Cross-specialty linkage and extrapolation of resource-based relative value scales

This article describes methods used to produce a RBRVS (resource-based relative value scales), a common scale from two specialties (internal medicine and general surgery) and explains the newly developed extrapolation process within each specialty. To produce a common scale, we selected six 'same' services as linking services common to both specialties. Then we used the bi-weighted least squares method to locate all the same services on a single, common scale. By using the same method, we tried to extrapolate all the services within each specialty, not by the method of Kelly et al, dividing all the services within the specialty into families (small homogeneous groups of services) to apply charge-based ratios. To compare both methods, we extrapolated all the services of general surgery according to each method. With the correlation analysis to compare both results to American RVUs, we found that general surgery's RVUs from our own extrapolation method turned out to be more highly correlated with American RVUs than from Kelly's extrapolation method. Consequently, extrapolation with bi-weighted least squares method gave reasonable results.