The authors of "Clinical Overlap Among Medical Diagnosis-Related Groups"1 have described, in my opinion, several serious limitations of the diagnosis-related group (DRG) patient classification. First, understanding the clinical interrelatedness of International Classification of Diseases—Ninth Revision— Clinical Modification (ICD-9-CM) diagnosis codes is critical to the design of a valid patient classification from both a clinical and payment perspective. Second, and equally important, is the way those codes are used to assign patients to categories. The DRG patient classification does not address either of these issues adequately. In fact, as Drs Iezzoni and Moskowitz point out, the problem of clinical overlap is intrinsic to the DRG methodology.
Although DRGs may predict hospital resources used on average (eg, average length of stay and average charges or costs of services), they do not define patient types that are clinically specific. Drs Iezzoni and Moskowitz indicate why this is true conceptually and provide examples
Young WW. Substitution Permissible. JAMA. 1986;255(7):942–943. doi:10.1001/jama.1986.03370070096036