APACHE II analysis of a surgical intensive care unit population in a tertiary care hospital in Karachi (Pakistan)

Purpose: Performance of prognostic models deteriorates over time by changes in case-mix and clinical practice. This study was conducted to describe the case-mix of a surgical intensive care unit (SICU) and assess the performance of APACHE II scoring system in this cohort. Methodology: We analyzed 213 adult patients admitted to the surgical intensive care unit (SICU) of Aga Khan University Hospital, from January 2011 to December 2012 and the performance of APACHE II scoring system was assessed in this population. Results: The mean age of patients was 46.31 years (SD +/-18.43), 67.1% patients were male and mean length of ICU stay was 6.54 days (SD +/-7.18). Admissions to SICU were from seven service departments with the highest admissions from general surgery followed by trauma and neurosurgery. The mean APACHE II score of this SICU population was 15.89 (SD +/-8.06), 12.88 (SD +/-6.29) in survivors and 22.24 (SD +/-7.66) in non-survivors (p < 0.01). The overall mortality was 33%, with SMR of 1.0. No patient survived with an APACHE II score of more than 34. In this SICU population the calibration and discrimination of the APACHE II scoring system was acceptable, i. e. [(H-LS 11.76 (p= 0.16)] and (area under the receiver operating curve = 0.83). Conclusion: APACHE II scoring system allows meaningful analysis of SICU population, therefore, it is recommend, that this simple and cost effective scoring system should be used to identify patients with high risk of death to justify the decisions of withholding expensive therapies in resource limited settings.