Purpose of review: In this article we discuss our experiences benchmarking eight ICUs in The Netherlands. Benchmarks must be carefully designed and implemented to generate meaningful results. We define prerequisites that we have identified for successful benchmarking and discuss the development, implementation and results of ICU benchmarks that we have completed.
Recent findings: Previous articles have discussed benchmarking ICUs, but there are still few studies of significant size and appropriate design that measure the impact of benchmarking on outcomes. Perhaps the most well known, and still best example of a benchmarking study designed to measure outcome improvements is the work of Pronovost et al. in Michigan ICUs.
Summary: Benchmarking is an increasingly common activity, however it is difficult to prove that benchmarks result in improved outcomes. Concurrent with our benchmarking activities the Standardized Mortality Ratio in Dutch ICUs has decreased. We have been able to show that larger ICUs in our benchmarks generally had improved outcomes despite a higher average patient severity. Quality assurance in healthcare is maturing and benchmarks will become an increasingly useful way of comparing performance between institutions.