Curtiss F R
Am J Hosp Pharm. 1985 Oct;42(10):2168-74.
The national hospital data used by the Health Care Financing Administration (HCFA) to construct the DRG-payment weights for the Medicare prospective-pricing system are analyzed and evaluated. The database represented a 20% sample of all Medicare hospital bills in 1981. Each record contained 96 bytes of data in 28 field elements, including DRG assignment, Standard Metropolitan Statistical Area code, length of stay (LOS), pharmacy charges, and total hospital charges. There was considerable variation in the relative weight of pharmacy charges to total ancillary charges among DRGs; however, the degree of variation appeared to be similar for both rural and urban hospitals. Examination of the pharmacy charges for the top-10 Medicare DRGs revealed that the median pharmacy charge per DRG was consistently less than the average pharmacy charge per DRG. Average pharmacy charges per DRG were generally 50% greater for urban hospitals than rural hospitals. Average and median LOS per DRG also differed substantially, and the LOS for urban hospitals was approximately 20% longer than it was in rural hospitals for all DRGs. The standards derived from the pharmacy-charge and LOS data used by HCFA in developing the DRG-payment weights should be used cautiously. The heterogeneity of these data confirms the imprecision in constructing these weights and the need to use median rather than average statistics as standards in use-review programs.
对医疗保健财务管理局（HCFA）用于构建医疗保险预期定价系统的DRG支付权重的国家医院数据进行了分析和评估。该数据库代表了1981年所有医疗保险医院账单的20%样本。每条记录在28个字段元素中包含96字节的数据，包括DRG分配、标准大都市统计区代码、住院时间（LOS）、药房费用和医院总费用。不同DRG之间药房费用占辅助总费用的相对权重存在相当大的差异；然而，农村和城市医院的差异程度似乎相似。对医疗保险前10大DRG的药房费用进行检查发现，每个DRG的药房费用中位数始终低于每个DRG的平均药房费用。城市医院每个DRG的平均药房费用通常比农村医院高50%。每个DRG的平均和中位数住院时间也有很大差异，并且所有DRG的城市医院住院时间比农村医院长约20%。HCFA在制定DRG支付权重时使用的药房费用和住院时间数据得出的标准应谨慎使用。这些数据的异质性证实了构建这些权重的不精确性，以及在使用审查程序中需要使用中位数而非平均数统计作为标准。
