Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Background: Hospitals are the most expensive component of the health care system and are increasingly held accountable for both efficiency and quality performance. This places a high priority on making the best possible use of resources and on understanding, measuring and improving hospital efficiency. Development and application of sound measures of hospital efficiency and quality will play a central role in funding and accountability initiatives. This study compared different approaches to efficiency performance and evaluated relationships between efficiency and quality performance.

Study Setting, Data and Analyses: We examined hospital efficiency and quality in 93 acute hospitals in Ontario, Canada. We evaluated the current measure of efficiency in Ontario: Unit Cost (UC) performance (deviation from average cost per weighted case). We also calculated efficiency scores using Data Envelopment Analysis (DEA). DEA is a linear optimization technique that is explicitly designed to examine production efficiency when there are both multiple inputs and multiple outputs. We compared UC with DEA efficiency performance scores and compared both to measures of clinical quality (adverse events and readmissions) and patient satisfaction. Efficiency was based on computer-audited financial statements, clinical quality on acute care discharge abstracts and patient satisfaction on a comprehensive survey of all acute hospital patients. Data are collected as part of the Ontario Hospital Report. All comparisons examined rankings and categorization in the top and bottom decile and quartile for each measure. Data from all sources was available for 93 acute hospitals (of 115 with more than 3,650 acute patient days) reflecting activity between April 2004 and March 2005.

Results: Ranking comparisons (using deciles) indicated that almost 60% of facilities were identically categorized using UC or DEA efficiency measures. Only one facility was ranked at the opposite performance category (efficient by DEA but inefficient by UC). DEA efficiency scores suggest variable returns to scale and more efficient hospitals in urban and teaching categories. DEA efficiency scores penalized adverse events but rewarded re-admissions. UC comparisons provided similar results but provided a greater reward to hospitals with more re-admissions. Nearly 50% of hospitals with the best UC performance had the highest rates of readmission. Both UC and DEA similarly indicated low correlation with patient satisfaction.

Implications: While, different methods of determining efficiency provide different performance rankings, DEA provided a lesser reward to hospitals with high readmission rates. Hospital efficiency measurement must consider the relationship with clinical quality performance.