Short-term and long-term outcome prediction with the Acute Physiology and Chronic Health Evaluation II system after orthotopic liver transplantation

Deep learning to predict long-term mortality in patients requiring 7 days of mechanical ventilation

Background

Among patients with acute respiratory failure requiring prolonged mechanical ventilation, tracheostomies are typically placed after approximately 7 to 10 days. Yet half of patients admitted to the intensive care unit receiving tracheostomy will die within a year, often within three months. Existing mortality prediction models for prolonged mechanical ventilation, such as the ProVent Score, have poor sensitivity and are not applied until after 14 days of mechanical ventilation. We developed a model to predict 3-month mortality in patients requiring more than 7 days of mechanical ventilation using deep learning techniques and compared this to existing mortality models.

Methods

Retrospective cohort study. Setting: The Medical Information Mart for Intensive Care III Database. Patients: All adults requiring ≥ 7 days of mechanical ventilation. Measurements: A neural network model for 3-month mortality was created using process-of-care variables, including demographic, physiologic and clinical data. The area under the receiver operator curve (AUROC) was compared to the ProVent model at predicting 3 and 12-month mortality. Shapley values were used to identify the variables with the greatest contributions to the model.

Results

There were 4,334 encounters divided into a development cohort (n = 3467) and a testing cohort (n = 867). The final deep learning model included 250 variables and had an AUROC of 0.74 for predicting 3-month mortality at day 7 of mechanical ventilation versus 0.59 for the ProVent model. Older age and elevated Simplified Acute Physiology Score II (SAPS II) Score on intensive care unit admission had the largest contribution to predicting mortality.

Discussion

We developed a deep learning prediction model for 3-month mortality among patients requiring ≥ 7 days of mechanical ventilation using a neural network approach utilizing readily available clinical variables. The model outperforms the ProVent model for predicting mortality among patients requiring ≥ 7 days of mechanical ventilation. This model requires external validation.

Validation of APACHE II, APACHE III and SAPS II scores in in-hospital and one year mortality prediction in a mixed intensive care unit in Poland: a cohort study

BACKGROUND

There are several scores used for in-hospital mortality prediction in critical illness. Their application in a local scenario requires validation to ensure appropriate diagnostic accuracy. Moreover, their use in assessing post-discharge mortality in intensive care unit (ICU) survivors has not been extensively studied. We aimed to validate APACHE II, APACHE III and SAPS II scores in short- and long-term mortality prediction in a mixed adult ICU in Poland. APACHE II, APACHE III and SAPS II scores, with corresponding predicted mortality ratios, were calculated for 303 consecutive patients admitted to a 10-bed ICU in 2016. Short-term (in-hospital) and long-term (12-month post-discharge) mortality was assessed.

RESULTS

Median APACHE II, APACHE III and SAPS II scores were 19 (IQR 12-24), 67 (36.5-88) and 44 (27-56) points, with corresponding in-hospital mortality ratios of 25.8% (IQR 12.1-46.0), 18.5% (IQR 3.8-41.8) and 34.8% (IQR 7.9-59.8). Observed in-hospital mortality was 35.6%. Moreover, 12-month post-discharge mortality reached 17.4%. All the scores predicted in-hospital mortality (p < 0.05): APACHE II (AUC = 0.78; 95%CI 0.73-0.83), APACHE III (AUC = 0.79; 95%CI 0.74-0.84) and SAPS II (AUC = 0.79; 95%CI 0.74-0.84); as well as mortality after hospital discharge (p < 0.05): APACHE II (AUC = 0.71; 95%CI 0.64-0.78), APACHE III (AUC = 0.72; 95%CI 0.65-0.78) and SAPS II (AUC = 0.69; 95%CI 0.62-0.76), with no statistically significant difference between the scores (p > 0.05). The calibration of the scores was good.

CONCLUSIONS

All the scores are acceptable predictors of in-hospital mortality. In the case of post-discharge mortality, their diagnostic accuracy is lower and of borderline clinical relevance. Further studies are needed to create scores estimating the long-term prognosis of subjects successfully discharged from the ICU.

The SAPS 3 score as a predictor of hospital mortality in a South African tertiary intensive care unit: A prospective cohort study

BACKGROUND

No African countries were included in the development of the Simplified Acute Physiology Score 3 (SAPS 3). This study aimed to assess the performance of the SAPS 3 as a predictor of hospital mortality in patients admitted to a multi-disciplinary tertiary intensive care unit (ICU) in South Africa.

METHODS

A prospective cohort study was undertaken in a tertiary single-centre closed multidisciplinary ICU with 16 beds over 12 months in 2017. First time admissions 12 years and over were included. Exclusions were patients who died within six hours of admission, incomplete data sets and unknown outcome after ICU discharge. Demographic data, clinical admission data and co-morbidities were recorded. The SAPS 3 score was calculated within the first hour of ICU admission. The highest Sequential Organ Failure Assessment score, vasopressor use, mechanical ventilation requirements and details of acute kidney injury, if present, were recorded. Discrimination of the model was evaluated using an area under the receiver operating characteristic curve (AUROC) and calibration by the Hosmer-Lemeshow (HL) Goodness of Fit Test (Ĉ and Ĥ statistic). The observed versus the SAPS 3 model predicted mortality ratios were compared and the standardized mortality ratio (SMR) was calculated.

RESULTS

A total of 829 admissions with a mean SAPS 3 (SD) of 48.1 (16) were included. Of patients with a known human immunodeficiency virus (HIV) status, 32,4% were positive. The ICU and hospital mortality rates were 13.3% and 21.4% respectively. The SAPS 3 model had a AUROC of 0.796 and HL Ĉ and Ĥ statistics were 12.1 and 11.8 (p-values 0.15 and 0.16). The SMR for the model was 1.002 (95%CI: 0.91-1.10). The mortality of 41% for the subgroup with sepsis/septic shock was higher than predicted with a SMR of 1.24 (95% CI 1.11-1.37).

CONCLUSIONS

The SAPS 3 model showed good calibration and fair discrimination when applied to the cohort. The SAPS 3 model can be used to describe the case mix in this African ICU with a high incidence of HIV. Ongoing efforts should be made to improve outcomes of septic patients.

Perioperative factors impacting intensive care outcomes following Whipple procedure: A retrospective study

Background and Aims

Whipple procedure is associated with perhaps the most perioperative morbidity and mortality amongst surgical procedures. Current data regarding their ICU profile and outcomes are lacking. Thus, in the present study, we aimed to determine perioperative factors affecting patient-centred outcomes following the Whipple procedure.

Methods

In a cohort of patients undergoing pylorus-sparing pancreaticoduodenectomies, we strove to determine perioperative variables that may impact outcomes. Unfavourable outcomes (composite of mortality, prolonged ICU stay of more than 14 days or ICU readmission) of patients who underwent the procedure were recorded and logistic regressions analysis of significant variables conducted.

Results

Around 68 patients recruited over a 20-month period which included 57 males (83.8%); mean age was 53.4(±11.2) with mean acute physiology and chronic health evaluation (APACHE) II score12.5 (±6.1). Nineteen patients remained intubated at the end of procedures (27.9%). Median ICU stay was 2 days (IQR 2-3). Unfavourable ICU outcomes were 14 in number (20.6%) and 2 (2.9%) hospital deaths occurred. Pulmonary complications occurred in 12 patients (17.7%) and non-pulmonary complications occurred in 41 patients (60.3%). In a multiple logistic regression analysis, the APACHE score 1.34 (1.09-1.64) and pulmonary complications 17.3 (2.1-145) were variables that were identified as predictors of unfavourable outcomes.

Conclusion

The APACHE II score may reliably predict adverse outcomes following Whipple procedure. Although non-pulmonary complications are common, pulmonary complications in these patients adversely impact patient outcomes.

Comparison of APACHE IV with APACHE II, SAPS 3, MELD, MELD-Na, and CTP scores in predicting mortality after liver transplantation

The Acute Physiology and Chronic Health Evaluation (APACHE) IV score and Simplified Acute Physiology Score (SAPS) 3 include liver transplantation as a diagnostic category. The performance of APACHE IV-liver transplantation (LT) specific predicted mortality, SAPS 3, APACHE II, Model for End-stage Liver Disease (MELD)-Na, MELD, and CTP scores in predicting in-hospital and 1 year mortality in liver transplant patients was compared using 590 liver transplantations in a single university hospital. In-hospital mortality and 1 year mortality were 2.9% and 4.2%, respectively. The APACHE IV-LT specific predicted mortality showed better performance in predicting in-hospital mortality (AUC 0.91, 95% CI [0.86–0.96]) compared to SAPS 3 (AUC 0.78, 95% CI [0.66–0.90], p = 0.01), MELD-Na (AUC 0.74, 95% CI [0.57–0.86], p = 0.01), and CTP (AUC 0.68, 95% CI [0.54–0.81], p = 0.01). The APACHE IV-LT specific predicted mortality showed better performance in predicting 1 year mortality (AUC 0.83, 95% CI [0.76–0.9]) compared to MELD-Na (AUC 0.67, 95% CI [0.55–0.79], p = 0.04) and CTP (AUC 0.64, 95% CI [0.53–0.75], p = 0.03), and also in all MELD groups and in both living and deceased donor transplantation. The APACHE IV-LT specific predicted mortality showed better performance in predicting in-hospital and 1 year mortality after liver transplantation.

Urinary cadmium levels predict mortality of patients with acute heart failure

BACKGROUND

Acute heart failure (AHF) is a serious condition that is associated with increased mortality in critically ill patients. Previous studies indicated that environmental exposure to cadmium increases mortality of general populations. However, the relationship of cadmium exposure and mortality is unclear for AHF patients.

MATERIALS AND METHODS

A total of 153 patients with AHF in intensive care units (ICUs) met the inclusion criteria and were followed up for 6 months. Demographic data, AHF etiology, hematological and biochemical data, and hospital mortality were recorded. The scores of two predictive systems (Sequential Organ Failure Assessment [SOFA], Acute Physiology and Chronic Health Evaluation II [APACHE II]) for mortality in critically ill patients were calculated, and urinary cadmium levels were recorded.

RESULTS

At the end of the follow-up period, the mortality rate was 24.8%. The survivors (n=115) had higher urinary cadmium levels on day 1 (D1UCd) of ICU admission than non-survivors (n=38). A multiple linear regression analysis revealed a positive correlation between D1UCd and acute kidney injury, but a negative correlation between D1UCd and the level of serum albumin. A multivariate Cox analysis indicated that D1UCd was an independent predictor of mortality in AHF patients. For each increment of 1 μg of D1UCd, the hazard ratio for ICU mortality was 1.20 (95% confidence interval [CI]: 1.09-1.32, P<0.001). The area under the receiver operating characteristic curve for D1UCd was 0.84 (95% CI: 0.78-0.91), better than the values for the SOFA and APACHE II systems.

CONCLUSION

The D1UCd may serve as a single predictor of hospital mortality for AHF patients in the ICU. Because of the high mortality and smaller sample size, more investigations are required to confirm these observations and elucidate the underlying mechanisms.

Predicting AKI in emergency admissions: an external validation study of the acute kidney injury prediction score (APS)

OBJECTIVES

Hospital-acquired acute kidney injury (HA-AKI) is associated with a high risk of mortality. Prediction models or rules may identify those most at risk of HA-AKI. This study externally validated one of the few clinical prediction rules (CPRs) derived in a general medicine cohort using clinical information and data from an acute hospitals electronic system on admission: the acute kidney injury prediction score (APS).

DESIGN, SETTING AND PARTICIPANTS

External validation in a single UK non-specialist acute hospital (2013-2015, 12 554 episodes); four cohorts: adult medical and general surgical populations, with and without a known preadmission baseline serum creatinine (SCr).

METHODS

Performance assessed by discrimination using area under the receiver operating characteristic curves (AUCROC) and calibration.

RESULTS

HA-AKI incidence within 7 days (kidney disease: improving global outcomes (KDIGO) change in SCr) was 8.1% (n=409) of medical patients with known baseline SCr, 6.6% (n=141) in those without a baseline, 4.9% (n=204) in surgical patients with baseline and 4% (n=49) in those without. Across the four cohorts AUCROC were: medical with known baseline 0.65 (95% CIs 0.62 to 0.67) and no baseline 0.71 (0.67 to 0.75), surgical with baseline 0.66 (0.62 to 0.70) and no baseline 0.68 (0.58 to 0.75). For calibration, in medicine and surgical cohorts with baseline SCr, Hosmer-Lemeshow p values were non-significant, suggesting acceptable calibration. In the medical cohort, at a cut-off of five points on the APS to predict HA-AKI, positive predictive value was 16% (13-18%) and negative predictive value 94% (93-94%). Of medical patients with HA-AKI, those with an APS ≥5 had a significantly increased risk of death (28% vs 18%, OR 1.8 (95% CI 1.1 to 2.9), p=0.015).

CONCLUSIONS

On external validation the APS on admission shows moderate discrimination and acceptable calibration to predict HA-AKI and may be useful as a severity marker when HA-AKI occurs. Harnessing linked data from primary care may be one way to achieve more accurate risk prediction.

Association of Urinary Cadmium with Mortality in Patients at a Coronary Care Unit

BACKGROUND

Determine the effect of the day 1 urinary excretion of cadmium (D1-UE-Cd) on mortality of patients admitted to a coronary care unit (CCU).

METHODS

A total of 323 patients were enrolled in this 6-month study. Urine and blood samples were taken within 24 h after CCU admission. Demographic data, clinical diagnoses, and hospital mortality were recorded. The scores of established systems for prediction of mortality in critically ill patients were calculated.

RESULTS

Compared with survivors (n = 289), non-survivors (n = 34) had higher levels of D1-UE-Cd. Stepwise multiple linear regression analysis indicated that D1-UE-Cd was positively associated with pulse rate and level of aspartate aminotransferase, but negatively associated with serum albumin level. Multivariate Cox analysis, with adjustment for other significant variables and measurements from mortality scoring systems, indicated that respiratory rate and D1-UE-Cd were independent and significant predictors of mortality. For each 1 μg/day increase of D1-UE-Cd, the hazard ratio for CCU mortality was 3.160 (95% confidence interval: 1.944-5.136, p < 0.001). The chi-square value of Hosmer-Lemeshow goodness-of-fit test for D1-UE-Cd was 10.869 (p = 0.213). The area under the receiver operating characteristic curve for D1-UE-Cd was 0.87 (95% confidence interval: 0.81-0.93).

CONCLUSIONS

The D1-UE-Cd, an objective variable with no inter-observer variability, accurately predicted hospital mortality of CCU patients and outperformed other established scoring systems. Further studies are needed to determine the physiological mechanism of the effect of cadmium on mortality in CCU patients.

The significance of regional hemoglobin oxygen saturation values and limb-to-arm ratios of near-infrared spectroscopy to detect critical limb ischemia

This study examines the application of near-infrared spectroscopy to noninvasively detect critical limb ischemia using regional hemoglobin oxygen saturation in percentage values and regional hemoglobin oxygen saturation limb-to-arm ratios. The regional hemoglobin oxygen saturation values and regional hemoglobin oxygen saturation limb-to-arm ratios were calculated in 61 patients with critical limb ischemia (group A). Measurements were performed in rest at four fixed spots at the most affected lower limb and at a reference spot at both upper arms. Similar measurements were performed in the left lower limb of 30 age-matched control patients without peripheral arterial disease (group B). The regional hemoglobin oxygen saturation values and regional hemoglobin oxygen saturation limb-to-arm ratios were significantly different at all measured spots between the groups (all p < 0.001), except for the regional hemoglobin oxygen saturation limb-to-arm ratios of the distal vastus lateralis (p = 0.056). However, a broad overlap of individual regional hemoglobin oxygen saturation values and regional hemoglobin oxygen saturation limb-to-arm ratios was found in both groups, which resulted in poor discriminative predictive value of single measurements. Single measurements of regional hemoglobin oxygen saturation values and regional hemoglobin oxygen saturation limb-to-arm ratios at all measured spots have poor discriminative predictive value in detection of critical limb ischemia. Measurement of regional hemoglobin oxygen saturation values and regional hemoglobin oxygen saturation limb-to-arm ratios at any of the measurement spots has no added value in detecting lower limb ischemia in individuals compared with current diagnostic modalities.

Predicting reintubation, prolonged mechanical ventilation and death in post-coronary artery bypass graft surgery: a comparison between artificial neural networks and logistic regression models

INTRODUCTION

In coronary artery bypass (CABG) surgery, the common complications are the need for reintubation, prolonged mechanical ventilation (PMV) and death. Thus, a reliable model for the prognostic evaluation of those particular outcomes is a worthwhile pursuit. The existence of such a system would lead to better resource planning, cost reductions and an increased ability to guide preventive strategies. The aim of this study was to compare different methods - logistic regression (LR) and artificial neural networks (ANNs) - in accomplishing this goal.

MATERIAL AND METHODS

Subjects undergoing CABG (n = 1315) were divided into training (n = 1053) and validation (n = 262) groups. The set of independent variables consisted of age, gender, weight, height, body mass index, diabetes, creatinine level, cardiopulmonary bypass, presence of preserved ventricular function, moderate and severe ventricular dysfunction and total number of grafts. The PMV was also an input for the prediction of death. The ability of ANN to discriminate outcomes was assessed using receiver-operating characteristic (ROC) analysis and the results were compared using a multivariate LR.

RESULTS

The ROC curve areas for LR and ANN models, respectively, were: for reintubation 0.62 (CI: 0.50-0.75) and 0.65 (CI: 0.53-0.77); for PMV 0.67 (CI: 0.57-0.78) and 0.72 (CI: 0.64-0.81); and for death 0.86 (CI: 0.79-0.93) and 0.85 (CI: 0.80-0.91). No differences were observed between models.

CONCLUSIONS

The ANN has similar discriminating power in predicting reintubation, PMV and death outcomes. Thus, both models may be applicable as a predictor for these outcomes in subjects undergoing CABG.

ICU director data: using data to assess value, inform local change, and relate to the external world

Improving value within critical care remains a priority because it represents a significant portion of health-care spending, faces high rates of adverse events, and inconsistently delivers evidence-based practices. ICU directors are increasingly required to understand all aspects of the value provided by their units to inform local improvement efforts and relate effectively to external parties. A clear understanding of the overall process of measuring quality and value as well as the strengths, limitations, and potential application of individual metrics is critical to supporting this charge. In this review, we provide a conceptual framework for understanding value metrics, describe an approach to developing a value measurement program, and summarize common metrics to characterize ICU value. We first summarize how ICU value can be represented as a function of outcomes and costs. We expand this equation and relate it to both the classic structure-process-outcome framework for quality assessment and the Institute of Medicine's six aims of health care. We then describe how ICU leaders can develop their own value measurement process by identifying target areas, selecting appropriate measures, acquiring the necessary data, analyzing the data, and disseminating the findings. Within this measurement process, we summarize common metrics that can be used to characterize ICU value. As health care, in general, and critical care, in particular, changes and data become more available, it is increasingly important for ICU leaders to understand how to effectively acquire, evaluate, and apply data to improve the value of care provided to patients.

Is SAPS 3 better than APACHE II at predicting mortality in critically ill transplant patients?

OBJECTIVES

This study compared the accuracy of the Simplified Acute Physiology Score 3 with that of Acute Physiology and Chronic Health Evaluation II at predicting hospital mortality in patients from a transplant intensive care unit.

METHOD

A total of 501 patients were enrolled in the study (152 liver transplants, 271 kidney transplants, 54 lung transplants, 24 kidney-pancreas transplants) between May 2006 and January 2007. The Simplified Acute Physiology Score 3 was calculated using the global equation (customized for South America) and the Acute Physiology and Chronic Health Evaluation II score; the scores were calculated within 24 hours of admission. A receiver-operating characteristic curve was generated, and the area under the receiver-operating characteristic curve was calculated to identify the patients at the greatest risk of death according to Simplified Acute Physiology Score 3 and Acute Physiology and Chronic Health Evaluation II scores. The Hosmer-Lemeshow goodness-of-fit test was used for statistically significant results and indicated a difference in performance over deciles. The standardized mortality ratio was used to estimate the overall model performance.

RESULTS

The ability of both scores to predict hospital mortality was poor in the liver and renal transplant groups and average in the lung transplant group (area under the receiver-operating characteristic curve = 0.696 for Simplified Acute Physiology Score 3 and 0.670 for Acute Physiology and Chronic Health Evaluation II). The calibration of both scores was poor, even after customizing the Simplified Acute Physiology Score 3 score for South America.

CONCLUSIONS

The low predictive accuracy of the Simplified Acute Physiology Score 3 and Acute Physiology and Chronic Health Evaluation II scores does not warrant the use of these scores in critically ill transplant patients.

Sluggish decline in a post-transplant model for end-stage liver disease score is a predictor of mortality in living donor liver transplantation

Background

The pre-transplant model for end-stage liver disease (pre-MELD) score is controversial regarding its ability to predict patient mortality after liver transplantation (LT). Prominent changes in physical conditions through the surgery may require a post-transplant indicator for better mortality prediction. We aimed to investigate whether the post-transplant MELD (post-MELD) score can be a predictor of 1-year mortality.

Methods

Perioperative variables of 269 patients with living donor LT were retrospectively investigated on their association with 1-year mortality. Post-MELD scores until the 30th day and their respective declines from the 1st day post-MELD score were included along with pre-MELD, acute physiology and chronic health evaluation (APACHE) II, and sequential organ failure assessment (SOFA) scores on the 1st post-transplant day. The predictive model of mortality was established by multivariate Cox's proportional hazards regression.

Results

The 1-year mortality rate was 17% (n = 44), and the leading cause of death was graft failure. Among prognostic indicators, only post-MELD scores after the 5th day and declines in post-MELD scores until the 5th and 30th day were associated with mortality in univariate analyses (P < 0.05). After multivariate analyses, declines in post-MELD scores until the 5th day of less than 5 points (hazard ratio 2.35, P = 0.007) and prolonged mechanical ventilation ≥24 hours were the earliest independent predictors of 1-year mortality.

Conclusions

A sluggish decline in post-MELD scores during the early post-transplant period may be a meaningful prognostic indicator of 1-year mortality after LT.

Looking forward, looking back: assessing variations in hospital resource use and outcomes for elderly patients with heart failure

BACKGROUND

Recent studies have found substantial variation in hospital resource use by expired Medicare beneficiaries with chronic illnesses. By analyzing only expired patients, these studies cannot identify differences across hospitals in health outcomes like mortality. This study examines the association between mortality and resource use at the hospital level, when all Medicare beneficiaries hospitalized for heart failure are examined.

METHODS AND RESULTS

A total of 3999 individuals hospitalized with a principal diagnosis of heart failure at 6 California teaching hospitals between January 1, 2001, and June 30, 2005, were analyzed with multivariate risk-adjustment models for total hospital days, total hospital direct costs, and mortality within 180-days after initial admission ("Looking Forward"). A subset of 1639 individuals who died during the study period were analyzed with multivariate risk-adjustment models for total hospital days and total hospital direct costs within 180-days before death ("Looking Back"). "Looking Forward" risk-adjusted hospital means ranged from 17.0% to 26.0% for mortality, 7.8 to 14.9 days for total hospital days, and 0.66 to 1.30 times the mean value for indexed total direct costs. Spearman rank correlation coefficients were -0.68 between mortality and hospital days, and -0.93 between mortality and indexed total direct costs. "Looking Back" risk-adjusted hospital means ranged from 9.1 to 21.7 days for total hospital days and 0.91 to 1.79 times the mean value for indexed total direct costs. Variation in resource use site ranks between expired and all individuals were attributable to insignificant differences.

CONCLUSIONS

California teaching hospitals that used more resources caring for patients hospitalized for heart failure had lower mortality rates. Focusing only on expired individuals may overlook mortality variation as well as associations between greater resource use and lower mortality. Reporting values without identifying significant differences may result in incorrect assumption of true differences.

APACHE III outcome prediction in patients admitted to the intensive care unit after liver transplantation: a retrospective cohort study

Background

The Acute Physiology and Chronic Health Evaluation (APACHE) III prognostic system has not been previously validated in patients admitted to the intensive care unit (ICU) after orthotopic liver transplantation (OLT). We hypothesized that APACHE III would perform satisfactorily in patients after OLT

Methods

A retrospective cohort study was performed. Patients admitted to the ICU after OLT between July 1996 and May 2008 were identified. Data were abstracted from the institutional APACHE III and liver transplantation databases and individual patient medical records. Standardized mortality ratios (with 95% confidence intervals) were calculated by dividing the observed mortality rates by the rates predicted by APACHE III. The area under the receiver operating characteristic curve (AUC) and the Hosmer-Lemeshow C statistic were used to assess, respectively, discrimination and calibration of APACHE III.

Results

APACHE III data were available for 918 admissions after OLT. Mean (standard deviation [SD]) APACHE III (APIII) and Acute Physiology (APS) scores on the day of transplant were 60.5 (25.8) and 50.8 (23.6), respectively. Mean (SD) predicted ICU and hospital mortality rates were 7.3% (15.4) and 10.6% (18.9), respectively. The observed ICU and hospital mortality rates were 1.1% and 3.4%, respectively. The standardized ICU and hospital mortality ratios with their 95% C.I. were 0.15 (0.07 to 0.27) and 0.32 (0.22 to 0.45), respectively.

There were statistically significant differences in APS, APIII, predicted ICU and predicted hospital mortality between survivors and non-survivors. In predicting mortality, the AUC of APACHE III prediction of hospital death was 0.65 (95% CI, 0.62 to 0.68). The Hosmer-Lemeshow C statistic was 5.288 with a p value of 0.871 (10 degrees of freedom).

Conclusion

APACHE III discriminates poorly between survivors and non-survivors of patients admitted to the ICU after OLT. Though APACHE III has been shown to be valid in heterogenous populations and in certain groups of patients with specific diagnoses, it should be used with caution – if used at all – in recipients of liver transplantation.

[POSSUM scoring system for predicting mortality in surgical patients]

This study evaluated the use of the POSSUM (Physiological and Operative Severity Score for Enumeration of Mortality and Morbidity) score for predicting mortality in surgical practice. In this study, 416 surgical patients admitted into ICUs for post-surgical care were analyzed. Both predicted and actual mortality rates were compared, according to four risk groups: 0-4%, 5-14%, 15-49%, 50% and over, and the area under the ROC curve of the POSSUM and APACHE II for mortality. The POSSUM and APACHE II scores overestimated the risk of death. The area under the ROC curve of the POSSUM was 0.762, and under APACHE II was 0.737, suggesting the use of POSSUM as an auxiliary tool to predict the risk of death in surgical patients.

Cadmium excretion predicting hospital mortality and illness severity of critically ill medical patients

OBJECTIVE

To determine the prognostic value of day 1 urine excretion of cadmium (1st DUE-Cd) for predicting outcomes in intensive care unit (ICU) patients.

DESIGN

Prospective study.

SETTING

ICUs in Chang Gung Memorial Hospital, Lin-Kou Medical Center, Taiwan, ROC.

PATIENTS

Two hundred one ICU patients.

INTERVENTIONS

Urine and blood samples were taken within 24 hours after admission.

MEASUREMENTS AND MAIN RESULTS

Disease severity, hospital mortality, and number of organ failures were evaluated in each medical ICU patient. Stepwise multiple linear regression analysis indicated that a history of chronic hepatitis, serum albumin, and glutamic-pyruvic transaminase were significantly related to 1st DUE-Cd after adjusting for other related variables. Cox multivariate analysis revealed that serum blood urea nitrogen level and ICU 1st DUE-Cd were significantly related to hospital mortality after other risk factors and scoring systems were adjusted. Each 1-microg increase in ICU 1st DUE-Cd was associated with a 7% increase in hospital mortality rate. All patients with poisoning magnitude of cadmium excretion (>10 microg/day) died, except one and those with normal cadmium excretion survived. Chi-square values of the Hosmer-Lemeshow goodness-of-fit test were 6.936 (p = 0.544), and area under the receiver operating characteristic curve was 0.868 (95% confidence intervals: 0.82-0.92) for ICU 1st DUE-Cd.

CONCLUSIONS

The ICU 1st DUE-Cd may predict hospital mortality in critically ill medical patients. Because of excess mortality and relatively small sample size, the predictive role of DUE-Cd needs further external validation.

Acute Physiology and Chronic Health Evaluation (APACHE) III outcome prediction after major vascular surgery

OBJECTIVE

To investigate the performance of the Acute Physiology and Chronic Health Evaluation (APACHE) III scoring system in patients admitted to the intensive care unit (ICU) after major vascular surgery.

DESIGN

Retrospective cohort study.

SETTING

A tertiary referral center.

PARTICIPANTS

Three thousand one hundred forty-eight patients who underwent major vascular surgery between October 1994 and March 2006.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

Data were abstracted from an institutional APACHE III database. Standardized mortality ratios (SMRs) (with 95% confidence intervals) were calculated. The area under the receiver operating characteristic curve (AUC) and Hosmer-Lemeshow C statistic were used to assess discrimination and calibration, respectively. The mean age of 3,148 patients studied was 70.5 years (+/- standard deviation 9.6). The mean Acute Physiology Score and the APACHE III score on the day of ICU admission were 31.0 (+/- 17.5) and 45.1 (+/- 18.8), respectively. The mean predicted ICU and hospital mortality rates were 3.2% (+/- 7.8%) and 5.0% (+/- 9.5%), respectively. The median (and interquartile range) ICU and hospital lengths of stay were 4.3 (3.6-5.1) and 14 days (11.9-16.8 days), respectively. The observed ICU mortality rate was 2.4% (75/3, 148 patients) and hospital mortality rate was 3.7% (116/3,148). The ICU and hospital SMRs were 0.74 (0.58-0.91) and 0.74 (0.61-0.88), respectively. The AUC of APACHE III-derived prediction of hospital mortality was 0.840 (95% confidence interval, 0.799-0.880), indicating excellent discrimination. The Hosmer-Lemeshow C statistic was 28.492, with a p value <0.01, indicating poor calibration.

CONCLUSIONS

The APACHE III scoring system discriminates well between survivors and nonsurvivors after major vascular surgery, but calibration of the model is poor.

The acute physiology and chronic health evaluation III outcome prediction in patients admitted to the intensive care unit after pneumonectomy

PURPOSE

The Acute Physiology and Chronic Health Evaluation (APACHE) III prognostic system has not been previously validated in patients admitted to the intensive care unit (ICU) after pneumonectomy. The purpose of this study was to determine if the APACHE III predicts hospital mortality after pneumonectomy.

METHODS

A retrospective review of all adult patients admitted to a single thoracic surgical intensive care unit after pneumonectomy between October 1994 and December 2004. Patient demographics, ICU admission day APACHE III score, actual and predicted hospital mortality, and length of hospital and ICU stay data were collected. Data on preoperative pulmonary function tests and smoking habits were also collected. Univariate statistical methods and logistic regression were used. The performance of the APACHE III prognostic system was assessed by the Hosmer-Lemeshow statistic for calibration and area under receiver operating characteristic curve (AUC) for discrimination.

RESULTS

There were 417 pneumonectomies performed during the study period, of which 281 patients were admitted to the ICU. The mean age was 61.1 years, and 67.2% were men; 88.2% were smokers with a median of 40.0 (interquartile range, 18-62) pack-years of tobacco use. The mean APACHE III score on the day of ICU admission was 37.7 (+/- standard deviation 17.8), and the mean predicted hospital mortality rate was 6.4% (+/-10.4). The median (and interquartile range) lengths of ICU and hospital stay were 1.7 (0.9-3.1) and 9.0 (7.0-17.0) days, respectively. The observed ICU and hospital mortality rates were 4.6% (13/281 patients) and 8.2% (23/281), respectively. The standardized ICU and hospital mortality ratios with their 95% confidence intervals (CIs) were 1.55 (0.71-2.39) and 1.27 (0.75-1.78), respectively. There were significant differences in the mean APACHE III score (p < 0.001) and the predicted mortality rate (p < .001) between survivors and nonsurvivors. In predicting mortality, the AUC of APACHE III prediction was 0.801 (95% CI, 0.711-0.891), and the Hosmer-Lemeshow statistic was 9.898 with a p value of 0.272. Diffusion capacity of the lung for carbon monoxide (DLCO) and percentage predicted DLCO were higher in survivors, but the addition of either of these variables to a logistic regression model did not improve APACHE III mortality prediction.

CONCLUSIONS

In patients admitted to the ICU after pneumonectomy, the APACHE III discriminates moderately well between survivors and nonsurvivors. The calibration of the model appears to be good, although the low number of deaths limits the power of the calibration analysis. The use of APACHE III data in outcomes research involving patients who have undergone pneumonectomy is acceptable.

Cost of Gram-negative resistance*

OBJECTIVE

It is unclear that infections with Gram-negative rods resistant to at least one major class of antibiotics (rGNR) have a greater effect on patient morbidity than infections caused by sensitive strains (sGNR). We wished to test the hypothesis that rGNR infections are associated with higher resource utilization.

DESIGN

Retrospective observational cohort study of prospectively collected data.

SETTING

University hospital surgical intensive care unit and ward.

PATIENTS

Surgical patients with at least one GNR infection.

MEASUREMENTS

We compared admissions treated for rGNR infection with those with sGNR infections. Primary outcomes were total hospital costs and hospital length of stay. Other outcomes included antibiotic treatment cost, in-hospital death, and intensive care unit length of stay. After univariate analysis comparing outcomes after rGNR infection with those after sGNR infection, multivariate linear regression models for hospital cost and length of stay were created to account for potential confounders.

MAIN RESULTS

Cost data were available for 604 surgical admissions treated for at least one GNR infection (Centers for Disease Control and Prevention criteria), 137 (23%) of which were rGNR infections. Admissions with rGNR infections were associated with a higher severity of illness at the time of infection (Acute Physiology and Chronic Health Evaluation II score, 17.6 +/- 0.6 vs. 13.9 +/- 0.3), had higher median hospital costs ($80,500 vs. $29,604, p < .0001) and median antibiotic costs ($2,607 vs. $758, p < .0001), and had longer median hospital length of stay (29 vs. 13 days, p < .0001) and median intensive care unit length of stay (13 days vs. 1 day, p < .0001). Infection with rGNR within the first 7 days of admission was independently predictive of increased hospital cost (incremental increase in median hospital cost estimated at $11,075; 95% confidence interval, $3,282-$20,099).

CONCLUSIONS

Early infection with rGNR is associated with a high economic burden, which is in part related to increased antibiotic utilization compared with infection with sensitive organisms. Efforts to control overuse of antibiotics should be pursued.

Readmission to Surgical Intensive Care Increases Severity-Adjusted Patient Mortality

BACKGROUND

This study aims to determine whether severity-adjusted outcomes including mortality are adversely impacted by readmission to a surgical intensive care unit (SICU) during the same hospital stay.

METHODS

The study included all patients admitted to the 20-bed tertiary care SICU in an urban teaching Level I trauma center and multiorgan transplant center from January 1, 1996 to December 31, 2001. This was a prospective observational study with secondary data analysis. Acute Physiology and Chronic Health Evaluation (APACHE II) and Simplified Acute Physiology (SAPS) severity scores were calculated by a clinical information system. Outcomes were extracted from a computerized data warehouse.

RESULTS

In-hospital mortality and SICU length of stay (LOS) were measured for patients admitted and readmitted to the SICU. Of 10,840 patients admitted to the SICU, 296 (2.73%) required readmission to the SICU during the same hospital stay. The length of the original SICU stay was 4.9 +/- 6.7 days for readmitted patients compared with 3.2 +/- 6.0 days for nonreadmitted patients (p < 0.001). Readmitted patients had a higher mean APACHE II score on the day of original SICU discharge compared with nonreadmitted patients, 15.7 +/- 6.7 versus 13.8 +/- 7.1 (p < 0.001). The average APACHE II score increased from 15.7 +/- 6.7 to 18.1 +/- 8.6 between the day of SICU discharge and readmission (p < 0.001) and SAPS increased from 12.2 +/- 4.8 to 13.5 +/- 5.4 (p < 0.001). The distributions of severity-adjusted hospital mortality for both APACHE II and SAPS revealed that readmission to the SICU significantly increased mortality independent of the admission severity score.

CONCLUSIONS

Readmission to the SICU significantly increases the risk of death beyond that predicted by the APACHE II or SAPS scores alone. Higher APACHE II and SAPS scores upon discharge from the SICU and longer SICU LOS are associated with an increased incidence of readmission to the SICU on the same hospital stay. These results may be used to optimize the timing of SICU discharge and reduce the chance of readmission to intensive care.

Prolonged intubation rates after coronary artery bypass surgery and ICU risk stratification score

OBJECTIVE

To determine prolonged intubation rates among patients undergoing coronary artery bypass graft (CABG) surgery, and to evaluate the ability of the Intensive Care Unit Risk Stratification Score (ICURSS) model to predict these events.

DESIGN

Prospective observational study.

SETTING

A 24-bed ICU in a tertiary referral university hospital.

PATIENTS

Five hundred sixty-nine patients undergoing CABG surgery.

INTERVENTIONS

Variables of the ICURSS model were recorded at ICU admission. Extubation was performed according to a standard protocol. Patients remaining intubated within 8 h after ICU admission were designated as having early extubation failure (EEF). The next evaluations at 16, 24, 48, 72, and 96 h designated patients as having a prolonged intubation period (PIP) and prolonged mechanical ventilation (PMV) for 24, 48, 72, and 96 h. The ability of the ICURSS model to predict extubation failure at different cutoff values was measured using the Hosmer-Lemeshow goodness-of-fit test and the area under the receiver operating characteristic curve.

MEASUREMENTS AND RESULTS

Prolonged intubation rates were as follows: EEF, 40.2%; PIP, 17.2%; PMV for 24 h, 10.4%; PMV for 48 h, 7.6%; PMV for 72 h, 6.5%; and PMV for 96 h, 6.0%. At every cutoff, the ICURSS showed poor discrimination to predict the failure to be extubated. Calibration was also poor, although some ability to predict both EEF and PMV at > or = 48 h was shown.

CONCLUSIONS

Prolonged intubation rates after undergoing CABG surgery in our setting were comparable with those of other reports from institutions where fast-track cardiac anesthesia is currently in practice. In our cohort, the ICURSS was not useful for the prediction of length of intubation.

Outcome of patients with end-stage renal disease admitted to the intensive care unit

OBJECTIVES

To describe the clinical course of patients with end-stage renal disease (ESRD) admitted to the intensive care unit (ICU) and to compare the performance of Acute Physiology and Chronic Health Evaluation (APACHE) III and Sequential Organ Failure Assessment (SOFA) in predicting their outcome.

PATIENTS AND METHODS

This retrospective cohort study consisted of patients with ESRD admitted to 3 ICUs between January 1, 1997, and November 30, 2002. Data on demographics, APACHE III score, SOFA score, development of sepsis and organ failure, use of mechanical ventilation, and mortality were collected.

RESULTS

Of the 476 patients with ESRD who underwent dialysis during the study period, 93 (20%) required admission to the ICU. The most common ICU admission diagnosis was gastrointestinal bleeding. The first day median (Interquartile range) APACHE III score, SOFA score, and APACHE III predicted hospital mortality rate were 64 (47-79), 6 (5-8), and 12.9% (4.2%-30.8%), respectively. The observed ICU, hospital, and 30-day mortality rates were 9%, 16%, and 22%, respectively. Nonrenal organ failure developed in 48 patients (52%) and sepsis in 15 patients (16%). Mechanical ventilation was required In 26 patients (28%). The area under the receiver operating characteristic curve for the first-day APACHE III probability of hospital death in predicting 30-day mortality was 0.78 (95% confidence interval, 0.68-0.86) compared with 0.66 (95% confidence interval, 0.55-0.76) for the SOFA score (P = .16).

CONCLUSIONS

The observed hospital mortality of patients with ESRD admitted to the ICU is relatively low. There is no statistically significant difference in the performance of APACHE III and SOFA prognostic models in discriminating between 30-day survivors and nonsurvivors.

How can clinicians measure safety and quality in acute care?

The demand for high quality care is increasing and warranted. Evidence suggests that the quality of care in hospitals can be improved. The greatest opportunity to improve outcomes for patients over the next quarter century will probably come not from discovering new treatments but from learning how to deliver existing effective therapies. To improve, caregivers need to know what to do, how they are doing, and be able to improve the processes of care. The ability to monitor performance, though challenging in healthcare, is essential to improving quality of care. We present a practical method to assess and learn from routine practice. Methods to evaluate performance from industrial engineering can be broadly applied to efforts to improve the quality of healthcare. One method that may help to provide caregivers frequent feedback is time series data--ie, results are graphically correlated with time. Broad use of these tools might lead to the necessary improvements in quality of care.

Developing and pilot testing quality indicators in the intensive care unit

PURPOSE

To develop and implement a set of valid and reliable yet practical measures of intensive care units (ICU) quality of care in a cohort of ICUs and to estimate, based on current performance, the potential opportunity to improve quality.

METHODS

We included 13 adult medical and surgical ICUs in urban community teaching and community hospitals. To monitor performance on previously identified quality measures, we developed 3 data collection tools: the Team Leader, Daily Rounding, and Infection Control forms. These tools were pilot tested, validated, and modified before implementation. We used published estimates of efficacy to estimate the clinical and economic effect of our current performance for each of the process measures: appropriate sedation, prevention of ventilator-associated pneumonia, appropriate peptic ulcer disease (PUD) prophylaxis, appropriate deep venous thrombosis (DVT) prophylaxis, and appropriate use of blood transfusions.

RESULTS

Performance varied widely among the 13 ICUs and within ICUs. The median percentage of days in which ventilated patients received therapies that ought to was 64% for appropriate sedation, 67% for elevating head of bed, 89% for PUD prophylaxis, and 87% for DVT prophylaxis. The median rate of appropriate transfusion was 33%. The failure to use these therapies may lead to excess morbidity, mortality, and ICU length of stay.

CONCLUSION

To improve quality of care, we must measure our performance. This pilot study suggests that it is feasible to implement a broad set of ICU quality measures in a cohort of hospitals. By improving performance on these measures, we may realize reduced mortality, morbidity, and ICU length of stay.

Intensive care unit support and Acute Physiology and Chronic Health Evaluation III performance in hematopoietic stem cell transplant recipients

OBJECTIVE

Hematopoietic stem cell transplant (HSCT) recipients admitted to the intensive care unit (ICU) have high mortality. The prognostic importance of peripheral blood stem cell source in critically ill HSCT recipients and the performance of Acute Physiology and Chronic Health Evaluation (APACHE) III have not been well studied. In a previous study, the hospital mortality rate of HSCT recipients admitted to our ICU was 77%. The objectives of this study were to describe the clinical course of HSCT recipients admitted to the ICU and to determine the performance of APACHE III in predicting their mortality.

DESIGN

Retrospective cohort study.

SETTING

Academic medical center.

PATIENTS

HSCT recipients admitted to the ICU.

MEASUREMENTS

Demographics, transplant type, stem cell source, APACHE II and III predicted mortality, development of sepsis and organ failure, use of mechanical ventilation, duration of hospital stay, and mortality.

RESULTS

Ninety-four percent of the 112 HSCT recipients were white and 64% male. The mean APACHE II and III scores were 25 and 44, respectively. The APACHE II and III hospital predicted mortality rates were 44% and 42%, respectively. Mechanical ventilation was provided to 63%. Organ failure developed in 94% and sepsis in 62%. The ICU, hospital, and 30-day mortality rates were 33%, 46%, and 52%, respectively. Allogeneic transplant and higher APACHE III scores, but not bone marrow stem cell source, were associated with increased mortality. Invasive mechanical ventilation, vasoactive medication use, sepsis, and organ failure during patients' ICU course were also associated with increased mortality. The area under the receiver operating characteristic curve for APACHE III hospital mortality prediction was 0.704 (95% confidence interval, 0.610-0.786). For APACHE III hospital mortality prediction, the value of the Hosmer-Lemeshow statistic showed good model fit.

CONCLUSIONS

Current mortality figures of HSCT recipients admitted to the ICU are better than previously reported. The APACHE III prognostic model has moderate discrimination and good calibration in predicting hospital mortality in these patients.

Assessment of six mortality prediction models in patients admitted with severe sepsis and septic shock to the intensive care unit: a prospective cohort study

INTRODUCTION

We conducted the present study to assess the validity of mortality prediction systems in patients admitted to the intensive care unit (ICU) with severe sepsis and septic shock. We included Acute Physiology and Health Evaluation (APACHE) II, Simplified Acute Physiology Score (SAPS) II, Mortality Probability Model (MPM) II0 and MPM II24 in our evaluation. In addition, SAPS II and MPM II24 were customized for septic patients in a previous study, and the customized versions were included in this evaluation.

MATERIALS AND METHOD

This cohort, prospective, observational study was conducted in a tertiary care medical/surgical ICU. Consecutive patients meeting the diagnostic criteria for severe sepsis and septic shock during the first 24 hours of ICU admission between March 1999 and August 2001 were included. The data necessary for mortality prediction were collected prospectively as part of the ongoing ICU database. Predicted and actual mortality rates, and standardized mortality ratio were calculated. Calibration was assessed using Lemeshow-Hosmer goodness of fit C-statistic. Discrimination was assessed using receiver operating characteristic curves.

RESULTS

The overall mortality prediction was adequate for all six systems because none of the standardized mortality ratios differed significantly from 1. Calibration was inadequate for APACHE II, SAPS II, MPM II0 and MPM II24. However, the customized version of SAPS II exhibited significantly improved calibration (C-statistic for SAPS II 23.6 [P = 0.003] and for customized SAPS II 11.5 [P = 0.18]). Discrimination was best for customized MPM II24 (area under the receiver operating characteristic curve 0.826), followed by MPM II24 and customized SAPS II.

CONCLUSION

Although general ICU mortality system models had accurate overall mortality prediction, they had poor calibration. Customization of SAPS II and, to a lesser extent, MPM II24 improved calibration. The customized model may be a useful tool when evaluating outcomes in patients with sepsis.

Recent innovations in intensive care unit risk-prediction models

During the past 20 years, ICU risk-prediction models have undergone significant development, validation, and refinement. Among the general ICU severity of illness scoring systems, the Acute Physiology and Chronic Health Evaluation (APACHE), Mortality Prediction Model (MPM), and the Simplified Acute Physiology Score (SAPS) have become the most accepted and used. To risk-adjust patients with longer, more severe illnesses like sepsis and acute respiratory distress syndrome, several models of organ dysfunction or failure have become available, including the Multiple Organ Dysfunction Score (MODS), the Sequential Organ Failure Assessment (SOFA), and the Logistic Organ Dysfunction Score (LODS). Recent innovations in risk adjustment include automatic physiology and diagnostic variable retrieval and the use of artificial intelligence. These innovations have the potential of extending the uses of case-mix and severity-of-illness adjustment in the areas of clinical research, patient care, and administration. The challenges facing intensivists in the next few years are to further develop these models so that they can be used throughout the IUC stay to assess quality of care and to extend them to more specific patient groups such as the elderly and patients with chronic ICU courses.

Qualitative review of intensive care unit quality indicators

PURPOSE

The purpose of this study was to (1) conduct a systematic review of the literature to identify interventions that improve patient outcomes in the intensive care unit (ICU); (2) evaluate potential measures of quality based on the impact, feasibility, variability, and the strength of evidence to support each measure and to categorize these measures as outcome, process, access, or complication measures; and (3) select a list of candidate quality measures that can be broadly applied to improve ICU care.

METHODS

We identified and independently reviewed all studies in Medline (1965-2000) and The Cochrane Library (Issue 3, 2001) that met the following criteria: design: observational studies, experimental trials, or systematic reviews; population: critically ill adults; and intervention: process or structure measure that was associated with improved patient outcomes: morbidity, mortality, complications, errors, costs, length of stay (LOS), and patient reported outcomes. Studies were grouped into categories by the type of outcome that was improved by the intervention. Potential quality measures were evaluated for: impact on morbidity, mortality, and costs; feasibility of the measure; and variability in the measure. We evaluated the strength of evidence for each intervention used to improve outcomes and using the Delphi method, assigned an over-all recommendation for each quality measure.

RESULTS

A total of 3,014 citations were identified. Sixty-six studies that met selection criteria reported on a variety of interventions that were associated with improved patient outcomes. We identified 6 outcome measures: ICU mortality rate, ICU LOS greater than 7 days, average ICU LOS, average days on mechanical ventilation, suboptimal management of pain, and patient/family satisfaction; 6 process measures: effective assessment of pain, appropriate use of blood transfusions, prevention of ventilator-associated pneumonia, appropriate sedation, appropriate peptic ulcer disease prophylaxis, and appropriate deep venous thrombosis prophylaxis; 4 access measures: rate of delayed admissions, rate of delayed discharges, cancelled surgical cases, and emergency department by-pass hours; and 3 complication measures: rate of unplanned ICU readmission, rate of catheter-related blood stream infections, and rate of resistant infections.

CONCLUSIONS

Further work is needed to create operational definitions and to pilot test the selected measures. The value of these measures will be determined by our ability to evaluate our current performance and implement interventions designed to improve the quality of ICU care.

Multivariate regression analysis on early mortality after orthotopic liver transplantation

AIM

To identify the risk factors relating to early mortality after orthotopic liver transplantation.

METHODS

Clinical data of 37 adult patients undergoing liver transplantation were retrospectively collected and divided into two groups: the survived group and the death group (survival time<30 d). The relationship between multivariate risk factors and early mortality after orthotopic liver transplantation were analyzed by stepwise logistic regression.

RESULTS

The survival rate was 73%. Early mortality rate was 27%. APACAE III, preoperative serum creatinine level and interoperative bleeding quantity had a significant independent association with early mortality. (R=0.1841, 0.2056 and 0.3738).

CONCLUSION

APACHE III,preoperative serum creatinine level and interoperative bleeding quantity are significant risk factors relating to early mortality after orthotopic liver transplantation. To improve the recipient's preoperative critical condition and renal function and to reduce interoperative bleeding quantity could lower the early mortality after orthotopic liver transplantation.

External validation of a modified model of Acute Physiology and Chronic Health Evaluation (APACHE) II for orthotopic liver transplant patients

INTRODUCTION

The purpose of the study was to validate the newly derived postoperative orthotopic liver transplantation (OLTX)-specific diagnostic weight for the Acute Physiology and Chronic Health Evaluation (APACHE) II mortality prediction system in independent databases.

METHODS

Medical records of 174 liver transplantation patients admitted postoperatively to the adult intensive care units at King Fahad National Guard Hospital and the University of Wisconsin were reviewed, and data on age, sex, the underlying liver disease, APACHE II scores and the hospital outcome were collected. Predicted mortality was calculated using: 1) the original APACHE II diagnostic weight of postoperative other gastrointestinal surgery and 2) the newly derived OLTX-specific diagnostic category weight. Standardized mortality ratio and 95% confidence intervals were calculated. Calibration was evaluated with the Hosmer-Lemeshow goodness-of-fit C-statistic. Discrimination was tested by 2 x 2 classification matrices and by computing the areas under the receiver operating characteristic curves. Patient characteristics and outcome data were compared between the two hospitals.

RESULTS

APACHE II significantly overestimated mortality when the original diagnostic weight was used, but provided a closer estimate of mortality with the OTLX-specific diagnostic weight. The C-statistic analysis showed better calibration for the new approach; discrimination was also improved. The performances of the prediction systems were similar in the two hospitals. The new model provided more accurate estimates of hospital mortality in each hospital.

DISCUSSION

APACHE II provided an accurate estimate of mortality in liver transplant patients when the OLTX-specific diagnostic weight was used. With the new model, APACHE II can be used as a valid mortality prediction system in this group of patients.

The customization of APACHE II for patients receiving orthotopic liver transplants

General outcome prediction models developed for use with large, multicenter databases of critically ill patients may not correctly estimate mortality if applied to a particular group of patients that was under-represented in the original database. The development of new diagnostic weights has been proposed as a method of adapting the general model - the Acute Physiology and Chronic Health Evaluation (APACHE) II in this case - to a new group of patients. Such customization must be empirically tested, because the original model cannot contain an appropriate set of predictive variables for the particular group. In this issue of Critical Care, Arabi and co-workers present the results of the validation of a modified model of the APACHE II system for patients receiving orthotopic liver transplants. The use of a highly heterogeneous database for which not all important variables were taken into account and of a sample too small to use the Hosmer-Lemeshow goodness-of-fit test appropriately makes their conclusions uncertain.

Developing and implementing measures of quality of care in the intensive care unit

As consumers, payers, and regulatory agencies require evidence regarding quality of care, the demand for intensive care unit (ICU) quality measures will likely grow. ICU providers and professional societies may need to partner with experts in quality measurement to develop and implement quality measures. This essay outlines the steps for developing and implementing quality measures and provides examples of potential ICU quality indicators. Outcome measures, in particular mortality rates, require risk adjustment, making data collection burdensome and broad implementation unlikely. On the other hand, structure and process measures may be feasible to implement broadly. Given the steps for developing quality measures outlined in this essay and the growing evidence in the literature regarding the impact of ICU care, the future should realize the development and implementation of ICU quality indicators that are rigorously developed and provide insights into opportunities to improve the quality of ICU care.