Specificity and reliability of prognostic indexes in intensive care evaluation: The spontaneous cerebral haemorrhage case

Objective To determine the reliability of a generic index such as Simplified Acute Physiology Score II (SAPSII), compared with a specific one Intra Cerebral Haemorrhage score (ICH score), as an intensive care unit (ICU) outcome predictor when evaluating a general facility that frequently treats a specific type of patients - those with spontaneous cerebral haemorrhage. Methods The study cohort consisted of a random sample of patients (81) admitted to Modena's Policlinico Teaching Hospital's ICU with spontaneous ICH over a 24-month period. Main outcome measure SAPSII, ICH score, overall mortality. Results The mean ICH score for the 32 surviving patients was 3.41 ± 1.012 while for the 49 deceased patients was of 4.24 ± 0.855 (P = 0.000). The mean SAPSII value for the 32 surviving patients was 49.09 ± 16.58 while for the 49 deceased patients was 49.51 ± 15.93. SAPSII, ICH scores were analysed for mortality, by receiver operating characteristic curves: the area under the curve was significant for ICH, not-significant for SAPSII. Conclusions Regional quality controls use generic prognostic indexes (SAPSII) in relation to mortality and outcome to assess ICUs, which is appropriate when dealing with a general facility when there is not a predominant type of patient, but it may bias the evaluation if the population with specific pathologies (ICH), not included in the general index, is statistically considerable, leading to an incorrect criticality assessment, an inappropriate strategic plan and the subsequent inefficient resource allocation. © 2009 Blackwell Publishing Ltd