3M Commentary

The Real Impact of DRGs: Meaningful Data that Changed Hospital Management

July 17, 2014 / By Richard Averill, MS

A recent blog by François de Brantes, executive director of HCI3, titled “Letting the Facts Get in the Way of So-called Truths,” is highly critical of the DRG based Medicare inpatient prospective payment system (PPS). He urges readers to discover the facts about DRGs, a system he describes as endorsed by “agents of the status quo” that produces “meaningless comparisons” of patient data, with hospitals “being hurt more than helped by false truths.” As a member of the research team that developed Diagnosis Related Groups in the late 1970s, I want to respond to his assertions. Mr. de Brantes’ blog is rife with errors and distortions of fact; any valid points are lost in a barrage of misinformation.

De Brantes begins with the claim that using coded inpatient stays “…to compare hospitals by grouping them according to the DRG, you end up with massive heterogeneity of patients that renders comparisons meaningless.” He points to DRG assignments for a “chest pain” diagnosis and a coronary artery bypass graft (CABG) procedure as evidence of this heterogeneity. His examples, however, instead reveal a limited understanding of official coding guidelines and coding practices.

Chest pain in de Brantes’ view is “…widely used as a DRG assignment, even when the treatments performed in the facility indicate something else is going on,” implying that if a more definitive diagnosis had been established based on the results of a diagnostic procedure, the chest pain would still be coded as principal diagnosis. This is simply not accurate.

The Uniform Hospital Discharge Data Set (UHDDS) for reporting inpatient data defines the principal diagnosis as “that condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care.”[i] If a diagnostic procedure yields a more definitive diagnosis (e.g., diaphragmatic hernia) then it would be coded as the principal diagnosis and not the chest pain. If the patient undergoes a significant therapeutic procedure such as a pacemaker insertion, then the assignment of the MS-DRG would be based on that procedure. Essentially, a patient would only be assigned to the MS-DRG for “chest pain” if none of the diagnostic tests yielded a definitive diagnosis and no significant therapeutic procedure was performed. In this situation, the coding guidelines are clear:

“While specific diagnosis codes should be reported when they are supported by the available medical record documentation and clinical knowledge of the patient’s health condition, there are instances when signs/symptoms or unspecified codes are the best choices for accurately reflecting the healthcare encounter… It would be inappropriate to select a specific code that is not supported by the medical record documentation or conduct medically unnecessary diagnostic testing in order to determine a more specific code.”[ii]

De Brantes claims that for patients who have a coronary artery bypass graft (CABG) “…there’s a difference between an elective bypass and one caused by a heart attack. Classifying all CABGs together, irrespective of whether it’s elective or not, doesn’t make much sense.” He suggests that these patients are all assigned to the same CABG MS-DRG, further contributing to “massive heterogeneity” in comparative data. In fact, CABGs performed on patients suffering acute myocardial infarctions (AMIs) and elective CABGs are assigned to distinct MS-DRGs.

There are six CABG MS-DRGs (231-236), four of which (231-234) require that either an inpatient cardiac catheterization or a percutaneous transluminal coronary angioplasty (PTCA) be performed in addition to the CABG. That leaves two CABG MS-DRGs (235-236), which are only assigned if neither of these two additional procedures is performed. AMI patients virtually always have an inpatient cardiac catheterization or a PTCA performed prior to a CABG, and thus would be assigned to MS-DRGs 231-234. Elective CABGs, on the other hand, would have a cardiac catheterization performed on an outpatient basis prior to admission, resulting in assignment to MS-DRGs 235-236. Note also that MS-DRGs 231-236 are further differentiated by whether the patient has a major complication or comorbidity.

Despite de Brante’s claim of “massive heterogeneity,” CMS has concluded that it was the clinical homogeneity of the DRGs that was crucial to the success of the Medicare prospective payment system.

“Central to the success of the Medicare inpatient hospital prospective payment system is that DRGs have remained a clinical description of why the patient required hospitalization.”[iii]

The inherent language and communications power of DRGs created a link between the clinical and financial aspects of care resulting in behavior changes that fundamentally changed how hospitals were managed.

De Brantes next states, “…And guess what, the DRG assigned is, wait for it, ‘optimized’ to the highest reimbursement,” the implication being that hospitals are reporting codes that are not fully justified. Is he suggesting that massive fraud is being perpetrated by hospitals relative to the assignment of DRGs?

Lack of proof of this claim notwithstanding, suggesting that a hospital’s coding function is driven exclusively by the desire to optimize DRG reimbursement is a disservice to hospitals and health information management (HIM) professionals who work diligently to ensure their organizations are compliant with all coding rules and guidelines. These professionals understand that coding is about more than just payment. DRG assignment not only determines reimbursement, but has all kinds of malpractice and quality of care implications. A reported diagnosis for which there was no active treatment can be a disaster in a malpractice suit. Over-reporting diagnoses can make a hospital appear to be a poor performer on complication rate comparisons, which would result in major financial consequences for hospitals in states like Maryland, New York and Texas where healthcare payment is incentivized based on quality outcomes.

Finally, de Brantes claims that DRGs were developed “…not as a mechanism to measure the performance of hospitals and associated physicians in the management of a condition or the treatment of a procedure.” My recommendation to de Brantes is that he read any of the original journal articles published on DRG research, which tell a different story.

Our research team at Yale University’s School of Management explicitly developed DRGs as a management tool. Payment applications were an afterthought. We described our intent as follows:

“It is expected that there are collections of customers [patients] who consume similar resources and for whom the ‘production’ process follows a statistically predictable path.  If such customer [patients] classes can be identified by their attributes, then one has the basis for control of the ‘production’ process.” [iv]

We were focused on developing a management tool that could be used as the basis for controlling a hospital’s “production process” – the process of patient care management – and in doing so, change the fundamental management structure of hospitals:

“Under the traditional organization structure of a hospital, there is no department delegated the specific responsibility of assuring that the individual patients are financially well managed.”[v]

Once we had developed DRGs, we did extensive research to understand the actual cost associated with patients in each DRG.

“Since the care of patients is the basic service of a hospital, the goal of case mix cost accounting is to provide a complete financial picture of the costs of treating individual patients grouped into similar classes based on use of resources… A case-mix cost accounting system will provide an integrated picture of the financial consequences of providing this care to individual patients.”[vi]

Suggesting that DRGs were initially developed for any purpose other than as a management tool simply is not historically accurate.

DRGs revolutionized hospital management by giving hospitals clinically meaningful data to manage product line P&Ls and improve patient care processes. The MS-DRG prices essentially represent an efficiency standard that allows hospitals to compare their production efficiency to a common national norm. By 1990, seven years after the adoption of DRGs for Medicare payment, Medicare was spending $17 billion less per year than it would have if the DRG PPS had not been implemented.[vii] Those savings kept the Medicare inpatient trust fund from certain bankruptcy. Furthermore, more than 20 other countries have adopted payment or budgeting systems based on DRGs. It’s hard to believe a “meaningless” system of “false truths,” as de Brantes deems it, would have that kind of impact.

Richard Averill, MS, is Senior Vice President, Clinical and Economic Research for 3M Health Information Systems.

[i] Uniform Hospital Discharge Data Set (UHDDS) “Definition of Principal and Other [Secondary] Diagnoses.” 1992. 50 Federal Register 31039; adopted 1986, revised 1992.

[ii] Centers for Medicare and Medicaid Services. ICD-10-CM/PCS: The Next Generation of Coding, April 2013.

[iii] Medicare Program; Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 2002 Rates, 66 Federal Register, (proposed May 4, 2001), pt. 22669.

[iv] Mills, Fetter, Riedel, Averill (1976, July).  An Interactive Computer System for the Analysis of Health Care Data, Medical Care, 14(7).

[v] Thompson, Averill, Fetter (1979). Planning, Budgeting, and Controlling- One Look at the Future: Case-Mix Costs Accounting.  Health Services Research.

[vi] ibid.

[vii] Russell, Manning (1989). The Effects of Prospective Payment on Medicare Expenditures,” New England Journal of Medicine, 320(7).