A Multivariate Approach for Classifying Hospitals and Computing Blended Payment Rates

Prospective payment for inpatient hospital care is based on the ideal that hospitals that produce similar outputs, as measured by the types of cases the hospital treats, should be paid similar prices. However, similar output is a multidimensional concept. Thus operationalization of this ideal will ultimately require a more complex framework for determining hospital payment rates than currently employed at either the federal or state level. This article illustrates a multidimensional approach to achieve this objective. This technique, called Grade of Membership, is used to generate a unique type of hospital group and to characterize individual hospitals in terms of their degree of similarity to these groups. In addition, a new concept of grouping is described, a variable set based on hospitals' internal cost structure is developed and used, and ordinary least squares regression is employed to compute prices for these groups. With the use of simulation analysis, these groups are compared with more conventional groups

Actuarial science · Econometrics · Economics · Ideal (ethics) · Multivariate analysis · Multivariate statistics · Operationalization · Ordinary least squares · Payment · Prospective payment system · Set (abstract data type) · Statistics · Computer Science · Efficiency Analysis Using DEA · Healthcare Policy and Management · Mathematics · Medicine

• Classification Model

  Open Access•David J Jackson, Max A Woodbury et al.•Bulletin of Sociological…•1988

  Add to list

• The Use of Grade of Membership Analysis to Evaluate and Modify Diagnosis-related Groups

  Kenneth G Manton, James C Vertrees•Medical Care•1984

  Add to list