Extrapolation of measures of work for surveyed services to other services
N. L. Kelly, W. C. Hsiao, P. Braun, A. Sobol and M. DeNicola
Department of Health Policy and Management, Harvard University School of Public Health, Boston, MA 02115.
A national survey of physicians produced detailed data on the work involved
in performing 372 different services. This article describes methods
developed to extrapolate the study data to a larger universe of services,
defined by the Physicians' Current Procedural Terminology, edition 4.
Because data measuring work inputs for nonsurveyed services presently are
unavailable, we devised an extrapolation method that makes use of available
charge data without building their inherent distortions into the
extrapolated scale. To neutralize the effect of these distortions, we used
small, homogeneous families of services as the basic units for the
extrapolations and assumed that charges are reasonable indicators of
relative work within such families. To produce extrapolated work values
within each family, we multiplied an estimate of work based on survey data
for a benchmark procedure by charge-based ratios that represent the
relationships between surveyed and nonsurveyed services. These
extrapolations can be used in constructing a Resource-Based Relative Value
Scale.
