The l f e expectancy for survivors of critical illness has not been reported in the United Kingdom. The aim of this study was to compare the long-term survival of intensive care patients with that of an age and sex-matched population. AN patients admitted to a general intensive therapy unit over 4 years (n = 1168) were included in the study. Details of the survivors were forwarded to the Registrar General for Scotland, who then issued copies of death certijicates as the survivors died. The survival curve of patients discharged from the intensive therapy unit was significantly diyerent from that of the normal population. The risk of dying in the first year after discharge was 3.4 times higher (95% confidence intervals 2.7-4.2) than that in the normal population. It is not until the start of the fourth year after discharge that the probability of death matched that of the normal population.
Key wordsIntensive therapy; survival.Effectiveness may be expressed as a balance between cost and outcome but assessing the effectiveness of intensive therapy is difficult because of wide variations in cost and outcome [ 11. Costs vary according to treatment regimens and different facilities offered by different intensive therapy units (ITUs). Outcome may be measured as the quality and duration of survival after the patient has returned home.Reports of quality of life (QOL) after intensive therapy are inconsistent; some studies suggest an improvement in QOL [2] while others suggest deterioration [3-51.Intensive care provides life-saving rather than specific therapy aimed at an underlying disease. Therefore, the best duration of survival 'of patients without chronic or malignant disease is normal life expectancy. Attempts to relate the numbers of survivors with normal life expectancy to the cost of critical illness may estimate effectiveness and yet partly overcome the variability of intensive therapy. Therefore the aims of this study were to compare the long-term survival of intensive care patients with an age and sexmatched population and attempt to apply long-term survival as a measure of effectiveness.
MethodsAll patients (1168 in total) admitted to the ITU of the Western Infirmary, Glasgow, between June 1985 and June 1989 were studied. The patients' age, diagnosis and APACHE I1 score [6] were recorded on admission. On discharge, details of the ITU survivors were forwarded to the Office of the Registrar General for Scotland. The Registrar General issued a copy of the death certificate when any patient within the study group died before 1 January 1993. Therefore patients were followed for between 3.5 and 7.5 years, depending upon the date of admission.Calculation of a theoretical survival curve for an age and sex-matched normal population is possible using statistics released by the Scottish Office [7J Similarly, expectation of life of patients with a known age and sex distribution can be calculated from abridged life tables [8].During the study period, the treatments of 90 patients were costed accurately on an individual and...