Baltimore, MD - A Johns Hopkins University study seriously challenges the validity of the disease classification system now used by the federal government to determine how much hospitals will be reimbursed for patient care.
The study written by Susan D Horn, PhD; Gregory Bulkley, MD; Phoebe Sharkey, PhD; Angela F Chambers, MS; Roger A Horn, MD; and Carl J Schramm, PhD, JD was published in the July 4 issue of the New England Journal of Medicine. The study demonstrated that Diagnosis Related Groups (DRGs) often feil to account for the severity of a patient's illness, a major determinant of the actual cost of hospital care.
"Our findings showed that patient classification by DRGs do not adequately reflect the complexity and severity of illness, and further indicated that prospective payment programs based upon DRGs alone often unfairly and adversely discriminated against certain hospitals," says Susan D. Horn, PhD.
In a related study, the Hopkins team compared rates of reimbursement based on government's approach with the actual extent or severity of illness of the patient population and found both gross overpayments and substernal underpayments to both teaching and community hospitals. The authors found some institutions received as much as 59% too much and others as much as 25% too little for care rendered patients with a wide variety of illnesses.
The NEJM study compared hospital patients admitted for medical, surgical, obstetrical, gynecological and pediatric care in three university teaching hospitals, two community teaching hospitals and one community non-teaching hospital. (Hopkins was not one of the hospitals studied).
The study concluded: "An explicit assumption that HCFA has made in designing the present prospective payment system is that DRGs will account for the major differences in the cost of treatment among patients due to the severity of illness. To our knowledge there is no published study that supports this assumption and our findings are in direct conflict with it."
Horn says the assumption also has been that teaching hospitals are more expensive than community hospitals because they subsidize medical education and are less efficient.
"It is well recognized that the average charge per case in teaching hospitals, particularly university teaching hospitals, is higher than in community hospitals," Horn said. "This greater expense had been attributed primarily to presumed inefficiencies of operation and to the pass-through of the costs of research and teaching to patient bills. The findings in our study, however, suggested that academic teaching hospitals, which function as tertiary referral centers, are burdened with a greater proportion of more severely ill patients than many other hospitals."
HCFA has tried to compensate for higher patient care costs in teaching hospitals by adjusting the rate of reimbursement according to the ratio of medical residents to hospital beds. This approach is based upon the unsupported assumption that these higher costs are due to research and backing. However, the Hopkins study showed that a major reason for these higher costs is the greater burden of illness demonstrated by the type of patient that is referred to a teaching hospital. Furthermore, even some nonuniversity hospitals, with low resident to bed ratios, were actually seeing a very severe case mix, and therefore die current reimbursement strategy also discriminated against mem.
"Ideally," Horn adds, "prospective payment programs should seek to identify costs directly generated by the patient's illness, as distinct from other costs."
Horn also noted that more man a third of the classifications of illness in this study showed significant variations in severity of the illness. "Only when we adjusted reimbursement rates to account for the severity could much of the variation in charges among hospitals be consistentiy explained," she says.
To account for these variations, Horn used a previously developed Severity of Illness Index. The Severity Index is compatible with the current method of international classification of diseases and could easily be integrated into the DRG system. "A computerized version of the Severity of Illness Index will be available in early 1986," Hora says.