Journal of the Intensive Care Society 2018, Vol. 19(2) Supplement 1–162, Abstract EP.83
Analysis of 705 emergency elderly admissions to ICU (Aintree University Hospital NHS Foundation Trust, Liverpool, UK) found that APACHE II score was a fair predictor of hospital mortality but a poor predictor of one year mortality.
Abstract
Introduction: Increasing numbers of elderly patients (>/=80 years old) are presenting as emergencies to critical care. Such patients frequently have significant pre-existing co-morbidities and frailty, with poor critical care outcomes (Flaatten ICM 2017). In this context, critical care admission requires careful consideration. However, the existing literature frequently explores hospital mortality, ignoring longer term outcomes.
Methods: A retrospective observational study of elderly patients (>/=80 years) with unplanned admissions to the critical care unit at Aintree University Hospital (January 2010 – December 2016). We collected clinical information on acute physiology and co-morbidities (Functional Co-morbidity Score). We recorded hospital discharge status and 12-month survival data. HRA approval reference: 220258.
Results: We identified 705 emergency elderly admissions (mean 84 years, SD 3.2) during the study period. There were 354 (50.2%) female patients and survival to hospital discharge was 60.9% and 37.6% at one year. The APACHE II score was a fair predictor (AUCROC 0.705) of hospital mortality but a poor predictor (AUCROC 0.668) of one year mortality. We constructed a stepwise multivariate logistic regression model to identify clinically pertinent predictors of hospital (Table 1, AUCROC 0.754) and one-year mortality (AUCROC 0.671) when making admission decisions.
Conclusions: The APACHE II score was a fair predictor of hospital mortality and a poor predictor of mortality at one year. Simple clinical tools calculated at the point of admission performed better than the APACHE II score in predicting hospital mortality but not one-year mortality. This is likely due to multiple additional factors that impact on longer term mortality for hospital survivors
Introduction: Increasing numbers of elderly patients (>/=80 years old) are presenting as emergencies to critical care. Such patients frequently have significant pre-existing co-morbidities and frailty, with poor critical care outcomes (Flaatten ICM 2017). In this context, critical care admission requires careful consideration. However, the existing literature frequently explores hospital mortality, ignoring longer term outcomes.
Methods: A retrospective observational study of elderly patients (>/=80 years) with unplanned admissions to the critical care unit at Aintree University Hospital (January 2010 – December 2016). We collected clinical information on acute physiology and co-morbidities (Functional Co-morbidity Score). We recorded hospital discharge status and 12-month survival data. HRA approval reference: 220258.
Results: We identified 705 emergency elderly admissions (mean 84 years, SD 3.2) during the study period. There were 354 (50.2%) female patients and survival to hospital discharge was 60.9% and 37.6% at one year. The APACHE II score was a fair predictor (AUCROC 0.705) of hospital mortality but a poor predictor (AUCROC 0.668) of one year mortality. We constructed a stepwise multivariate logistic regression model to identify clinically pertinent predictors of hospital (Table 1, AUCROC 0.754) and one-year mortality (AUCROC 0.671) when making admission decisions.
Conclusions: The APACHE II score was a fair predictor of hospital mortality and a poor predictor of mortality at one year. Simple clinical tools calculated at the point of admission performed better than the APACHE II score in predicting hospital mortality but not one-year mortality. This is likely due to multiple additional factors that impact on longer term mortality for hospital survivors
