Volume 50 | Number 6 | December 2015

Abstract List

Kris C. L. Lee Ph.D., Kannan Sethuraman Ph.D., Jongsay Yong Ph.D.


To evaluate the relationship between hospital volume and outcome by focusing on alternative measures of volume that capture specialization and overall throughput of hospitals.

Data Sources/Study Setting

Hospital administrative data from the state of Victoria, Australia; data contain 1,798,474 admitted episodes reported by 135 public and private acute‐care hospitals.

Study Design

This study contrasts the volume–outcome relationship using regression models with different measures of volume; two‐step and single‐step risk‐adjustment methods are used.

Data Collection/Extraction Methods

The sample is restricted to ischemic heart disease () patients (‐10 codes: I20–I25) admitted during 2001/02 to 2004/05.

Principal Findings

Overall hospital throughput and degree of specialization display more substantive implications for the volume–outcome relationship than conventional caseload volume measure. Two‐step estimation when corrected for heteroscedasticity produces comparable results to single‐step methods.


Different measures of volume could lead to vastly different conclusions about the volume–outcome relationship. Hospital specialization and throughput should both be included as measures of volume to capture the notion of size, focus, and possible congestion effects.