Resurgence of Debate over Sepsis Definitions

I was recently reviewing the Work Plan for the U.S. Department of Health and Human Services (HHS) Office of Inspector General (OIG). Let me start by saying that your guess is as good as mine if the OIG will be allowed to continue completing planned reviews, but let’s say for the sake of argument that they will be able to do so. The planned review I want to discuss in this article is titled “Medicare Inpatient Hospital Billing for Sepsis.”

Some of you may already be aware of this planned audit but perhaps didn’t read the specifics. The rationale for this audit is that the Centers for Medicare & Medicaid Services (CMS) does not use the most recent definition of sepsis, e.g., Sepsis 3 or SOFA (Sequential Organ Failure Assessment), that was introduced in 2016, which better differentiates sepsis from a general infection.

Because sepsis is a frequently billed Medicare diagnosis, the OIG is concerned “that hospitals may be taking advantage of this broader definition, as they have financial incentive to do so.”

The OIG will analyze Medicare claim patterns for the billing of sepsis in 2023 for short-term acute-care inpatients, to examine how billing varies across hospitals. The OIG will also estimate how much could be saved if Medicare adopted the narrower, most recent definition of sepsis. Basically, the OIG is conducting this study to see if Medicare should follow in the footsteps of most private payors and audit sepsis claims using Sepsis 3 criteria.

Most clinical documentation integrity (CDI) and coding professionals are acutely aware of the struggle hospitals have faced since 2016, as they oscillate between using Sepsis 3 criteria to perform better on quality measures and SIRS (Systemic Inflammatory Response Syndrome) or Sepsis 2 criteria for better financial performance. Perhaps many thought this battle was put to rest within their facility, but it is likely to resurface, in light of this OIG audit, as well as inclusion of the sepsis bundle in the Hospital Value-Based Purchasing Program, effective with CMS fiscal year (FY) 2026.

According to Definitive Healthcare, which published the top 25 Diagnosis-Related Groups (DRGs) by diagnosis volume for 2023, MS-DRG 871, septicemia, or severe sepsis without mechanical ventilation of 96 or more hours with a major complication/comorbidity, comprised 8.6 percent of all DRGs and is the top MS-DRG by almost double the percentage of DRG 291, Heart Failure and Shock with a MCC, which is second-most prevalent at 4.5 percent. MS-DRG 872, septicemia, or severe sepsis without mechanical ventilation for 96 or more hours without an MCC, is tied for sixth, with 1.4 percent.

These claims likely represent patients diagnosed with SIRS or Sepsis 2 criteria, and would be the most at risk for denials, should Medicare update their definition of sepsis.

The prevalence of sepsis MS-DRGs should be considered in the context of Medicare’s Comprehensive Error Rate Testing (CERT) program. This program allows the CERT contractor to randomly review claims to measure the improper payment rate within the Medicare fee-for-service program. The projected improper payments for MS-DRGs 871 and 872 for 2024 is almost $66 million. The overwhelming cause of improper payments is coding errors, at 90.8 percent.

The problem with a high volume of sepsis cases based on the broader definitions is that these cases are likely to lower performance on the sepsis bundle quality measure. Under the Inpatient Quality Reporting Program, this measure had little impact on hospital finances. However, as it is incorporated into Hospital Value-Based Purchasing, low performance is likely to be heavily scrutinized, especially if it contributes to the hospital being penalized. Performance on the sepsis bundle will contribute to the safety domain, which is mostly comprised of healthcare-associated infections. In FY 2026, the safety domain contributes 25 percent to the overall performance under Hospital Value-Based Purchasing.

Sepsis is the only safety measure that is coding-based. Performance on the healthcare-associated measures is determined by routine surveillance data reported to the Centers for Disease Control and Prevention (CDC). ICD-10-CM codes are used to determine the measure population for the sepsis bundle. Although the sepsis bundle is supposed to focus on adults with a diagnosis of severe sepsis or shock, the logic used to identify inclusion cases is flawed.

The denominator includes inpatients 18 and over with an ICD-10-CM Principal or Other Diagnosis Code of Sepsis, Severe Sepsis, or Septic Shock, and without a code for COVID-19. Consequently, patients who are diagnosed with sepsis using SIRS or Sepsis 2 criteria can be included in the measure population, even though they may not have severe sepsis, which is defined as sepsis with organ failure.

The issue is that under Sepsis 3 criteria, what has historically been referred to as “severe sepsis or sepsis with organ failure” is simply sepsis, because organ failure attributable to sepsis must be present to diagnosis sepsis using SOFA criteria. In my opinion, this issue could be easily fixed by updating the denominator to require severe sepsis or septic shock as a secondary diagnosis, since it can’t be assigned as a principal diagnosis.

This would eliminate cases that are diagnosed using the older, broader definition because they are unlikely to meet severe sepsis criteria, and therefore are likely to contribute to lower performance.

The bottom line is that questions regarding how to define sepsis are likely to resurface, between inclusion of the sepsis bundle measure in Hospital Value-Based Purchasing and the OIG audit findings expected to be released in 2026.

Programming note:

Listen live this morning for the CDI report with Cheryl Ericson on Talk Ten Tuesday with Chuck Buck and Angela Comfort at 10 Eastern.