Verification of the Acute Physiology and Chronic Health Evaluation scoring system in a Hong Kong intensive care unit

OBJECTIVES: To validate the Acute Physiology and Chronic Health Evaluation (APACHE II) severity of illness scoring system in Chinese patients in a multidisciplinary intensive care unit (ICU) in Hong Kong. To audit the service and utilization of an ICU with a low ICU to hospital bed ratio. DESIGN: Prospective data collection and review. SETTING: A 12-bed multidisciplinary ICU within a 1,430-bed tertiary care university hospital. PATIENTS: Data from 1,573 of 1,814 consecutive patients admitted to the ICU from May 1988 to November 1990 were studied. The patients were all Chinese. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The patients' clinical details and APACHE II scores were recorded on day 2 of admission and reviewed at hospital discharge or after death. The APACHE II scores, risk of death values, age, and length of ICU stay between survivors and nonsurvivors were compared by two-sample t-tests. Relationships between mortality and APACHE II score, risk of death, and results from previous studies were investigated using the Pearson product-moment coefficient and regression analysis. Predictive capacity of risk of death was assessed by receiver operating characteristic curve analysis. The hospital mortality rate for study patients was 36%. Survivors were younger, had shorter ICU stays, lower APACHE scores, and lower risk of death values than nonsurvivors (p < .001). There was close correlation (r2 = .81, .77, and .76 for all patients, operative group, nonoperative group, respectively) between APACHE II scores and predicted risk of death values. Risk of death was an accurate group predictor of death in all patients and in separate operative and nonoperative groups. Areas under the receiver operating characteristic curves were 0.89 (all patients), 0.85 (operative), and 0.88 (nonoperative). Neither the Apache II scores nor risk of death scores were sufficiently accurate to predict outcome of individual patients. There was close concordance between observed and predicted mortality of patient groups. Mortality ratio was 0.97 (all patients), 0.89 (operative group), and 1.02 (nonoperative group). Chronological age, per se, was not a good predictor of mortality. The audit of the ICU service showed a short length (4.2 days) of ICU stay and high bed occupancy (80%). Subgroups of low-risk, postoperative patients with good outcomes and poor-risk patients admitted after cardiopulmonary arrest with a high mortality rate were identified. CONCLUSIONS: The APACHE II scoring system was an accurate predictor of group outcome in a Chinese population, making it suitable for comparisons between countries. Application of the APACHE II scoring system in a clinical audit facilitates critical appraisal of an ICU service. Problems identified by the study were a shortage of ICU beds and delayed referrals of patients.