The ABCs of DRGs

Learn about the history and significance of diagnosis-related groups (DRGs).


Diagnosis-related groups (DRGs) were originally developed in the early 1980s as a collaborative project between Robert B. Fetter, PhD, and John D. Thompson, MPH, of Yale University. The DRG classification is intended to categorize patients by their similar clinical characteristics and costs.

Image by Getty Images
Image by Getty Images

In 1983, Medicare adopted the DRG methodology (now known as CMS-DRGs) for hospital inpatient care reimbursement, with the intention of curbing skyrocketing health care costs. Since 1983, many other DRG systems have been developed and used throughout the world, most notably the 3M All Patient Refined (APR) DRG system, which is widely used in the United States for non-Medicare patients.

In 2007, CMS adopted Medicare Severity DRGs (MS-DRGs) to better differentiate patients' severity of illness and associated costs of care. Each of the original CMS-DRGs had either one (singlet) or two (doublet) levels of severity and reimbursement. In contrast, most MS-DRGs have three (triplet) levels, although there are still some singlet and duplet MS-DRGs.

An MS-DRG is determined by the principal diagnosis, the principal procedure, if any, and certain secondary diagnoses identified by CMS as comorbidities and complications (CCs) and major comorbidities and complications (MCCs). A comorbidity is a condition that existed before admission; a complication, in this context, is simply any condition occurring after admission, not necessarily a complication of care. Over 14,000 ICD-10-CM diagnosis codes are designed by CMS as CCs and about 3,200 codes are MCCs.

Every year, CMS assigns a “relative weight” to every DRG. The relative weight determines the reimbursement associated with that DRG and reflects the patient's severity of illness and cost of care during hospitalization. A higher relative weight is associated with longer length of stay, greater severity of illness, and higher reimbursement. For example, DRG 189 (respiratory failure) has a relative weight of 1.2353 and DRG 312 (syncope) is 0.8015.

The principal diagnosis is the condition established, after complete evaluation, to be primarily responsible for and the primary focus of the admission. The condition, or at least some signs or symptoms (including test results) of it, must have been present on admission. It often takes several days to identify the actual cause of signs, symptoms, and abnormal findings that were present on admission.

The “focus” of an admission is an important concept, which should guide the selection of a primary diagnosis. Factors such as severity, risks, complexity of evaluation and care, medications (IV vs. oral) and their risks, diagnostic procedures, number of consultants, and intensity of monitoring (e.g., frequency of vital signs or neuro-checks; nursing time; intensive care) should be considered.

Take, for example, a patient admitted for heart failure and pneumonia where the heart failure responded quickly to IV furosemide with no other particular management required, but the pneumonia was prolonged, requiring pulmonary and infectious disease consults and IV administration of potentially nephrotoxic antibiotics. Pneumonia must be assigned as the primary diagnosis, not heart failure.

The primary diagnosis is determined by the coder, who reviews the entire chart, including documentation by clinicians, and strictly follows ICD-10 instructions and other coding guidelines with which clinicians are typically unfamiliar. This is why accurate, precise, and specific documentation by clinicians is so important. Because of coding rules and definitions, the coded primary diagnosis may be different than the clinical impression of the primary reason for admission.

As an example, consider the patient admitted with bilateral staphylococcal pneumonia, causing a mild degree of sepsis and respiratory failure. Pursuant to coding rules, as counterintuitive as it may seem to a clinician, sepsis must be assigned as the primary diagnosis.

Inpatient procedures are coded on hospital claims using the ICD-10 Procedural Coding System, not the AMA's Current Procedural Terminology, 4th Edition, which is used for all clinician services. CMS requires all “significant” procedures to be coded by the hospital. A significant procedure is one that is surgical in nature, carries a procedural or anesthetic risk, or requires specialized training.

Most significant procedures are designated as “OR procedures” by the MS-DRG, and most of these are performed in an operating room. However, some OR procedures (e.g., endotracheal intubation, transbronchial biopsy or bedside excisional wound debridement) do not have to be performed in an operating room. Some common procedures that are designated as non-OR include esophagogastroscopy, colonoscopy, diagnostic bronchoscopy, and endobronchial biopsy.

There are two clinical types of DRG. A medical DRG is one where no OR procedure is performed. When an OR procedure is performed, a surgical DRG is assigned. As an example, sepsis due to a stage 4 pressure ulcer with cellulitis would be assigned to DRG 871 (sepsis with MCC). If excisional debridement is performed, the DRG changes to 853 (infectious disease with OR procedure with MCC).

How do CCs and MCCs fit in? CCs and MCCs are secondary diagnoses that may impact the DRG assignment (see examples in Table). In most cases, a CC increases the relative weight and an MCC results in an even higher weight that impacts severity and reimbursement. For example, DRG 293 (heart failure without CC/MCC) has a relative weight of 0.6656 whereas DRG 291 (heart failure with MCC) is 1.3454.

Ask Dr. Pinson

Q: What do you recommend clinicians document to capture chronic myocardial injury? Is that still considered a type of demand ischemia, or do you recommend something like “non-myocardial infarction (MI) troponin elevation”?

A: Thanks for this interesting question. By definition, in “chronic” myocardial injury, troponin levels are chronically elevated above the 99th percentile, usually due to chronic conditions like end-stage heart failure, myocarditis, cardiomyopathy, certain chemotherapeutic agents, or infiltrative diseases (e.g., amyloidosis, sarcoidosis). Also keep in mind that troponin levels are often chronically elevated in chronic kidney disease without any myocardial injury present.

The terms chronic myocardial injury, demand ischemia, and “non-MI troponin elevation” are problematic. The only ICD-10-CM codes available for myocardial “injury” (acute or chronic) are trauma codes that rarely describe the patient's actual medical condition. Demand ischemia describes an acute supply/demand mismatch without elevation of troponin above the 99th percentile and therefore is not myocardial injury at all. “Non-MI troponin elevation” describes non-ischemic myocardial injury that can be either acute or chronic and is coded as an abnormal blood test, using code R79.89 (Other specified abnormal findings of blood chemistry), not a diagnosis.

In all cases of chronic myocardial injury, the confirmed or suspected most likely cause should be specified by clinicians.

Q: I am a hospitalist and have a question. A heated debate arose among my group last week regarding the following scenario: I admit a patient who is extremely ill or moribund and needs a very high level of care (such as pressors, intubation, etc.), which would normally require a stay longer than two midnights, but in my professional medical opinion I honestly expect them to die within 24 hours. Should I admit that patient under inpatient or observation status? And what about a similar situation where I expect a patient to go formally on hospice within 24 hours, but the patient or family is not ready to do so at the time of admission? Thank you!

A: First of all, let me say that Medicare's two-midnight rule does not determine the medical necessity of an inpatient admission. It simply provides that an admission expected to last over two midnights is presumed to be medically necessary if reviewed by a Medicare contractor.

Many admissions that are medically necessary do not necessarily last over a period of two midnights. Any patient who is extremely ill or moribund should be admitted as an inpatient, not as observation, whether a hospice patient or not.