Prospective payment and the Medicare Hospice Benefit

The objective of this study was to determine the effects of very high cost patients on hospice financial status. Ten Pennsylvania hospices dually certified by Medicare were randomly selected and agreed to participate. Patient age, sex, diagnosis, length of stay and payer were fairly uniform across hospices. Payments varied by diagnosis and payer.

High cost patients were irregularly found in hospices; low cost patients were commonly and regularly distributed. Every hospice had at least one high cost patient. In one, the uncompensated payment for the 6.6 percent of patients defined as high cost ($7,300 and above) would have been 14.7 percent of total annual revenues. In another, uncompensated payments for high cost patients (9.8 percent) would have accounted for 17.2 percent of revenue. In 96.3 percent of the instances patients utilized less than the Medicare Hospice Benefit maximum allowable cost ($7,300); and, 98.8 percent of the time patients stayed less than the maximum allowable length of time of 210 days.

A logistic regression model found long length of stay (p < 0.0001), Medicare hospice benefit as primary payer (p <0.0001), any hospitalization during hospice stay (p < 0.003) and cerebrovascular disease diagnosis (p < 0.02) to be significantly related to high cost.

Between the time the study was planned and completed, Medicare instituted a reinsurance program allowing unused funds below the maximum allowable limit from one patient to be used for patients who exhausted their benefits. Thus, no study hospice was adversely affected by high cost patients. However, it should serve as an object lesson to Medicare in using prospective payment. A normal or near-normal distribution of patients by cost cannot be assumed for small institutions.