Concept: Low-Variation / High-Variation Hospital Utilization

Introduction

A low-variation / high-variation framework was developed (by Wennberg (1984) ) to identify both medical conditions and surgical procedures for which hospital utilization is fairly stable across small areas ("low-variation") and for which utilization can vary a great deal across small areas ("high-variation"). Rates of hospitalization for high-variation conditions such as pneumonia tend to reflect both physician decision-making (physician practice style) differences in treatment options and the influence of factors other than illness. High-variation conditions thus might serve as indicators, or red flags, for potentially reducible rates of hospitalizations. On the other hand, rates of hospitalization for low-variation conditions tend to reflect the incidence / prevalence of such conditions in the population and a high degree of professional consensus on treatment decisions. They might serve as indicators of necessary hospitalizations.

SAS FORMATS

The most recent low-variation/high-variation format is LHVMS.

NOTES

A comparison of the most recent information from Dartmouth, New Hampshire (Diagnosis Related Groupings [DRGs™] 1-494, 1994) regarding high-variation/low-variation groupings with the LHVMS format was performed. Aside from the pregnancy/childbirth DRGs (370-391), mental health DRGs (425-432), and DRGs 468, 469, and 477 which were re-assigned for the LHVMS format, the following outlines the differences between the most recent Dartmouth format and the LHVMS format:


* 351 (Sterilization, male) is a high-variation medical condition (the LHVMS format has it as a high-variation surgical procedure (minor)).

* 109 and 474 are no longer valid in DRG 11.

As well, there are some DRGs that are not currently part of the LHVMS format and have been assigned to variation categories by Dartmouth differently.