Measuring the effects of casemix on outcomes

Participation in the measurement of population health and health care outcomes has become an explicit professional duty for doctors, but comparisons are difficult to make because outcomes are conceptually complex and largely qualitative. Observational data, particularly from routine hospital statistics, are useful complements to experimental data provided that their variable quality is taken into account and adjustments are made to minimize bias and confounding and to allow for the effects of differences in casemix, which are problematic because of the nature of severity.