Intensive care unit length of stay: recent changes and future challenges.

OBJECTIVE: To compare case-mix adjusted intensive care unit (ICU) length of stay for critically ill patients with a variety of medical and surgical diagnoses during a 5-yr interval. DESIGN: Nonrandomized cohort study. SETTING: A total of 42 ICUs at 40 US hospitals during 1988-1990 and 285 ICUs at 161 US hospitals during 1993-1996. PATIENTS: A total of 17,105 consecutive ICU admissions during 1988-1990 and 38,888 consecutive ICU admissions during 1993-1996. MEASUREMENTS AND MAIN RESULTS: We used patient demographic and clinical characteristics to compare observed and predicted ICU length of stay and hospital mortality. Outcomes for patients studied during 1993-1996 were predicted using multivariable models that were developed and cross-validated using the 1988-1990 database. The mean observed hospital length of stay decreased by 3 days (from 14.8 days during 1988-1990 to 11.8 days during 1993-1996), but the mean observed ICU length of stay remained similar (4.70 vs. 4.53 days). After adjusting for patient and institutional differences, the mean predicted 1993-1996 ICU stay was 4.64 days. Thus, the mean-adjusted ICU stay decreased by 0.11 days during this 5-yr interval (T-statistic, 4.35; p < .001). The adjusted mean ICU length of stay was not changed for patients with 49 (75%) of the 65 ICU admission diagnoses. In contrast, the mean observed hospital length of stay was significantly shorter for 47 (72%) of the 65 admission diagnoses, and no ICU admission diagnosis was associated with a longer hospital stay. Aggregate risk-adjusted hospital mortality during 1993-1996 (12.35%) was not significantly different during 1988-1990 (12.27%, p = .54). CONCLUSIONS: For patients admitted to ICUs, the pressures associated with a decrease in hospital length of stay do not seem to have influenced the duration of ICU stay. Because of the high cost of intensive care, reduction in ICU stay may become a target for future cost-cutting efforts.

Evaluation of acute physiology and chronic health evaluation III predictions of hospital mortality in an independent database.

OBJECTIVE: To assess the accuracy and validity of Acute Physiology and Chronic Health Evaluation (APACHE) III hospital mortality predictions in an independent sample of U.S. intensive care unit (ICU) admissions. DESIGN: Nonrandomized, observational, cohort study. SETTING: Two hundred eighty-five ICUs in 161 U.S. hospitals, including 65 members of the Council of Teaching Hospitals and 64 nonteaching hospitals. PATIENTS: A consecutive sample of 37,668 ICU admissions during 1993 to 1996; including 25,448 admissions at hospitals with >400 beds and 1,074 admissions at hospitals with <200 beds. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We used demographic, clinical, and physiologic information recorded during ICU day 1 and the APACHE III equation to predict the probability of hospital mortality for each patient. We compared observed and predicted mortality for all admissions and across patient subgroups and assessed predictive accuracy using tests of discrimination and calibration. Aggregate hospital death rate was 12.35% and predicted hospital death rate was 12.27% (p =.541). The model discriminated between survivors and nonsurvivors well (area under receiver operating curve = 0.89). A calibration curve showed that the observed number of hospital deaths was close to the number of deaths predicted by the model, but when tested across deciles of risk, goodness-of-fit (Hosmer-Lemeshow statistic, chi-square = 48.71, 8 degrees of freedom, p< .0001) was not perfect. Observed and predicted hospital mortality rates were not significantly (p < .01) different for 55 (84.6%) of APACHE III's 65 specific ICU admission diagnoses and for 11 (84.6%) of the 13 residual organ system-related categories. The most frequent diagnoses with significant (p < .01) differences between observed and predicted hospital mortality rates included acute myocardial infarction, drug overdose, nonoperative head trauma, and nonoperative multiple trauma. CONCLUSIONS: APACHE III accurately predicted aggregate hospital mortality in an independent sample of U.S. ICU admissions. Further improvements in calibration can be achieved by more precise disease labeling, improved acquisition and weighting of neurologic abnormalities, adjustments that reflect changes in treatment outcomes over time, and a larger national database.

Neurological intensive care admissions: identifying candidates for intermediate care and the services they receive.

OBJECTIVE: The high cost and scarcity of intensive care unit (ICU) beds has resulted in a need for improved utilization. This study describes the characteristics of patients who are admitted to the ICU for neurosurgical and neurological care, identifies patients who might receive all or most of their care in an intermediate care unit, and describes the services the patients would receive in an intermediate care unit. METHODS: We describe patients who received neurological care and who were part of a prospective study of 17,440 patients admitted to 42 ICUs at 40 United States hospitals. We identified patients who received only monitoring during ICU Day 1 and then used a previously validated equation to distinguish which patients were at low risk (< 10%) for subsequent active life-supporting therapy. We also describe the services these patients received during their ICU stay. RESULTS: Among 3000 patients admitted to the ICU for neurological care, 1350 received active therapy and 1650 (55%) underwent monitoring and received concentrated nursing care on ICU Day 1. After excluding those patients who received active therapy at admission, 1288 (78%) of the 1650 patients who underwent monitoring at admission were at low risk (< 10%) for subsequent active therapy; 95.8% received no active therapy. These patients who were at low risk for subsequent active therapy were significantly (P < 0.001) more often admitted postoperatively, were younger and less severely ill, and had lower ICU and hospital mortality rates (0.9 and 3.9%, respectively) than patients who received active treatment at admission. CONCLUSIONS: Patients receiving neurological care at an ICU who receive only monitoring during their 1st ICU day and have a less than 10% predicted risk of active treatment can be safely transferred to an intermediate care unit. Some of these patients may not require ICU admission. We suggest guidelines for equipping and staffing neurological intermediate care units based on the type and amount of therapy received by these patients.

Planning patient services for intermediate care units: insights based on care for intensive care unit low-risk monitor admissions.

OBJECTIVE: To describe the technology and nursing services that would be required to care for intensive care unit (ICU) low-risk monitor admissions in an intermediate unit. DESIGN: Prospective, multicenter, inception cohort analysis. SETTING: Forty U.S. hospitals with > 200 beds, including 26 hospitals that were randomly selected and 14 that volunteered for the study. PATIENTS: A sample of 8,040 ICU patients admitted to the ICU for monitoring, who received no active life-support treatment on ICU day 1. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Demographic, physiologic, and treatment information were obtained during ICU days 1 to 7. A previously validated multivariate equation was used to identify 6,180 monitor admissions at low (< 10%) risk for receiving active treatment during their entire ICU stay. We used daily Therapeutic intervention Scoring System (TISS) data to identify the equipment, type and amount of nursing care, and the types of active treatment that would have been used had these ICU patients been admitted to an intermediate care unit. Mean day-1 ICU TISS scores were as follows: 16.4 for all patients; 18.3 for surgical patients; and 13.5 for medical admissions. Concentrated nursing care accounted for 89% and technologic monitoring for 11% of day-1 TISS points. Surgical admissions had a 2.8-day mean ICU length of stay and received an average of 16.5 TISS points per patient per day. Medical admissions had a 2.7-day mean ICU length of stay and received an average of 12.3 TISS points per patient per day. Subsequent active life-support therapy was received by 4.4% of these ICU low-risk monitor admissions. CONCLUSIONS: The services received by ICU low-risk monitor admissions provide insight regarding the equipment and nursing care that might be required, and the kinds of emergencies that might occur, if these patients were cared for in medical and surgical intermediate care units. Our data suggest that if ICU low-risk monitor patients were admitted to an intermediate care unit, they would mainly require concentrated nursing care (nurse/patient ratio of 1:3 to 1:4) and limited technologic monitoring.

Predicting the duration of mechanical ventilation. The importance of disease and patient characteristics.

STUDY OBJECTIVE: To analyze the determinants of an individual patient's duration of mechanical ventilation and assess interhospital variations for average durations of ventilation. DESIGN: Prospective, multicenter, inception, cohort study. SETTING: Forty-two ICUs at 40 US hospitals. PATIENTS: A total of 5,915 patients undergoing mechanical ventilation on ICU day 1 selected from the acute physiology and chronic health evaluation (APACHE) III database of 17,440 admissions. INTERVENTIONS: None. MEASUREMENTS AND RESULTS: Utilizing APACHE III data collected on the 5,915 patients, multivariate regression analysis was performed on selected patients and disease characteristics to determine which variables were significantly associated with the duration of mechanical ventilation. An equation predicting duration of ventilation was then developed using the significant predictor variables and its accuracy was evaluated. Variables significantly associated with duration of ventilation included primary reason for ICU admission, day 1 acute physiology score (APS) of APACHE III, age, prior patient location and hospital length of stay, activity limits due to respiratory disease, serum albumin, respiratory rate, and PaO2/FIo2 measurements. Using an equation derived from these variables, predicted durations of ventilation were then calculated and compared with actual observed durations for each of the 42 ICUs. Average duration of ventilation for the 42 ICUs ranged from 2.6 to 7.9 days, but 60% of this variation was accounted for by differences in patient characteristics. CONCLUSIONS: For patients admitted to the ICU and ventilated on day 1, total duration of ventilation is primarily determined by admitting diagnosis and degree of physiologic derangement as measured by APS. An equation developed using multivariate regression techniques can accurately predict average duration of ventilation for groups of ICU patients, and we believe this equation will be useful for comparing ventilator practices between ICUs, controlling for patient differences in clinical trials of new therapies or weaning techniques, and as a quality improvement mechanism.

Use of predicted risk of mortality to evaluate the efficacy of anticytokine therapy in sepsis. The rhIL-1ra Phase III Sepsis Syndrome Study Group.

OBJECTIVES: To investigate a novel anticytokine therapy in patients with sepsis syndrome, and the relationship between a patient's baseline mortality risk and survival benefit. DESIGN: Data from a recent phase III, double-blind, placebo-controlled, multicenter clinical trial with patients randomized to three treatment arms: an intravenous loading dose of recombinant human interleukin-1-receptor antagonist (rhIL-1ra) or placebo, followed by a continuous infusion of rhIL-1ra (1.0 mg/kg/hr, or 2.0 mg/kg/hr), or placebo for 72 hrs. SETTING: Sixty-three investigative centers in eight countries. PATIENTS: The study population consisted of 893 patients: 302 placebo patients; 298 patients treated with 1.0 mg/kg/hr of rhIL-1ra; and 293 patients treated with 2.0 mg/kg/hr of rhIL-1ra. MEASUREMENTS AND MAIN RESULTS: An independent, sepsis-specific, log-normal regression model that predicts the risk of mortality over 28 days was applied to all patients enrolled into the rhIL-1ra sepsis study. The ability of the Predicted Risk of Mortality model to predict 28-day mortality in the placebo patients was determined and the relationship between mortality risk and efficacy of rhIL-1ra was investigated. The trial data were also analyzed using two other risk-assessment models for comparison with Predicted Risk of Mortality. A significant increase in survival time was demonstrated for all patients treated with rhIL-1ra (n = 893, p < .02 Predicted Risk of Mortality log-normal), but patients with a Predicted Risk of Mortality of < 24% derived little benefit. Retrospective examination of time-to-death data demonstrated that rhIL-1ra reduced risk of death in the first 2 days for patients with > or = 24% Predicted Risk of Mortality (n = 580, p < .005 Predicted Risk of Mortality log-normal). This same effect was not present in patients with a Predicted Risk of Mortality of < 24% on entry into the study. The Predicted Risk of Mortality model predicted a 28-day mortality rate of 35% for placebo patients compared with 34% observed and accurately stratified patients along the full range of risks. There was a wide distribution of individual patient risks for 28-day mortality for all patients, as well as within categorical subgroups, such as shock and organ system dysfunction. Two alternate risk models were assessed and the Acute Physiology Score of Acute Physiology and Chronic Health Evaluation III also demonstrated a statistically significant survival benefit for rhIL-1ra (p = .04 Predicted Risk of Mortality log-normal) for all patients treated. CONCLUSIONS: Using an appropriate analytic model, a statistically significant increase in survival time from rhIL-1ra was measured. A direct relationship was found between a patient's Predicted Risk of Mortality at study entry to efficacy of rhIL-1ra. Individual risk or severity assessment may be a useful tool for evaluating the clinical benefit of new therapeutic approaches to sepsis and for monitoring outcomes at the bedside.

The use of risk predictions to identify candidates for intermediate care units. Implications for intensive care utilization and cost.

OBJECTIVE: To develop a predictive equation that estimates the probability of life-supporting therapy among ICU monitor admissions and to explore its potential for reducing cost and improving ICU utilization. DESIGN: Prospective inception cohort analysis. PARTICIPANTS: Forty-two ICUs in 40 US hospitals with more than 200 beds and a consecutive sample of 17,440 ICU admissions. INTERVENTIONS: A multivariate equation was developed to estimate the probability of life support for ICU monitoring admissions during an entire ICU stay. MEASUREMENTS: Demographic, physiologic, and treatment information obtained during the first 24 h in the ICU and over the first 7 ICU days. RESULTS: The most important determinants of subsequent risk for life-supporting (active) treatment were diagnosis, the acute physiology score of APACHE III, age, operative status, and the patient's location and hospital length of stay before ICU admission. Among 8,040 ICU monitoring admissions, 6,180 (76.8%) had a low (< 10%) risk for receiving active treatment during the ICU stay; 95.6% received no subsequent active treatment. Review of outcomes and the type and amount of therapy received suggest that most low-risk ICU monitor admissions could be safely cared for in an intermediate care setting. CONCLUSION: Objective predictions can accurately identify groups of ICU admissions who are at a low risk for receiving life support. This capability can be used to assess ICU resource use and develop strategies for providing graded critical care services at a reduced cost.

The use of APACHE III to evaluate ICU length of stay, resource use, and mortality after coronary artery by-pass surgery.

OBJECTIVE: To identify patient characteristics that are associated with increased ICU length of stay, resource use, and hospital mortality after coronary artery bypass surgery. DESIGN: Prospective, multicenter study. SETTING: Six tertiary care hospitals. PARTICIPANTS: A consecutive sample of 2,435 unselected ICU admissions following coronary artery by-pass surgery. MATERIALS AND METHODS: Demographic, operative characteristics and APACHE III score were collected during the first postoperative day; and APACHE III scores and therapeutic interventions during the first three postoperative days. Hospital survival and ICU length of stay were also recorded. Multivariate equations were derived and cross-validated to predict hospital mortality, ICU length of stay, and ICU resource use. RESULTS: Unadjusted hospital mortality rate was 3.9% (range 1.0% to 6.0%), mean ICU length of stay was 3.7 days (range 3.2 to 4.7 days), and first 3-day ICU resource use (TISS points) was 99 (range 68 to 116). The range of actual to predicted ICU length of stay varied from 0.86 to 1.26; and resource use from 0.71 to 1.16. CONCLUSIONS: A limited number of operative characteristics, the post-operative acute physiology score (APS) of APACHE III and patient demographic data can predict hospital death rate, ICU length of stay, and resource use immediately following coronary by-pass surgery. These estimates may compliment assessments based on pre-operative risk factors in order to more precisely evaluate and improve the efficacy and efficiency of cardiovascular surgery.

Improving intensive care unit discharge decisions: supplementing physician judgment with predictions of next day risk for life support.

OBJECTIVE: To develop predictive equations, estimating the probability that an individual intensive care unit (ICU) patient will receive life support within the next 24 hrs. DESIGN: Prospective, multicenter, inception cohort study. SETTING: Forty-two ICUs in 40 U.S. hospitals, including 26 that were randomly selected and 14 volunteer hospitals, primarily university or large tertiary care centers. PATIENTS: A consecutive sample of 17,440 ICU admissions. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A series of multivariate equations were developed to create daily estimates of probability of life support in the next 24 hrs. These equations used demographic, physiologic, and treatment information obtained at the time of ICU admission and during the first 7 ICU days. The most important determinants of next day risk for life support were the current day's therapy and Acute Physiology Score of the Acute Physiology and Chronic Health Evaluation (APACHE) III score. Other predictor variables included diagnosis, age, chronic health status, emergency surgery, previous day Acute Physiology Score, and hospital stay and location before ICU admission. The cross-validated ICU day 1, 2, and 3 predictive equations had receiver operating characteristic areas of 0.90. Survival, ICU readmission rate, and the number and type of therapies received by patients predicted at < 10% risk for active treatment suggest that discharge of patients meeting these criteria to an intermediate care unit or hospital ward could reduce ICU bed demand without compromising patient safety. CONCLUSIONS: Accurate, objective predictions of next day risk for life support can be developed, using readily available patient information. Supplementing physician judgment with these objective risk assessments deserves evaluation for the role of these assessments in enhancing patient safety and improving ICU resource utilization.

The performance of intensive care units: does good management make a difference?

A significant portion of health care resources are spent in intensive care units with, historically, up to two-fold variation in risk-adjusted mortality. Technological, demographic, and social forces are likely to lead to an increased volume of intensive care in the future. Thus, it is important to identify ways of more efficiently managing intensive care units and reducing the variation in patient outcomes. Based on data collected from 17,440 patients across 42 ICUs, the present study examines the factors associated with risk-adjusted mortality, risk-adjusted average length of stay, nurse turnover, evaluated technical quality of care, and evaluated ability to meet family member needs. Using the Apache III methodology for risk-adjustment, findings reveal that: 1) technological availability is significantly associated with lower risk-adjusted mortality (beta = -.42); 2) diagnostic diversity is significantly associated with greater risk-adjusted mortality (beta = .46); and 3) caregiver interaction comprising the culture, leadership, coordination, communication, and conflict management abilities of the unit is significantly associated with lower risk-adjusted length of stay (beta = .34), lower nurse turnover (beta = -.36), higher evaluated technical quality of care (beta = .81), and greater evaluated ability to meet family member needs (beta = .74). Furthermore, units with greater technological availability are significantly more likely to be associated with hospitals that are more profitable, involved in teaching activities, and have unit leaders actively participating in hospital-wide quality improvement activities. The findings hold a number of important managerial and policy implications regarding technological adoption, specialization, and the quality of interaction among ICU team members. They suggest intervention "leverage points" for care givers, managers, and external policy makers in efforts to continuously improve the outcomes of intensive care.

Intensive care at two teaching hospitals: an organizational case study.

OBJECTIVE: To examine structural and organizational characteristics at two ICUs with marked differences in risk-adjusted survival. METHODS: We performed on-site organizational analysis in two ICUs at two major teaching hospitals. Our main outcome measures were interviews and direct observations by a team of clinical and organizational researchers; demographic, clinical, and survival data for 888 ICU admissions; and questionnaire responses from 70 nurses and 42 physicians on ICU structure and organization. ICU performance was measured using risk-adjusted survival and the ratios of actual to predicted ICU length of stay and resource use. RESULTS: Structural and organizational questionnaires, self-evaluation by staff members, and the research team's implicit judgments following detailed on-site analysis failed to distinguish units with higher and lower risk-adjusted survival. Both units exhibited practices to emulate and practices to avoid. CONCLUSIONS: The methods used in this study can identify organizational problems and potential means for improvement. The best practices and suggestions for improvement at these units provide examples of methods for improving ICU management.

Do-not-resuscitate orders in intensive care units. Current practices and recent changes.

OBJECTIVES: To describe the characteristics of patients with do-not-resuscitate (DNR) orders and the frequency and timing of these orders in a representative sample of intensive care units (ICUs) and to compare practices from 1980 to 1990. DESIGN: Prospective inception cohort. SETTING: A total of 42 ICUs in 40 US hospitals with 200 or more beds; 26 randomly selected hospitals and 14 large, tertiary care hospitals that volunteered to be studied. PARTICIPANTS: A consecutive sample of 17,440 ICU admissions from 1988 to 1990. MEASUREMENTS: Patient demographic characteristics, comorbid conditions, disease, and physiological abnormalities. MAIN OUTCOME MEASURES: Frequency and timing of DNR orders; ICU resource use before and after DNR orders; and patients' hospital and ICU discharge status. RESULTS: Physicians wrote DNR orders for 1577 ICU admissions (9%) (hospital range, 1.5% to 22%). Patients with ICU DNR orders were older, more functionally impaired, had more comorbid illness, a higher severity of illness, and required the use of more ICU resources compared with patients without DNR orders. Compared with data from a similar survey from 1979 to 1982, ICU DNR orders were more frequent in 1988 to 1990 (9% vs 5.4%; P < .001) and preceded 60% of all in-unit deaths compared with only 39% in 1979 to 1982 (P < .001). Do-not-resuscitate orders were written sooner (for 3.6% vs 2.0% of patients on day 1 in the ICU) and patients with DNR orders remained in the ICU longer in 1988 to 1990 (2.8 vs 1.4 days) than in 1979 to 1982, and had lower ICU and hospital mortality rates (64% vs 74%, P < .001; and 85% vs 94%, P < .001). CONCLUSIONS: Over the last decade physicians and patients' families set limits earlier and more frequently in cases likely to have poor outcomes. We attribute this change to a greater dialogue about setting limits to care and a greater knowledge of treatment outcomes among physicians and families. These changes in practice preceded implementation of the Patient Self-determination Act, designed to ensure patient autonomy for decisions about life-sustaining therapy.

Value and cost of teaching hospitals: a prospective, multicenter, inception cohort study.

OBJECTIVE: To examine variations in case-mix, structure, resource use, and outcome performance among teaching and nonteaching intensive care units (ICU). DESIGN: Prospective inception cohort study. PATIENTS: A consecutive sample of 15,297 patients at 35 hospitals, which compared 8,269 patients admitted to 20 teaching ICUs at 18 hospitals vs. 7,028 patients admitted to 17 non-teaching ICUs at 17 hospitals. INTERVENTIONS: None. MEASUREMENTS: We selected demographic, physiologic, and treatment information for an average of 415 patients at each ICU, and collected data on hospital and ICU structure. Outcomes were compared using ratios of observed to risk-adjusted predicted hospital death rates, ICU length of stay, and resource use. MAIN RESULTS: When compared to nonteaching ICUs, teaching ICUs had twice the number of physicians who regularly provided services and cared for significantly younger and more severely ill (p < .001) patients. Risk-adjusted ICU length of stay was similar, but resource use was significantly (p < .001) greater in teaching ICUs, with $3,000 (10.5%) of estimated total costs for an average ICU admission related to increased use of diagnostic testing and invasive procedures in teaching ICUs. Risk-adjusted hospital death rates were not significantly different (p = .1) between all teaching and nonteaching ICUs, but were significantly (p < .05) better in four teaching ICUs, but in only one nonteaching ICU. The 14 hospitals that were members of the Council of Teaching Hospitals had significantly better risk-adjusted outcome in their 16 ICUs than all others (odds ratio = 1.21, confidence interval 1.06 to 1.38, p = .004). CONCLUSIONS: Teaching ICUs care for more complex patients in a substantially more complicated organizational setting. The best risk-adjusted survival rates occur at teaching ICUs, but production cost is higher in teaching units, secondary to increased testing and therapy. Teaching ICUs are also successfully transferring knowledge to trainees who, after their training, are achieving equivalent results at slightly lower cost in nonteaching ICUs.

Improving intensive care: observations based on organizational case studies in nine intensive care units: a prospective, multicenter study.

OBJECTIVE: To examine organizational practices associated with higher and lower intensive care unit (ICU) outcome performance. DESIGN: Prospective multicenter study. Onsite organizational analysis; prospective inception cohort. SETTING: Nine ICUs (one medical, two surgical, six medical-surgical) at five teaching and four nonteaching hospitals. PARTICIPANTS: A sample of 3,672 ICU admissions; 316 nurses and 202 physicians. MATERIALS AND METHODS: Interviews and direct observations by a team of clinical and organizational researchers. Demographic, physiologic, and outcome data for an average of 408 admissions per ICU; and questionnaires on ICU structure and organization. The ratio of actual/predicted hospital death rate was used to measure ICU effectiveness; the ratio of actual/predicted length of ICU stay was used to assess efficiency. MEASUREMENTS AND MAIN RESULTS: ICUs with superior risk-adjusted survival could not be distinguished by structural and organizational questionnaires or by global judgment following on-site analysis. Superior organizational practices among these ICUs were related to a patient-centered culture, strong medical and nursing leadership, effective communication and coordination, and open, collaborative approaches to solving problems and managing conflict. CONCLUSIONS: The best and worst organizational practices found in this study can be used by ICU leaders as a checklist for improving ICU management.

The clinical evaluation of new drugs for sepsis. A prospective study design based on survival analysis.

OBJECTIVE: To develop a survival model and severity assessment method to estimate the 28-day mortality risk for patients with sepsis syndrome entering phase 2 and 3 drug evaluations. DESIGN: Retrospective analysis of intensive care unit admissions with sepsis syndrome by means of log-normal regression to identify risk factors for 28-day mortality. Prospective application of the model to patients with gram-negative infection meeting sepsis syndrome criteria from separate data collection (validation group). PATIENTS: A total of 58,737 intensive care unit admissions at 107 hospitals in the United States and Western Europe screened to yield 1195 patients meeting entry criteria for the sepsis syndrome study for the original model; 295 hospitalized patients with gram-negative infection meeting criteria for sepsis syndrome for validation. MAIN OUTCOME MEASURES: Survival time and mortality at 28 days after fulfillment of the sepsis syndrome criteria. RESULTS: Acute physiologic abnormalities were the most important prognostic factors influencing outcome (82% of total chi 2). Specific disease resulting in intensive care unit admission and the time the patient was in the hospital and intensive care unit before qualification were also independent risks, as were age and a clinical history of cirrhosis. The model's overall classification accuracy was a Somers' Dyx of .52 (rank correlation between predicted risk and 28-day mortality) (receiver operating characteristic area, 0.76), with equal accuracy (Dyx = .59; receiver operating characteristic area, 0.80) in the independent group of patients. CONCLUSIONS: We created an accurate independent estimate for 28-day mortality risk for patients with sepsis syndrome (severe sepsis). This estimate could improve the evaluation of new drugs by investigating whether the drug's benefit varies by patient risk and then determining the amount of benefit for individual patients.

Variations in mortality and length of stay in intensive care units.

OBJECTIVE: To evaluate the amount of variation in in-hospital mortality and length of intensive care unit (ICU) stay that can be accounted for by clinical data available at ICU admission. DESIGN: Inception cohort study. SETTING: Forty-two ICUs in 40 hospitals, including 26 hospitals that were randomly selected and 14 large tertiary care hospitals that volunteered for the study. PARTICIPANTS: A consecutive sample of 16,622 patients and 17,440 ICU admissions. MEASUREMENTS AND MAIN OUTCOMES: Data on selected demographic characteristics, comorbidity, and specific physiologic variables were recorded during the first ICU day for an average of 415 admissions at each ICU; hospital discharge status (dead or alive) and length of ICU stay were recorded for individual patients; and the ratio of actual to predicted in-hospital mortality, standardized mortality ratios, and the ratio of actual to predicted length of ICU stay were recorded for individual ICUs. RESULTS: Unadjusted in-hospital mortality rates for the 42 units varied from 6.4% to 40%, and 90% (R2 = 0.90) of this variation was attributable to patient characteristics at admission. The standard mortality ratio varied from 0.67 to 1.25. The mean unadjusted length of ICU stay varied from 3.3 to 7.3 days, and 78% of the variation (R2 = 0.78) was attributed to patient and selected institutional characteristics. The best performing unit had a length of stay ratio of 0.88, whereas the poorest performing unit had a ratio of 1.21. CONCLUSIONS: Clinicians can use readily available admission data to adjust for considerable variations in patient severity and type in different ICUs. Such data should permit precise evaluation and comparison of ICU effectiveness and efficiency, which varied substantially in this study, and result in improved methods of risk prediction and evaluation of new medical practices.

Reliability of a measure of severity of illness: acute physiology of chronic health evaluation--II.

This study examines how reliably the components of the APACHE II score (Acute Physiology Score (APS), age and chronic health) are abstracted from the medical record in terms of inter-rater reproducibility (Intraclass Correlation Coefficient [ICC], kappa). In the sample studied, assignment of the APS is highly reproducible (ICC = 0.90). Reproducibility of the age variable (ICC = 0.998) suggests that age is accurately abstracted. Chronic health data does not fare as well as the APS and age (kappa = 0.66). This study suggests that the components of the APACHE II score can be collected reliably.

The APACHE III prognostic system. Risk prediction of hospital mortality for critically ill hospitalized adults.

The objective of this study was to refine the APACHE (Acute Physiology, Age, Chronic Health Evaluation) methodology in order to more accurately predict hospital mortality risk for critically ill hospitalized adults. We prospectively collected data on 17,440 unselected adult medical/surgical intensive care unit (ICU) admissions at 40 US hospitals (14 volunteer tertiary-care institutions and 26 hospitals randomly chosen to represent intensive care services nationwide). We analyzed the relationship between the patient's likelihood of surviving to hospital discharge and the following predictive variables: major medical and surgical disease categories, acute physiologic abnormalities, age, preexisting functional limitations, major comorbidities, and treatment location immediately prior to ICU admission. The APACHE III prognostic system consists of two options: (1) an APACHE III score, which can provide initial risk stratification for severely ill hospitalized patients within independently defined patient groups; and (2) an APACHE III predictive equation, which uses APACHE III score and reference data on major disease categories and treatment location immediately prior to ICU admission to provide risk estimates for hospital mortality for individual ICU patients. A five-point increase in APACHE III score (range, 0 to 299) is independently associated with a statistically significant increase in the relative risk of hospital death (odds ratio, 1.10 to 1.78) within each of 78 major medical and surgical disease categories. The overall predictive accuracy of the first-day APACHE III equation was such that, within 24 h of ICU admission, 95 percent of ICU admissions could be given a risk estimate for hospital death that was within 3 percent of that actually observed (r2 = 0.41; receiver operating characteristic = 0.90). Recording changes in the APACHE III score on each subsequent day of ICU therapy provided daily updates in these risk estimates. When applied across the individual ICUs, the first-day APACHE III equation accounted for the majority of variation in observed death rates (r2 = 0.90, p less than 0.0001).

Utilizing findings from the APACHE III research to develop operational information system for the ICU--the APACHE III ICU Management System.

The APACHE III data base reflects the disease, physiologic status, and outcome data from 17,400 ICU patients at 40 hospitals, 26 of which were randomly selected from representative geographic regions, bed size, and teaching status. This provides a nationally representative standard for measuring several important aspects of ICU performance. Results from the study have now been used to develop an automated information system to provide real time information about expected ICU patient outcome, length of stay, production cost, and ICU performance. The information system provides several new capabilities to ICU clinicians, clinic, and hospital administrators. Among the system's capabilities are: the ability to compare local ICU performance against predetermined criteria; the ability to forecast nursing requirements; and, the ability to make both individual and group patient outcome predictions. The system also provides improved administrative support by tracking ICU charges at the point of origin and reduces staff workload eliminating the requirement for several manually maintained logs and patient lists. APACHE III has the capability to electronically interface with and utilize data already captured in existing hospital information systems, automated laboratory information systems, and patient monitoring systems. APACHE III will also be completely integrated with several CIS vendors' products.