Utilization Strategies for Intensive Care Units

To the Editor.—  The article "Utilization Strategies for Intensive Care Units" by Drs Kalb and Miller' fails to address the important question of how to evaluate the effects of rationing critical care resources in a hospital. This entails collecting data on the impact of the intensive care unit (ICU) on the pool of patients both inside and outside the ICU when attendant bed shortages occur. Studying only patients in the ICU, the common methodology cited in the article by Drs Kalb and Miller, introduces selection bias into ICU research and neglects a basic tenet of epidemiology.As an example, take a hypothetical 100-bed hospital with a single 10-bed ICU. This system will generate descriptive data such as case mix and severity for ICU and non-ICU patients. Outcome data (including ICU, ward, and hospital mortality; occupancy rates; and lengths of stay) can be collected that will be unique to this particular