Search PubMed⌕ Search

Biomedical subjects

J R Le Gall

Publications and source records attributed to J R Le Gall.

At least 19 recordsLinked to original sources
PubMed2001

Evaluation of the logistic organ dysfunction system for the assessment of organ dysfunction and mortality in critically ill patients.

OBJECTIVES: To evaluate the performance of the logistic organ dysfunction (LOD) system for the assessment of morbidity and mortality in multiple organ dysfunction/failure (MOD/F) in an independent database and to evaluate the use of sequential LOD measurements for the prediction of outcome. DESIGN AND SETTING: Prospective, multicentric cohort study in 13 adult medical, surgical, and mixed intensive care units (ICUs) in Austria. PATIENTS: A total of 2,893 consecutive admissions to the ICUs. MEASUREMENTS AND MAIN RESULTS: Patient vital status at ICU and hospital discharge was recorded. Univariate analysis showed that the LOD was able to distinguish between survivors and nonsurvivors (2 vs. 6 median score). Within organ systems, higher levels of the severity of organ dysfunction were consistently associated with higher mortality. For the prediction of hospital mortality, the original prognostic LOD model did not perform well in our patients, as indicated by the goodness-of-fit C statistic. Using multiple logistic regression we developed a prognostic model with a satisfactory fit in our patients. The integration of further measurements during the ICU stay increased discrimination but not calibration. CONCLUSIONS: The LOD system is well correlated well with the numbers and levels of organ dysfunctions and discriminates well between survivors and nonsurvivors. It can thus be used to quantify the baseline severity of organ dysfunction. Moreover, after customization of the predictive equation the LOD predicted hospital mortality in our patients with high precision. It thus provides a combined measure of morbidity and mortality for critically ill patients with MOD/F.

Aged↗
PubMed2001

Routine exploratory thoracentesis in ICU patients with pleural effusions: results of a French questionnaire study.

PURPOSE: The purpose of this study was to report the opinions of intensivists regarding pleural effusions in patients in the intensive care unit (ICU). MATERIALS AND METHODS: Questionnaires were sent to 1,032 intensivists, who were members of the French Society of Critical Care. RESULTS: Four hundred thirty-one questionnaires (41.7%) were returned. Overall, the respondents' estimated the incidence of pleural effusion in ICU patients to be 22.19 +/- 17%, whereas 37 +/- 27% considered that exploratory thoracentesis was likely to determine the cause of the effusion, and 17.36 +/- 16% considered that its results were likely to result in a change in their therapeutic attitude. Sixty-five (15%) physicians, chiefly pulmonologists, performed exploratory thoracentesis routinely (Group 1). Compared with those who did not perform routine thoracentesis (Group 2), they ascribed a higher proportion of pleural effusions to infection (31.3% vs. 13.5%) and were more likely to consider that exploratory thoracentesis had a diagnostic and therapeutic contribution (51.2% vs. 34% and 23% vs. 16%, respectively). In addition to the respiratory medicine subspecialty, the practice of routine exploratory thoracentesis was significantly related to seniority, to the frequency of the suspicion of an infectious cause in the physician's practice, and to his or her appreciation of the risks associated with exploratory thoracentesis. Physicians from Group 1 were also more likely to describe exploratory thoracentesis as a noninvasive procedure. CONCLUSIONS: The beliefs and attitudes of intensivists regarding pleural effusions and exploratory thoracentesis are divergent. This may be due to the lack of precise guidelines on the topic and prompt the design of further studies to establish precisely the epidemiology and causes of pleural effusions in ICU patients.

Attitude of Health Personnel↗
PubMed2001

Improved survival in cancer patients requiring mechanical ventilatory support: impact of noninvasive mechanical ventilatory support.

OBJECTIVE: When a cancer patient becomes critically ill, mechanical ventilation (MV) is often considered futile. However, recent studies have found that outcomes of critically ill cancer patients have been improving over the years and that classic predictors of high mortality have lost their relevance. DESIGN: We retrospectively determined outcomes and predictors of 30-day mortality in 237 mechanically-ventilated cancer patients admitted to the intensive care unit (ICU). PATIENTS: The 132 (55.7%) patients who were admitted between 1990 and 1995 were compared with 105 (44.3%) patients who were admitted between 1996 and 1998. The malignancy was leukemia/lymphoma in 119 (50.3%) patients, myeloma in 50 (21%), and a solid tumor in 68 (28.7%). Forty-two (17.7%) patients had bone marrow transplantation, and 91 (38.4%) were neutropenic. Median Simplified Acute Physiology Score II (SAPS II) was 58 (range, 40-75). Reasons for MV were acute hypoxemic respiratory failure in 148 (62.5%) patients, coma in 54 (22.8%), and cardiogenic pulmonary edema in 35 (14.7%). Conventional MV was used first in 189 (79.8%) patients, and noninvasive MV (NIMV) was used in 48 (20.2%). Overall mortality rate was 72.5% (172 deaths). RESULTS: Logistic regression identified three variables associated with mortality: ICU admission between 1996 and 1998 (odds ratio [OR], 0.24; 95% confidence interval [CI], 0.12-0.50) and the use of NIMV (OR, 0.34; 95% CI, 0.16-0.73) were protective, and the SAPS II was aggravating (OR, 1.04 per point; 95% CI, 1.02-1.06). To better define the impact of NIMV, we performed a pairwise-matched exposed-unexposed analysis. Forty-eight patients who did and 48 who did not receive NIMV as the first ventilation method were matched for SAPS II, type of malignancy, and period of ICU admission. Crude ICU mortality rates from exposed patients and controls were 43.7% and 70.8%, respectively. NIMV remained protective from mortality after adjustment for matching variables (OR, 0.31; 95% CI, 0.12-0.82). CONCLUSION: Our results confirm that mortality has improved over the past decade in critically ill cancer patients, even those who require MV, and suggest that this may be, in part, because of a protective effect of NIMV.

APACHE↗
PubMed2001

Compliance with triage to intensive care recommendations.

DESIGN: Recommendations for triage to intensive care units (ICUs) have been issued but not evaluated. SETTING: In this prospective, multicenter study, all patients granted or refused admission to 26 ICUs affiliated with the French Society for Critical Care were included during a 1-month period. Characteristics of participating ICUs and patients, circumstances of triage, and description of the triage decision with particular attention to compliance with published recommendations were recorded. RESULTS: During the study period, 1,009 patients were and 283 were not admitted to the participating ICUs. Refused patients were more likely to be older than 65 yrs (odds ratio [OR], 3.53; confidence interval [CI], 1.98-5.32) and to have a poor chronic health status (OR, 3.09; CI, 2.05-4.67). An admission diagnosis of acute respiratory or renal failure, shock, or coma was associated with admission, whereas chronic severe respiratory and heart failure or metastatic disease without hope of remission were associated with refusal (OR, 2.24; CI, 1.38-3.64). Only four (range, 0-8) of the 20 recommendations for triage to ICU were observed; a full unit and triage over the phone were associated with significantly poorer compliance with recommendations (0 [0-2] vs. 6 [2-9], p =.0003; and 1 [0-6] vs. 6 [1-9], p <.0001; respectively). CONCLUSION: Recommendations for triage to intensive care are rarely observed, particularly when the unit is full or triage is done over the phone. These recommendations may need to be redesigned to improve their practicability under real-life conditions, with special attention to phone triage and triaging to a full unit.

Age Factors↗
PubMed2001

Meeting the needs of intensive care unit patient families: a multicenter study.

Intensive care unit (ICU) caregivers should seek to develop collaborative relationships with their patients' family members, based on an open exchange of information and aimed at helping family members cope with their distress and allowing them to speak for the patient if necessary. We conducted a prospective multicenter study of family member satisfaction evaluated using the Critical Care Family Needs Inventory. Forty-three French ICUs participated in the study. ICU characteristics, patient and family member demographics, and data on satisfaction were collected. Factors associated with satisfaction were identified using a Poisson regression model. A total of 637 patients were included in the study, and 920 family members completed the questionnaire. Seven predictors of family satisfaction were found: one family-related factor, namely, family of French descent and six caregiver-related factors, namely, no perceived contradictions in information given by caregivers; information provided by a junior physician; patient to nurse ratio </= 3; knowledge of the specific role of each caregiver; help from the family's own doctor; and sufficient time spent giving information. Predictors of satisfaction are amenable to intervention and deserve to be investigated further with the goal of improving the satisfaction of ICU patients' family members.

Adult↗
PubMed2000

Infection as a trigger of diabetic ketoacidosis in intensive care-unit patients.

We determined the prevalence and indicators of infection in intensive care unit (ICU) patients with diabetic ketoacidosis (DKA) by performing a retrospective analysis of 123 episodes of DKA (in 113 patients) managed in a medical ICU between 1990 and 1997. In univariate analysis, features associated with infection were female sex, neurological symptoms at admission, fever during the week before admission, a need for colloids, a high blood lactate level at admission, and lack of complete clearance of ketonuria within 12 h. Multivariate analysis identified 3 independent predictors of infection: female sex (odds ratio [OR], 2.31; confidence interval [CI], 1.05-5.35), neurological symptoms at admission (OR, 2.83; CI, 1.18-6.8), and lack of complete clearance of ketonuria within 12 h (OR, 3.73; CI, 1.58-9.09). Infection is the leading trigger of DKA in ICU patients. Neurological symptoms at admission and lack of complete clearance of ketonuria within 12 h are useful warning signals of infection.

Adult↗
PubMed2000

Rapid diagnosis of infectious pleural effusions by use of reagent strips.

Reagent strips have not yet been tested for use in the diagnosis of infectious pleural effusions. A reagent strip was used to evaluate 82 patients with pleural effusions: 20 patients had transudative effusions, 35 had infectious exudative effusions (empyema in 14 and parapneumonic effusion in 21), and 27 had noninfectious exudative effusions. Pleural fluid protein, as evaluated by the reagent strip, proved accurate for the detection of exudative effusions (sensitivity, 93.1%; specificity, 50%; positive predictive value, 84.3%; negative predictive value, 71.5%; odds ratio [OR], 6.77; and 95% confidence interval [CI], 1.87-24). The reagent strip leukocyte esterase test effectively detected infectious exudative effusions (sensitivity, 42.8%; specificity, 91.3%; positive predictive value, 88.2%; negative predictive value, 51.2%; OR, 4.46; and 95% CI, 1.2-16.4). Pleural pH was significantly predicted by the reagent strip but was of no assistance in categorization of exudative effusions as infectious or noninfectious. Compared with physical, laboratory, and microbiological data, the reagent strip was as accurate for estimation of percentages of infectious and noninfectious exudative effusions. Thus, reagent strips may be a rapid, easy-to-use, and inexpensive technique for discriminating transudative from exudative pleural effusions and for categorizing exudative pleural effusions as infectious or noninfectious.

Adult↗
PubMed2000

Ratios of observed to expected mortality are affected by differences in case mix and quality of care.

OBJECTIVES: To validate SAPS II-AM, a recently customized version of the Simplified Acute Physiology Score II (SAPS II) in a larger cohort of Austrian intensive care patients and to evaluate the effect of the customization process on the ratio of observed to expected mortality. DESIGN: Prospective, multicentric cohort study. PATIENTS AND SETTING: A total of 2,901 patients consecutively admitted to 13 adult medical, surgical, and mixed intensive care units (ICUs) in Austria. MEASUREMENTS AND RESULTS: After the database was divided randomly into a development sample (n = 1,450) and a validation sample (n = 1,451), logistic regression was used to develop a new model (SAPS II-AM2). The original SAPS II, the SAPS IIAM, and the newly developed SAPS II-AM2 were then compared by means of calibration, discrimination and O/E ratios. Differences in O/E ratios before and after customization (deltaO/E) were calculated. The Hosmer-Lemeshow goodness-of-fit H and C statistics revealed poor calibration of the original SAPS II on the database. The new model, SAPS II-AM2, performed better than the SAPS II-AM and excellent in the validation data set. However, mean O/E ratios varied widely among diagnostic categories (range 0.55-1.05 for the SAPS II). Moreover, the deltaO/E of the 13 ICUs ranged from -3.6 % to +25 %. CONCLUSIONS: Today's severity scoring systems, such as the SAPS II, are limited by not measuring (and adjusting for) a profound part of what constitutes case mix. Changes in the distribution of patient characteristics (known and unknown) therefore affect prognostic accuracy. First-level customization was not able to solve all these problems. Using O/E ratios for quality of care comparisons one must therefore be critical when using these data and should search for possible confounding factors. In the case of unsatisfactory calibration, customized severity of illness models may be useful as an adjunct for quality control.

APACHE↗
PubMed2000

Predictors of short-term mortality in critically ill patients with solid malignancies.

UNLABELLED: Admission of cancer patients with serious medical complications to the ICU remains controversial primarily because of the high short-term mortality rates in these patients. However, the cancer patient population is heterogeneous regarding age, underlying conditions, and curability of their disease, suggesting that large variations may occur in the effectiveness of intensive care within this subgroup of critically ill patients. OBJECTIVES: To identify factors predicting 30-day mortality in patients with solid tumors admitted to a medical ICU. PATIENTS AND METHODS: We conducted a retrospective study in 120 consecutive cancer patients (excluding patients with hematological malignancies) admitted to the medical ICU of a 650-bed university hospital between January 1990 and July 1997. Medical history, physical and laboratory test findings at admission, and therapeutic interventions within the first 24 h in the ICU were recorded. The study endpoint was vital status 30 days after ICU admission. Stepwise logistic regression was used to identify independent prognostic factors. RESULTS: The observed 30-day mortality rate was 58.7 % (n = 68), with most deaths (92 %) occurring in the ICU. Univariate predictors of 30-day mortality were either protective [prior surgery for the cancer (p = 0.01) and complete remission (p = 0.01)] or associated with higher mortality [Knaus scale C or D (p = 0.02), shock (p = 0.04), need for vasopressors (p = 0.0006) or for mechanical ventilation (p = 0.0001), SAPS II score greater than 36 (p = 0.0001), LOD score greater than 6 (p = 0.0001), and ODIN score > 2 (p = 0.0001)]. Three variables were independent predictors: previous surgery for the cancer (OR 0.20, 95 % CI 0.07-0.58), LOD score > 6 (OR 1.26, 95 % CI 1.09-1.44), and need for mechanical ventilation (OR 3.55, 95 % CI; 1.26-6.7). Variables previously thought to be indicative of a poor prognosis (i. e., advanced age, metastatic or progressive disease, neutropenia or bone marrow transplantation) were not predictive of outcome. CONCLUSION: When transfer to an ICU is considered an option by patients and physicians, 30-day mortality is better estimated by an evaluation of acute organ dysfunction than by the characteristics of the underlying malignancy.

APACHE↗
PubMed2000

Half the families of intensive care unit patients experience inadequate communication with physicians.

OBJECTIVE: Effective communication of simple, clear information to families of intensive care unit (ICU) patients is a vital component of quality care. The purpose of this study was to identify factors associated with poor comprehension by family members of the status of ICU patients. DESIGN: Prospective study. SETTING: University-affiliated medical intensive care unit. PATIENTS AND METHODS: A total of 102 patients admitted to an ICU for >2 days. INTERVENTION: The representatives of 76 patients who were visited by at least one person during their ICU stay were interviewed. RESULTS: Mean patient age was 54+/-17 yrs and mean Simplified Acute Physiology Score II at admission was 40+/-20. The representative was the spouse in 47 cases (62%). Among representatives, 25 (33%) were of foreign descent and 12 (16%) did not speak French. Mean duration of the first meeting with a physician was 10+/-6 mins. In 34 cases (54%), the representative failed to comprehend the diagnosis, prognosis, or treatment of the patient. Factors associated with poor comprehension by representatives included patient-related, family-related, and physician-related factors. Patient-related factors included age <50 yrs (p = .03), unemployment (p = .01), referral from a hematology or oncology ward (p = .006), admission for acute respiratory failure (p = .005) or coma (p = .01), and a relatively favorable prognosis (p = .04). Family-related factors were foreign descent (p = .007), no knowledge of French (p = .03), representative not the spouse (p = .03), and no healthcare professional in the family (p = .01). Physician-related factors were first meeting with representative <10 mins (p = .03) and failure to give the representative an information brochure (p = .02). Moreover, after the first meeting, caregivers accurately predicted poor comprehension by representatives (p = .03). CONCLUSIONS: Patient information is frequently not communicated effectively to family members by ICU physicians. Physicians should strive to identify patients and families who require special attention and to determine how their personal style of interrelating with family members may impair communication.

Communication↗
PubMed2000

[The performance of intensive care services].

To evaluate the performance of intensive care unit (ICU) the severity scores are measured on the first day; organ system dysfunctions are measured several times during the stay. The severity scores are developed from large data bases of thousands of patients. They include the patient age, previous health status, severity and sometimes the main diagnosis. Prediction models are published to evaluate the risk of death for each patient. By adding the risk of death of each patient and dividing by the number of patients, the expected mortality rate is calculated. By comparing the observed and expected mortality rates the Standard Mortality Ratio (SMR) may be measured. The data collection must be rigorous, the studied population must be similar to the population of the large data bases. Other elements of performance may be evaluated, such as the cost-efficiency or the quality of life or surviving patients.

APACHE↗
PubMed1999

Changing use of intensive care for hematological patients: the example of multiple myeloma.

OBJECTIVE: Intensivists generally view patients with hematological malignancies as poor candidates for intensive care. Nevertheless, hematologists have recently developed more aggressive treatment protocols capable of achieving prolonged complete remissions in many of these patients. This change mandates a reappraisal of indications for ICU admission in each type of hematological disease. Improved knowledge of the prognosis is of assistance in making treatment decisions. PATIENTS AND METHODS: The records of 75 myeloma patients consecutively admitted to our ICU between 1992 and 1998 were reviewed retrospectively and predictors of 30-day mortality were identified using stepwise logistic regression. RESULTS: The median age was 56 years (37-84). Chronic health status (Knaus scale) was C or D in 39 cases. Fifty-five patients (73%) had stage III disease and 17 had a complete or partial remission. Autologous bone marrow transplantation had been performed in 28 patients (37%). ICU admission occurred between 1992 and 1995 in 41 patients (54.7%), and between 1996 and 1998 in 34 patients (45.3%). The median SAPS II and LOD scores were 60 (23-107) and 7 (0-21), respectively. Reasons for ICU admission were acute respiratory failure in 39 patients (52%) and shock in 31 (41%). Forty-six patients (61%) required mechanical ventilation. Fifty patients (66%) received vasopressors and 24 dialysis. Thirty-day mortality was 57%. Only five parameters were independently associated with 30-day mortality in the multivariate model: female gender (OR = 5.12), mechanical ventilation (OR = 16.7) and use of vasopressor agents (OR = 5.67) were associated with a higher mortality rate, whereas disease remission (OR = 0.16) and ICU admission between 1996 and 1998 (OR = 0.09) were associated with a lower one. CONCLUSION: The prognosis for myeloma patients in the ICU is improving over time. This may reflect either recent therapeutic changes in hematological departments and ICUs or changes in patient selection for ICU admission. Hematologists and intensivists should work closely together to select hematological patients likely to benefit from ICU admission.

Adult↗
PubMed1998

Estimation of direct cost and resource allocation in intensive care: correlation with Omega system.

OBJECTIVE: An instrument able to estimate the direct costs of stays in Intensive Care Units (ICUs) simply would be very useful for resource allocation inside a hospital, through a global budget system. The aim of this study was to propose such a tool. DESIGN: Since 1991, a region-wide common data base has collected standard data of intensive care such as the Omega Score, Simplified Acute Physiologic Score, length of stay, length of ventilation, main diagnosis and procedures. The Omega Score, developed in France in 1986 and proved to be related to the workload, was recorded on each patient of the study. SETTING: Eighteen ICUs of Assistance Publique-Hôpitaux de Paris (AP-HP) and suburbs. PATIENTS: 1) Hundred twenty-one randomly selected ICU patients; 2) 12,000 consecutive ICU stays collected in the common data base in 1993. MEASUREMENTS: 1) On the sample of 121 patients, medical expenditure and nursing time associated with interventions were measured through a prospective study. The correlation between Omega points and direct costs was calculated, and regression equations were applied to the 12,000 stays of the data base, leading to estimated costs. 2) From the analytic accounting of AP-HP, the mean direct cost per stay and per unit was calculated, and compared with the mean associated Omega score from the data base. In both methods a comparison of actual and estimated costs was made. RESULTS: The Omega Score is strongly correlated to total direct costs, medical direct costs and nursing requirements. This correlation is observed both in the random sample of 121 stays and on the data base' stays. The discrepancy of estimated costs through Omega Score and actual costs may result from drugs, blood product underestimation and therapeutic procedures not involved in the Omega Score. CONCLUSIONS: The Omega system appears to be a simple and relevant indicator with which to estimate the direct costs of each stay, and then to organise nursing requirements and resource allocation.

Critical Care↗
PubMed1997

[Pheochromocytoma mimicking septic shock].

A tumour arising from chromaffin cells, the pheochromocytoma may be associated or present with hypotension and result in a lethal shock. The authors report a rare case of pheochromocytoma simulating septic shock. They review the mechanisms by which hypersecretion of catecholamines may cause shock. The detection of pheochromocytoma in its atypical presentation is essential for early diagnosis and the reduction of mortality related to this condition.

Adrenal Gland Neoplasms↗
PubMed1996

The Logistic Organ Dysfunction system. A new way to assess organ dysfunction in the intensive care unit. ICU Scoring Group.

OBJECTIVE: To develop an objective method for assessing organ dysfunction among intensive care unit (ICU) patients on the first day of the ICU stay. DESIGN AND SETTING: Physiological variables defined dysfunction in 6 organ systems. Logistic regression techniques were used to determine severity levels and relative weights for the Logistic Organ Dysfunction (LOD) score and for conversion of the LOD score to a probability of mortality. PATIENTS: A total of 13 152 consecutive admission to 137 adult medical/surgical ICUs in 12 countries from the European/North American Study of Severity Systems. OUTCOME MEASURES: Patient vital status at hospital discharge. RESULTS: The LOD System identified from 1 to 3 levels of organ dysfunction for 6 organ systems: neurologic, cardiovascular, renal, pulmonary, hematologic, and hepatic. From 1 to 5 LOD points were assigned to the levels of severity, and the resulting LOD scores ranged from 0 to 22 points. Model calibration was very good in the developmental and validation samples (P=.21 and P=.50, respectively), as was model discrimination (area under the receiver operating characteristic curves of 0.843 and 0.850, respectively). CONCLUSION: The LOD System provides an objective tool for assessing severity levels for organ dysfunction in the ICU, a critical component in the conduct of clinical trials. Neurologic, cardiovascular, and renal dysfunction were the most severe organ dysfunctions, followed by pulmonary and hematologic dysfunction, with hepatic dysfunction the least severe. The LOD System takes into account both the relative severity among organ systems and the degree of severity within an organ system.

Adult↗
PubMed1996

Life expectancy in hospitalized patients with AIDS: prognostic factors on admission.

In the course of the inevitable hospitalizations of AIDS patients, many difficult questions concerning curative and palliative approaches to care have to be answered. In order to guide these strategies, we conducted a prospective study to identify those variables which can be easily quantified on admission which might be predictive of patient outcomes. Between 1 June 1990 and 25 April 1991, 140 consecutive hospitalizations of 83 AIDS patients were recorded. Demographic, clinical, and biological data were collected within 48 hours of admission. Probable (p <0.10) or definite (p <0.05) factors contributing to a higher mortality included type of opportunistic infections, serum albumin level, total lymphocyte count, weight, CD4 cell count, and neurological manifestations. In the multivariate proportional hazards model, two factors were significantly and independently predictive of lower survival: body weight less than 90% of ideal body weight and neurologic manifestations. The probability of survival was significantly affected by the number of predictive factors present on admission, and patients were significantly more likely to die when these latter two factors were present concomitantly. These factors might be useful to define the optimal mode of care for hospitalized AIDS patients, considering both patients' wishes and an objective assessment of the prognosis.

Acquired Immunodeficiency Syndrome↗
PubMed1995

Customized probability models for early severe sepsis in adult intensive care patients. Intensive Care Unit Scoring Group.

OBJECTIVE: To develop customized versions of the Simplified Acute Physiology Score II (SAPS II) and the 24-hour Mortality Probability Model II (MPM II) to estimate the probability of mortality for intensive care unit patients with early severe sepsis. DESIGN AND SETTING: Logistic regression models developed for patients with severe sepsis in a database of adult medical and surgical intensive care units in 12 countries. PATIENTS: Of 11,458 patients in the intensive care unit for at least 24 hours, 1130 had severe sepsis based on criteria of the American College of Chest Physicians and the Society of Critical Care Medicine (systemic inflammatory response syndrome in response to infection, plus hypotension, hypoperfusion, or multiple organ dysfunction). RESULTS: In patients with severe sepsis, mortality was higher (48.0% vs 19.6% among other patients) and 28-day survival was lower. The customized SAPS II was well calibrated (P = .92 for the goodness-of-fit test) and discriminated well (area under the receiver operating characteristic [ROC] curve, 0.78). Performance in the validation sample was equally good (P = .85 for the goodness-of-fit test; area under the ROC curve, 0.79). The customized MPM II was well calibrated (P = .92 for the goodness-of-fit test) and discriminated well (area under the ROC curve, 0.79). Performance in the validation sample was equally good (P = .52 for the goodness-of-fit test; area under the ROC curve, 0.75). The models are independent of each other; either can be used alone to estimate the probability of mortality of patients with severe sepsis. CONCLUSIONS: Customization provides a simple technique to apply existing models to a subgroup of patients. Accurately assessing the probability of hospital mortality is a useful adjunct for clinical trials.

Adult↗