key. Documentation of every patient’s unique narrative needs to reflect the clinical truth.. There is no way around it.

It is of the utmost importance that we accurately call sepsis when it is present and not call every kind of patient who comes through the hospital doors with an infection septic. This will ultimately impact on the future of healthcare, the use of hospital resources, and patient outcomes. If we send patients with every infection to an inpatient bed even if he or she does not need to be hospitalized, we are going to expose them to risks that can kill them and deprive those who are in dire need of the services they require.

And that is the real bottom line.