(p. 125) Severity of illness scoring systems 

Clinical outcome comparisons for research and quality assurance require risk adjustment measures validated in the population of interest. There are many scoring systems using intensive care unit (ICU)-specific or administrative data sets, or both. Risk-adjusted ICU and hospital mortality outcome measures may be not granular enough or may be censored before the absolute risk of the studied outcome reaches that of the population at large. Data linkage methods may be used to examine longer-term outcomes. Organ failure scores provide a method for assessing the intra-episode time course of illness and scores using treatment variables may be useful for assessing care requirements. Each adjustment system has specific merits and limitations, which must be understood for appropriate use. Graphical representations of the comparisons facilitate understanding and time-appropriate response to variations in outcome. There are, as yet, no universally-accepted measures for severity of illness and risk adjustment in deteriorating patients outside the ICU.