Validation and use of resource utilization groups as a case-mix measure for long-term care

A companion article describes the development of a patient classification system for long-term care patients, Resource Utilization Groups (RUGs). Three potential limitations of this system and its development are addressed here: the use of a subjectively determined dependent variable, geographic limitation of the data to Connecticut skilled nursing facilities, and limited assessment of the quality of the facilities studied. Additional systems of Resource Utilization Groups were derived, using the same clustering technique but employing two separate data sets from the Battelle Human Affairs Research Center. These data bases provided an objective dependent variable, wide geographic distribution of both skilled nursing facilities and intermediate care facilities, and homes specifically selected on the basis of quality. The RUGs derived from the two sets of Battelle data and the initial RUG system showed remarkable similarity in their patient groupings and in the case-mix indexes developed for nursing homes. The concurrence of the results obtained for these three systems greatly strengthens the basis for the use of this classification system as a case mix measure for long-term care.