Prognostic evaluation of critically ill patients from the intensive care unit of the Hospital Beneficence Portuguese of Ribeirão Preto
© BioMed Central Ltd 2007
Published: 19 June 2007
A prognostic evaluation system was developed to measure the clinical severity of patients and to evaluate assistance quality, among other objectives. The APACHE II score (APII) analyses 12 clinical, physiological and laboratorial variables, through which the risk of death can be obtained, translating the patient severity into numerical values. The evaluation of the patient prognosis and the prediction of the risk of death for seriously ill patients are of great importance, requiring adequate intensive assistance.
To characterize the severity of patients, comparing the observed versus expected mortality, and to evaluate the ICU performance regarding assistance.
A prospective study in which the APII was calculated and the outcome of the patients admitted to our ICU from 24 June 2006 to 19 October 2006 was studied. For such calculation, the largest discrepancy from the reference values was considered in the first 24 hours in the ICU. The study included 202 patients, 116 men (57.4%) and 86 women (42.6%), with age varying between 22 and 94 years (mean age: 63.9 years). The most prevalent diseases were: postoperative cardiac care (14.8%), congestive heart failure (13.4%), acute arterial insufficiency (11.9%) and pneumonia (9.4%). The length of stay in the ICU varied from 0 to 63 days (average = 5.7 days). The cutoff value of the APII was 25 and the results of RO were 36.8%.
According to the ROC curve, a sensitivity of 87.5% and a specificity of 77% was observed for an APII cutoff value of 25; thus, 87% of the patients who died had APII ≥25 and 77% of the patients classified as high severity presented APII <25. The curve also showed that 88% of the patients who died presented RO ≥36.8% and 74% of patients classified as high severity had a RO <36.8%. When correlating the cutoff value of 25 from the APII with 36.8% for the RO, it was noticed that 96.9% of patients with APII ≥25 and RO ≥36.8% evolved to death and that 93.7% of patients with APII <25 and RO <36.8% were classified as high severity; there was only 4.95% of inconsistency. The global expected mortality was 44% while the observed mortality was 37%. APII <12 excluded death, and APII ≥45 confirmed death.
The population studied included patients of higher severity when compared with those described in the general literature. The observed mortality was less than the expected mortality, suggesting adequate assistance. The APII is a good prognostic index, and when used in the first 24 hours of internment presents high specificity and sensitivity to calculate the death risk.