Daily assessment of severity of illness and mortality prediction for individual patients

Trajectories of hepatic and coagulation dysfunctions related to a rapidly fatal outcome among hospitalized patients with dengue fever in Tainan, 2015

Background

Hepatic dysfunction and coagulopathy are common in acute dengue illness. We analyzed the trajectories of the above parameters in the survivors and fatal patients in the outbreak in Tainan, 2015.

Methods

A retrospective study was conducted using data from a tertiary hospital between January and December 2015. Multilevel modeling (MLM) was used to identify the changes in aminotransferase (AST), alanine aminotransferase (ALT), activated partial thromboplastin time (aPTT), and platelet counts from Day 0 to Day 7 of the onset of dengue infection. The machine-learning algorithm was used by purity measure assumption to calculate the accuracy of serum transaminases and coagulation variables to discriminate between the fatal and survival groups.

Results

There were 4,069 dengue patients, of which 0.9% died in one week after illness onset (i.e., early mortality). Case fatality rate was the highest for those aged ≥70 years. Both AST and ALT values of the fatal group were significantly higher than those of the survivor group from Day 3 (AST median, 624 U/L vs. 60 U/L, p < 0.001; ALT median, 116 U/L vs. 29 U/L, p = 0.01) of illness onset and peaked on Day 6 (AST median, 9805 U/L vs. 90 U/L, p < 0.001; ALT median, 1504 U/L vs. 49 U/L, p < 0.001). AST ≥ 203 U/L, ALT ≥ 55 U/L, AST²/ALT criteria ≥337.35, or AST/platelet count ratio index (APRI) ≥ 19.18 on Day 3 of dengue infection had a high true positive rate, 90%, 78%, 100%, or 100%, respectively, of early mortality. The platelet counts of the fatal group declined significantly than those of the survivor group since Day 3 of illness onset (median, 19 x10³/μl vs. 91 x10³/μl, p < 0.01), and aPTT values of the fatal group significantly prolonged longer since Day 5 (median, 68.7 seconds vs. 40.1 seconds, p < 0.001).

Conclusions

AST, ALT, and platelet counts should be monitored closely from Day 0 to Day 3 of dengue infection, and aPTT be followed up on Day 5 of infection to identify the individuals at risk for early mortality.

Dengue fever (DF) is currently one of the most severe public health problems. Clinical presentations of dengue are diverse and non-specific, often with unpredictable clinical progression and outcome. Hepatic dysfunction and abnormal coagulation factors are common in acute dengue illness, reflected by abnormal alanine aminotransferase (AST), aspartate aminotransferase (ALT), activated partial thromboplastin time (aPTT), and platelet counts. However, there is no information available about the monitoring frequency required, which could help identify those dengue patients who are likely to die, especially during epidemic outbreaks with limited healthcare resources. We examined all the laboratory-confirmed dengue patients who admitted to the major tertiary hospital in Tainan during the 2015 dengue outbreak, and the different trajectories of hepatic function and coagulation factors between survivors and rapidly fatal dengue patients were analyzed. Although there were no differences in AST, ALT, aPTT, and platelet counts between the survivor and fatal groups on the day DF symptoms first appeared, the differences increased from the early stages of infection and became more prominent during the early stages of the illness. The necessity of monitoring the AST, ALT, aPTT, and platelet count frequently during the febrile phase is emphasized by this study.

The value of daily platelet counts for predicting dengue shock syndrome: Results from a prospective observational study of 2301 Vietnamese children with dengue

BACKGROUND

Dengue is the most important mosquito-borne viral infection to affect humans. Although it usually manifests as a self-limited febrile illness, complications may occur as the fever subsides. A systemic vascular leak syndrome that sometimes progresses to life-threatening hypovolaemic shock is the most serious complication seen in children, typically accompanied by haemoconcentration and thrombocytopenia. Robust evidence on risk factors, especially features present early in the illness course, for progression to dengue shock syndrome (DSS) is lacking. Moreover, the potential value of incorporating serial haematocrit and platelet measurements in prediction models has never been assessed.

METHODOLOGY/PRINCIPAL FINDINGS

We analyzed data from a prospective observational study of Vietnamese children aged 5-15 years admitted with clinically suspected dengue to the Hospital for Tropical Diseases in Ho Chi Minh City between 2001 and 2009. The analysis population comprised all children with laboratory-confirmed dengue enrolled between days 1-4 of illness. Logistic regression was the main statistical model for all univariate and multivariable analyses. The prognostic value of daily haematocrit levels and platelet counts were assessed using graphs and separate regression models fitted on each day of illness. Among the 2301 children included in the analysis, 143 (6%) progressed to DSS. Significant baseline risk factors for DSS included a history of vomiting, higher temperature, a palpable liver, and a lower platelet count. Prediction models that included serial daily platelet counts demonstrated better ability to discriminate patients who developed DSS from others, than models based on enrolment information only. However inclusion of daily haematocrit values did not improve prediction of DSS.

CONCLUSIONS/SIGNIFICANCE

Daily monitoring of platelet counts is important to help identify patients at high risk of DSS. Development of dynamic prediction models that incorporate signs, symptoms, and daily laboratory measurements, could improve DSS prediction and thereby reduce the burden on health services in endemic areas.

Patient length of stay and mortality prediction: A survey

Over the past few years, there has been increased interest in data mining and machine learning methods to improve hospital performance, in particular hospitals want to improve their intensive care unit statistics by reducing the number of patients dying inside the intensive care unit. Research has focused on prediction of measurable outcomes, including risk of complications, mortality and length of hospital stay. The length of stay is an important metric both for healthcare providers and patients, influenced by numerous factors. In particular, the length of stay in critical care is of great significance, both to patient experience and the cost of care, and is influenced by factors specific to the highly complex environment of the intensive care unit. The length of stay is often used as a surrogate for other outcomes, where those outcomes cannot be measured; for example as a surrogate for hospital or intensive care unit mortality. The length of stay is also a parameter, which has been used to identify the severity of illnesses and healthcare resource utilisation. This paper examines a range of length of stay and mortality prediction applications in acute medicine and the critical care unit. It also focuses on the methods of analysing length of stay and mortality prediction. Moreover, the paper provides a classification and evaluation for the analytical methods of the length of stay and mortality prediction associated with a grouping of relevant research papers published in the years 1984 to 2016 related to the domain of survival analysis. In addition, the paper highlights some of the gaps and challenges of the domain.

Applicability of Prediction Rules in Patients With Community-Acquired Pneumonia Requiring Intensive Care: A Pilot Study

Little attention has been paid to developing prediction rules that could assist in deciding which patients with community-acquired pneumonia (CAP) need intensive care. Four existing prediction rules were examined to determine if any could predict the need for intensive care in these patients. The prediction rules studied were British Thoracic Society (BTS), Conte et al, Leroy et al, and Fine et al. Thirty-two patients admitted to the medical or coronary intensive care unit (ICU) during 1 year with pneumonia Diagnosis Related Group 079 or 089 were evaluated. The sensitivity of each rule for identifying a need for ICU admission in our group was BTS .72 using both rules together, Conte et al .47, Leroy et al .56, and Fine et al .84. It was concluded that these rules poorly identify the need for ICU admission for patients with severe CAP. Of the 4 rules tested, the BTS rule was the simplest, and the Fine et al rule was the most sensitive. None of them performed well enough to be used for decision making in individual patients.

Quadrimodal distribution of death after trauma suggests that critical injury is a potentially terminal disease

BACKGROUND

Patterns of death after trauma are changing due to advances in critical care. We examined mortality in critically injured patients who survived index hospitalization.

METHODS

Retrospective analysis of adults admitted to a Level-1 trauma center (1/1/2000-12/31/2010) with critical injury was conducted comparing patient characteristics, injury, and resource utilization between those who died during follow-up and survivors.

RESULTS

Of 1,695 critically injured patients, 1,135 (67.0%) were discharged alive. As of 5/1/2012, 977/1,135 (86.0%) remained alive; 75/158 (47.5%) patients who died during follow-up, died in the first year. Patients who died had longer hospital stays (24 vs. 17 days) and ICU LOS (17 vs. 8 days), were more likely to undergo tracheostomies (36% vs. 16%) and gastrostomies (39% vs. 16%) and to be discharged to rehabilitation (76% vs. 63%) or skilled nursing (13% vs. 5.8%) facilities than survivors. In multivariable models, male sex, older age, and longer ICU LOS predicted mortality. Patients with ICU LOS >16 days had 1.66 odds of 1-year mortality vs. those with shorter ICU stays.

CONCLUSIONS

ICU LOS during index hospitalization is associated with post-discharge mortality. Patients with prolonged ICU stays after surviving critical injury may benefit from detailed discussions about goals of care after discharge.

[Verification of validity of MPM II for neurological patients in intensive care units]

PURPOSE

Mortality Probability Model (MPM) II is a model for predicting mortality probability of patients admitted to ICU. This study was done to test the validity of MPM II for critically ill neurological patients and to determine applicability of MPM II in predicting mortality of neurological ICU patients.

METHODS

Data were collected from medical records of 187 neurological patients over 18 yr of age who were admitted to the ICU of C University Hospital during the period from January 2008 to May 2009. Collected data were analyzed through χ(2) test, t-test, Mann-Whiteny test, goodness of fit test, and ROC curve.

RESULTS

As to mortality according to patients' general and clinically related characteristics, mortality was statistically significantly different for ICU stay, hospital stay, APACHE III score, APACHE predicted death rate, GCS, endotracheal intubation, and central venous catheter. Results of Hosmer-Lemeshow goodness-of-fit test were MPM II(0) (χ(2)=0.02, p=.989), MPM II(24) (χ(2)=0.99 p=.805), MPM II(48) (χ(2)=0.91, p=.822), and MPM II(72) (χ(2)=1.57, p=.457), and results of the discrimination test using the ROC curve were MPM II(0), .726 (p<.001), MPM II(24), .764 (p<.001), MPM II(48), .762 (p<.001), and MPM II(72), .809 (p<.001).

CONCLUSION

MPM II was found to be a valid mortality prediction model for neurological ICU patients.

A predictive model for the early identification of patients at risk for a prolonged intensive care unit length of stay

BACKGROUND

Patients with a prolonged intensive care unit (ICU) length of stay account for a disproportionate amount of resource use. Early identification of patients at risk for a prolonged length of stay can lead to quality enhancements that reduce ICU stay. This study developed and validated a model that identifies patients at risk for a prolonged ICU stay.

METHODS

We performed a retrospective cohort study of 343,555 admissions to 83 ICUs in 31 U.S. hospitals from 2002-2007. We examined the distribution of ICU length of stay to identify a threshold where clinicians might be concerned about a prolonged stay; this resulted in choosing a 5-day cut-point. From patients remaining in the ICU on day 5 we developed a multivariable regression model that predicted remaining ICU stay. Predictor variables included information gathered at admission, day 1, and ICU day 5. Data from 12,640 admissions during 2002-2005 were used to develop the model, and the remaining 12,904 admissions to internally validate the model. Finally, we used data on 11,903 admissions during 2006-2007 to externally validate the model.

RESULTS

The variables that had the greatest impact on remaining ICU length of stay were those measured on day 5, not at admission or during day 1. Mechanical ventilation, PaO2: FiO2 ratio, other physiologic components, and sedation on day 5 accounted for 81.6% of the variation in predicted remaining ICU stay. In the external validation set observed ICU stay was 11.99 days and predicted total ICU stay (5 days + day 5 predicted remaining stay) was 11.62 days, a difference of 8.7 hours. For the same patients, the difference between mean observed and mean predicted ICU stay using the APACHE day 1 model was 149.3 hours. The new model's r2 was 20.2% across individuals and 44.3% across units.

CONCLUSIONS

A model that uses patient data from ICU days 1 and 5 accurately predicts a prolonged ICU stay. These predictions are more accurate than those based on ICU day 1 data alone. The model can be used to benchmark ICU performance and to alert physicians to explore care alternatives aimed at reducing ICU stay.

[Scoring systems for daily assessment in intensive care medicine. Overview, current possibilities and demands on new developments]

Scoring systems are a fixed element of modern diagnostics and are integrated in the diagnosis-related groups (DRG) billing system as well as quality assurance projects. The ongoing developments require classification according to the terms of use in order to maintain an overview of the numerous systems available. In the area of intensive care medicine scoring systems can be divided into admission scores and progress scores, whereby the scores for daily assessment can be further subdivided into five categories, depending on the target criteria: objective description of the grade of organ dysfunction, progression in intensive care therapy, evaluation of the degree of nursing care, determination of outcome/mortality risk, and grouping of patient collectives for clinical trials. In future developments it will be necessary to generate new strategies to adequately describe the progress of a patient. Not only will mortality be challenged as a target criterion but also the handling of missing data and the simplification of reality by categorization practised so far that can be found in all established scoring systems as far as calculation of predictive values regarding a defined result.

The logistic organ dysfunction score as a tool for making ethical decisions

PURPOSE

We examined whether the change of the logistic organ dysfunction score (LOD) between the first and the fourth day in the intensive care unit (ICU) could be predictive of death in the ICU. The LOD could then be used to help make decisions concerning therapeutic limitations (TL).

METHODS

One hundred fifty-four patients were included. Exclusion criteria were: discharge from the ICU or TL before the 72nd hr. Ninety-three patients remained for evaluation. The LOD was calculated on the day of admission (LOD1) and between the 72nd and 96th hr (LOD4). The DeltaLOD = LOD4-LOD1 index was calculated for survivors and non-survivors; sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated.

RESULTS

Sixteen patients died in the ICU, they had a higher DeltaLOD (0 vs -2; P = 0.0046) than the survivors. After logistic regression, a high DeltaLOD was associated with a higher risk of death in the ICU independent of the initial severity of disease. The PPV concerning death in the ICU was 0.66 for a DeltaLOD > or = 4 cut-off. The NPV was 0.89 for a cut-off of > or = 1.

CONCLUSION

DeltaLOD appears to be a predictor of death in the ICU, independent of the initial severity of disease. The PPV is not high enough to assist with making individual TL decisions. The NPV can help to identify patients at low risk of death. The DeltaLOD deserves to be evaluated in a population exhibiting greater severity of disease.

Excess risk of death from intensive care unit-acquired nosocomial bloodstream infections: a reappraisal

BACKGROUND

Overall rates of bloodstream infection (BSI) are often used as quality indicators in intensive care units (ICUs). We investigated whether ICU-acquired BSI increased mortality (by > or = 10%) after adjustment for severity of infection at ICU admission and during the pre-BSI stay.

METHODS

We conducted a matched, risk-adjusted (1:n), exposed-unexposed study of patients with stays longer than 72 h in 12 ICUs randomly selected from the Outcomerea database.

RESULTS

Patients with BSI after the third ICU day (exposed group) were matched on the basis of risk-exposure time and mortality predicted at admission using the Three-Day Recalibrated ICU Outcome (TRIO) score to patients without BSI (unexposed group). Severity was assessed daily using the Logistic Organ Dysfunction (LOD) score. Of 3247 patients with ICU stays of >3 days, 232 experienced BSI by day 30 (incidence, 6.8 cases per 100 admissions); among them, 226 patients were matched to 1023 unexposed patients. Crude hospital mortality was 61.5% among exposed and 36.7% among unexposed patients (P<.0001). Attributable mortality was 24.8%. The only variable associated with both BSI and hospital mortality was the LOD score determined 4 days before onset of BSI (odds ratio [OR], 1.10; 95% confidence interval [CI], 1.03-1.16; P = .0025). The adjusted OR for hospital mortality among exposed patients (OR, 3.20; 95% CI, 2.30-4.43) decreased when the LOD score determined 4 days before onset of BSI was taken into account (OR, 3.02; 95% CI, 2.17-4.22; P<.0001). The estimated risk of death from BSI varied considerably according to the source and resistance of organisms, time to onset, and appropriateness of treatment.

CONCLUSIONS

When adjusted for risk-exposure time and severity at admission and during the ICU stay, BSI was associated with a 3-fold increase in mortality, but considerable variation occurred across BSI subgroups. Focusing on BSI subgroups may be valuable for assessing quality of care in ICUs.

Use of the Sequential Organ Failure Assessment score as a severity score

OBJECTIVE

To evaluate whether the SOFA score can be used to develop a model to predict intensive care unit (ICU) mortality in different countries.

DESIGN AND SETTING

Analysis of a prospectively collected database. Patients with ICU stay longer than 2 days were studied to develop a mortality prediction model based on measurements of organ dysfunction.

PATIENTS

748 patients from six countries.

MEASUREMENTS AND RESULTS

Two logistic regression models were constructed, one based on the SOFA maximum (SOFA Max model) and the other on variables identified by multivariate regression (SOFA Max-infection model). The H and C statistics had a p value above 0.05 for both models, but the D statistics showed a poor performance on the SOFA Max model when stratified for the presence of infection. Subsequent analysis was performed with SOFA Max-infection model. The area under the curve was 0.853. There were no statistically significant differences in observed and predicted mortalities except for one country which had a higher than predicted ICU mortality both in the overall population (28.3 vs. 19.1%) and in the noninfected patients (21.4 vs. 12.6%).

CONCLUSIONS

The SOFA Max adjusted for age and the presence of infection can predict mortality in this population, but in one country the ICU mortality was higher than expected. Our data do not allow us to determine the reasons behind these differences, and further studies to detect differences in mortality between countries and to elucidate the basis for these differences should be encouraged.

Calibration and discrimination by daily Logistic Organ Dysfunction scoring comparatively with daily Sequential Organ Failure Assessment scoring for predicting hospital mortality in critically ill patients

OBJECTIVE

The Logistic Organ Dysfunction (LOD) score has been proved effective in evaluating severity during the first day in an intensive care unit but has not been evaluated later. To evaluate attributable mortality related to nosocomial events, organ dysfunction scores that remain accurate throughout the intensive care unit stay are needed. The objective of this study was to evaluate how accurately daily LOD scoring predicts mortality comparatively with daily Sequential Organ Failure Assessment (SOFA) scoring.

DESIGN

Prospective multicenter study.

SETTING

Six intensive care units in France.

PATIENTS

A total of 1685 patients with intensive care unit stays longer than 48 hrs were included in this study (511 hospital deaths). Median age was 66 yrs, and median Simplified Acute Physiology Score II at admission was 38. For each patient, a senior physician recorded the variables needed to compute organ dysfunction scores daily throughout the intensive care unit stay.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

SOFA and LOD scores were computed daily during the first 7 days. Calibration was evaluated based on goodness-of-fit by the Hosmer-Lemeshow chi-square statistic (lower chi-square values and higher values indicate better fit) and discrimination based on the receiver operating characteristics (ROC) area under the curve (AUC; a ROC-AUC of 1 indicates faultless discrimination and a ROC-AUC of 0.5 indicates the effects of chance alone). Because calibration of both scores was poor at all time points ( p<.001), customization was performed using the total score (model 1) or separate introduction of each dysfunction (model 2). The performance of customized LOD and SOFA scores on a given day in predicting mortality was assessed in those patients who spent at least one more calendar day in the intensive care unit. The original LOD and SOFA scores had satisfactory ROC-AUC values (0.720 to 0.766). Internal consistency of both scores was acceptable ( p< 10(-4) for each organ dysfunction). After customization, the original scores calibrated well between days 1 and 7. Discrimination by both scores was better with model 2 (AUC-ROC, 0.729-0.784).

CONCLUSION

Daily LOD and SOFA scores showed good accuracy and internal consistency, and they could be used to adjust severity for events occurring in the intensive care unit.

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.

New issues in severity scoring: interfacing the ICU and evaluating it

Since the development of the first general outcome prediction models, these instruments have been widely used in the intensive care unit (ICU), both for patient evaluation and for ICU evaluation. Since some of these uses have been serious questioned, we assisted in the last years to the exploration of alternative paths for increasing the predictive power of the models and to enhance their applicability and utility in the real world. Part of these efforts focused on the exploration of more meaningful outcomes (clinical and non-clinical) with a strong tonic into the relation between outcomes and resources use. Also, since it is now widely recognized that the ICU is not an island, but it is integrated in a continuum of care, more and more efforts are being made to optimize and evaluate the interface between the ICU and the hospital, both at ICU admission and at ICU discharge. The objective of this review is to present and discuss, to the clinician working in the ICU, these emerging issues.