Decision support issues using a physiology based score

The Various Scoring Systems in Pediatric Intensive Care Units: A Prospective Observational Study

BACKGROUND

The discrimination power of the pediatric risk of mortality (PRISM), pediatric index of mortality (PIM), sequential organ failure assessment (SOFA), and pediatric logistic organ dysfunction (PELOD) may not always be true for countries such as India due to differences in factors from those nations where these scoring systems were validated. Therefore, this study was undertaken to determine and compare severity, course of illness, and outcomes in critically ill children admitted to the pediatric intensive care unit (PICU) using different scoring systems such as PRISM 4, PIM 3, PELOD 2, and the pediatric sequential organ failure assessment (pSOFA ) score, and to analyze the clinical spectrum and demographic profile of children admitted to the PICU.

MATERIALS AND METHOD

This was a prospective, single-center, observational study conducted in the PICU of the Indira Gandhi Institute of Medical Science, Patna, India, over two years. Two hundred children in the age group of one month to 14 years admitted to the PICU were recruited into the study. Prognostic scoring systems, including PRISM4 and PIM3, were used to compare the outcome, mortality, and length of PICU stay, whereas PELODS and pSOFA were descriptive scores that assessed the multiorgan dysfunction. A correlation between the different scoring systems and the outcome was determined.

RESULTS

The majority of children (26.5%, n=53) were one to three years of age. The maximum number of patients was male (66.5%, n=133). Renal complications were the predominant admission diagnosis in 19% (n=38) of children. The mortality rate was found to be 18.5%. The mortality was most common in infants <1 year of age (n=11, 29.73%) and those of the male sex (n=22, 59.46%). A significant correlation was found between length of stay and mortality (p<0.00001). A significant positive correlation was observed between mortality and PRISM 4, PIM 3, PELOD 2, and pSOFA scores on the first day of admission (p<0.00001). The pSOFA and PELOD2 showed better discrimination power (area under the curve (AUC): 0.77 and 0.74, respectively).

CONCLUSION

The study concluded that the pSOFA and PELOD2 scores are reliable predictors of mortality in critically ill children.

Comparison of pediatric scoring systems for mortality in septic patients and the impact of missing information on their predictive power: a retrospective analysis

Background

Scores can assess the severity and course of disease and predict outcome in an objective manner. This information is needed for proper risk assessment and stratification. Furthermore, scoring systems support optimal patient care, resource management and are gaining in importance in terms of artificial intelligence.

Objective

This study evaluated and compared the prognostic ability of various common pediatric scoring systems (PRISM, PRISM III, PRISM IV, PIM, PIM2, PIM3, PELOD, PELOD 2) in order to determine which is the most applicable score for pediatric sepsis patients in terms of timing of disease survey and insensitivity to missing data.

Methods

We retrospectively examined data from 398 patients under 18 years of age, who were diagnosed with sepsis. Scores were assessed at ICU admission and re-evaluated on the day of peak C-reactive protein. The scores were compared for their ability to predict mortality in this specific patient population and for their impairment due to missing data.

Results

PIM (AUC 0.76 (0.68-0.76)), PIM2 (AUC 0.78 (0.72-0.78)) and PIM3 (AUC 0.76 (0.68-0.76)) scores together with PRSIM III (AUC 0.75 (0.68-0.75)) and PELOD 2 (AUC 0.75 (0.66-0.75)) are the most suitable scores for determining patient prognosis at ICU admission. Once sepsis is pronounced, PELOD 2 (AUC 0.84 (0.77-0.91)) and PRISM IV (AUC 0.8 (0.72-0.88)) become significantly better in their performance and count among the best prognostic scores for use at this time together with PRISM III (AUC 0.81 (0.73-0.89)). PELOD 2 is good for monitoring and, like the PIM scores, is also largely insensitive to missing values.

Conclusion

Overall, PIM scores show comparatively good performance, are stable as far as timing of the disease survey is concerned, and they are also relatively stable in terms of missing parameters. PELOD 2 is best suitable for monitoring clinical course.

Can Pediatric Risk of Mortality Score (PRISM III) Be Used Effectively in Initial Evaluation and Follow-up of Critically Ill Cancer Patients Admitted to Pediatric Oncology Intensive Care Unit (POICU)? A Prospective Study, in a Tertiary Cancer Center in Egypt

Pediatric Risk of Mortality Score (PRISM III-12) is a physiology-based predictor for risk of mortality. We conducted prospective study from January 1, 2014 to 2015 in pediatric oncology intensive care unit (POICU) at South Egypt Cancer Institute, Egypt to explore the ability of 1st PRISM III-12 to predict the risk of mortality in critically ill cancer patients and the ability of serial PRISM III measured every 72 hours to follow-up the patients' clinical condition during POICU stay. In total, 123 (78 males) children were included. Median age was 5 years (1 to 15 y). Death rate was 20%. 1st PRISM III-12 mean was 19 (0 to 61). The mean 1st PRISM III-12 for survivors was significantly higher compared with nonsurvivors (15 vs. 37 respectively; P<0.001). 1st PRISM III-12 mean was significantly correlated to the reasons for admission and organ failures' number (P<0.001 and <0.001). 1st PRISM III-12 correlated weakly positive with the length of stay (r=0.2; P=0.024). Receiver operator curve for 1st PRISM III-12 was 0.913 (95% confidence interval, 0.85-0.98; P<0.001). Decline in serial PRISM III was significantly correlated with favorable (survivor) outcome (P<0.001). We concluded that PRISM III-12 can be used effectively in predicting the risk of mortality and following the clinical condition of patients during POICU stay.

Development and Validation of the Neonatal Risk Estimate Score for Children Using Extracorporeal Respiratory Support

OBJECTIVE

To develop and validate the Neonatal Risk Estimate Score for Children Using Extracorporeal Respiratory Support, which estimates the risk of in-hospital death for neonates prior to receiving respiratory extracorporeal membrane oxygenation (ECMO) support.

STUDY DESIGN

We used an international ECMO registry (2008-2013); neonates receiving ECMO for respiratory support were included. We divided the registry into a derivation sample and internal validation sample, by calendar date. We chose candidate variables a priori based on published evidence of association with mortality; variables independently associated with mortality in logistic regression were included in this parsimonious model of risk adjustment. We evaluated model discrimination with the area under the receiver operating characteristic curve (AUC), and we evaluated calibration with the Hosmer-Lemeshow goodness-of-fit test.

RESULTS

During 2008-2013, 4592 neonates received ECMO respiratory support with mortality of 31%. The development dataset contained 3139 patients treated in 2008-2011. The Neo-RESCUERS measure had an AUC of 0.78 (95% CI 0.76-0.79). The validation cohort had an AUC = 0.77 (0.75-0.80). Patients in the lowest risk decile had an observed mortality of 7.0% and a predicted mortality of 4.4%, and those in the highest risk decile had an observed mortality of 65.6% and a predicted mortality of 67.5%.

CONCLUSIONS

Neonatal Risk Estimate Score for Children Using Extracorporeal Respiratory Support offers severity-of-illness adjustment for neonatal patients with respiratory failure receiving ECMO. This score may be used to adjust patient survival to assess hospital-level performance in ECMO-based care.

Performance of PRISM III and PELOD-2 scores in a pediatric intensive care unit

The study aims were to compare two models (The Pediatric Risk of Mortality III (PRISM III) and Pediatric Logistic Organ Dysfunction (PELOD-2)) for prediction of mortality in a pediatric intensive care unit (PICU) and recalibrate PELOD-2 in a Portuguese population. To achieve the previous goal, a prospective cohort study to evaluate score performance (standardized mortality ratio, discrimination, and calibration) for both models was performed. A total of 556 patients consecutively admitted to our PICU between January 2011 and December 2012 were included in the analysis. The median age was 65 months, with an interquartile range of 1 month to 17 years. The male-to-female ratio was 1.5. The median length of PICU stay was 3 days. The overall predicted number of deaths using PRISM III score was 30.8 patients whereas that by PELOD-2 was 22.1 patients. The observed mortality was 29 patients. The area under the receiver operating characteristics curve for the two models was 0.92 and 0.94, respectively. The Hosmer and Lemeshow goodness-of-fit test showed a good calibration only for PRISM III (PRISM III: χ (2) = 3.820, p = 0.282; PELOD-2: χ (2) = 9.576, p = 0.022).

CONCLUSIONS

Both scores had good discrimination. PELOD-2 needs recalibration to be a better reliable prediction tool.

WHAT IS KNOWN

• PRISM III (Pediatric Risk of Mortality III) and PELOD (Pediatric Logistic Organ Dysfunction) scores are frequently used to assess the performance of intensive care units and also for mortality prediction in the pediatric population. • Pediatric Logistic Organ Dysfunction 2 is the newer version of PELOD and has recently been validated with good discrimination and calibration. What is New: • In our population, both scores had good discrimination. • PELOD-2 needs recalibration to be a better reliable prediction tool.

Power and limitations of daily prognostications of death in the medical ICU for outcomes in the following 6 months

OBJECTIVES

We tested the power of clinicians' predictions that a medical ICU patient would "die before hospital discharge" for both survival to discharge and for outcomes at 6 months.

DESIGN

We restricted our analyses to patients who had been in the medical ICU at least 72 hours and for whom we had follow-up at 6 months after medical ICU admission. For 350 medical ICU patients, on each medical ICU day, we asked their attending physician, fellow, resident, and primary nurse one question-"do you think this patient will die in hospital or survive to be discharged"? We correlated these responses with 6-month outcomes (death and/or Barthel score for survivors).

RESULTS

We obtained over 6,000 predictions on 2,271 medical ICU patient-days. Of 350 medical ICU patients who stayed more than 72 hours, 143 patients (41%) had discordant predictions-that is, on the same medical ICU day, at least one provider predicted survival, whereas another predicted death before discharge. As we have shown previously, predictions of "death before discharge" were imperfect-only 104 of 187 of patients with a prediction of death (56%) actually died in hospital. However, this is the central finding of our study, and predictions of death before discharge were much more accurate for 6-month outcomes. Of 120 patients with a corroborated prediction of death before discharge (93%), 112 patients had died within 6 months of medical ICU discharge, and only 4% were functioning with a Barthel score more than 70. In contrast, 67 of 163 patients who did not have any prediction of death before discharge (41%) were alive with Barthel score more than 70 at 6 months.

CONCLUSIONS

Fewer than 4% of medical ICU patients who required 72 hours of medical ICU care and had a corroborated prediction of death before discharge were alive at 6 months and functioning with a Barthel score more than 70.

Development of a daily mortality probability prediction model from Intensive Care Unit patients using a discrete-time event history analysis

As studies have pointed out, severity scores are imperfect at predicting individual clinical chance of survival. The clinical condition and pathophysiological status of these patients in the Intensive Care Unit might differ from or be more complicated than most predictive models account for. In addition, as the pathophysiological status changes over time, the likelihood of survival day by day will vary. Actually, it would decrease over time and a single prediction value cannot address this truth. Clearly, alternative models and refinements are warranted. In this study, we used discrete-time-event models with the changes of clinical variables, including blood cell counts, to predict daily probability of mortality in individual patients from day 3 to day 28 post Intensive Care Unit admission. Both models we built exhibited good discrimination in the training (overall area under ROC curve: 0.80 and 0.79, respectively) and validation cohorts (overall area under ROC curve: 0.78 and 0.76, respectively) to predict daily ICU mortality. The paper describes the methodology, the development process and the content of the models, and discusses the possibility of them to serve as the foundation of a new bedside advisory or alarm system.

Are all ICUs the same?

The ability to compare intensive care units (ICUs) and determine whether they provide the same level of care with regard to efficacy, efficiency, and quality is a cornerstone of understanding critical care and improving the quality of that care. Without collecting high-quality data, adjusted for severity of illness and analyzed in a comparative fashion, it would not be possible to describe best practices objectively, to identify which ICUs are doing a good job or to learn from those units that are. This review article discusses how and why ICUs are compared. Particular attention is focused on the severity of illness scores, standardized mortality, and comparative reporting. A data collecting network, Virtual Pediatric Systems, limited liability corporation (VPS, LLC), designed for the purposes of determining where differences in critical care can be identified and the value that this adds in improving quality is discussed. Finally, results from this large data sharing collaborative describing the practice of pediatric critical care are included for the purpose of pediatric intensive care units practice benchmarks.

Trauma intensive care unit survival: how good is an educated guess?

BACKGROUND

Resource utilization in medicine is becoming a more and more urgent issue with ongoing national discussions on healthcare coverage. In the management of a trauma system, large amounts of resources and money are expended on individual patients in hope of a "great save." In addition, those of us caring for these patients are required to estimate outcomes daily to the family in an effort to choose the best course of care for an individual patient. Hence, we undertook a study to analyze the accuracy of outcomes predictions of various members of the healthcare team.

METHODS

During a period of 38 months (July 2005 to August 2008), an observational study of patients admitted to a Level I Trauma Center Intensive Care Unit (ICU) was undertaken. Institutional Review Board permission was obtained before starting the study. Only patients older than 18 years were included. Patients who were moribund or expected discharge within 72 hours were excluded.Our traumatized ICU patients are cared for by a multidisciplinary team consisting of a trauma/ICU attending, all of whom have additional certification in surgical critical care and who rotate through the ICU on a weekly basis, a surgical ICU fellow, residents and medical students of several levels of training who rotate on a monthly basis, trauma advanced-level practitioners who rotate weekly, and bedside ICU nurses who work routine shifts. Respiratory therapists, nutritionists, ICU pharmacists, and other members of the rounding team were not included in the study because they do not provide global patient care. Regardless of admitting physician, the patients are managed by the team, and our practice of care is similar across the group, based on protocols and consensus.For each of the study patients, a survey tool was filled out by the ICU rounding team on hospital day 1 and hospital day 3. The tool was completed by members of the team providing global care to the patient and varied depending on the members of the group at each day's rounds. All current and admission data on injuries, study and laboratory results, and current patient status were available to all members of the team. Each member was expected to fill out the survey tool independently, and the results of the tool were not discussed during rounds.Concurrently, data were collected by the ICU fellow and research nurse. These data and the results of the survey tools were entered in a database for analysis after patient discharge. A retrospective analysis was undertaken to analyze the relative accuracy of the care, team members' assessment, and actual survival. Statistical analysis was done using by-chance accuracy comparisons.

RESULTS

Two hundred twenty-three patients had 326 observations performed. Day 3 accuracy improved for most groups. In all groups, accuracy was found to be statistically significantly better than by-chance accuracy. Given that the majority of patients in the trauma population are survivors, sensitivity and positive predictive value of the observer's ability to predict death were also evaluated.

CONCLUSIONS

Although significantly better than chance prediction, the ability of members of the ICU team to predict survival of trauma patients remains poor, particularly on initial evaluation. A period of clinical observation improves the accuracy. Unfortunately, experience of the observer does not seem to improve accuracy of survival prediction. This data indicate that care must be taken when describing likely outcomes to patient family members.

Admission chemokine (C-C motif) ligand 4 levels predict survival in pediatric septic shock

OBJECTIVE

Stratification with an effective outcome biomarker could improve the design of interventional trials in pediatric septic shock. The objective of this study was to test the usefulness of chemokine (C-C motif) ligand 4 as an outcome biomarker for mortality in pediatric septic shock.

DESIGN

A cross-sectional, observational study.

SETTING

Eighteen pediatric intensive care units in the United States.

PATIENTS

One hundred fifty-six pediatric patients with septic shock.

INTERVENTIONS

Serum samples were obtained within 24 hrs of admission to the pediatric intensive care unit. Serum levels of chemokine (C-C motif) ligand 4 were measured by enzyme-linked immunosorbent assay and compared with mortality in a training set of 34 patients. These data were used to generate a cutoff value whose usefulness was evaluated through prospective application-without post hoc modification-to a larger validation set of 122 patients.

MEASUREMENTS AND MAIN RESULTS

On inspection of the training set data, a cutoff value of 140 pg/mL was chosen. When applied to the validation set, serum chemokine (C-C motif) ligand 4 levels >140 pg/mL yielded a sensitivity of 92% and a specificity of 40% for mortality. A serum level of < or =140 pg/mL had a negative predictive value for mortality of 98%.

CONCLUSIONS

A serum level of chemokine (C-C motif) ligand 4 of < or =140 pg/mL, when obtained within 24 hrs of admission, predicts a very high likelihood of survival in pediatric septic shock. Exclusion of patients with a chemokine (C-C motif) ligand 4 level of < or =140 pg/mL from interventional clinical trials in pediatric septic shock could create a study population in which survival benefit from the study agent could be more readily demonstrated.

Performance of PRISM (Pediatric Risk of Mortality) score and PIM (Pediatric Index of Mortality) score in a tertiary care pediatric ICU

OBJECTIVE

To validate Pediatric Risk of Mortality (PRISM) and Pediatric Index of Mortality (PIM) score.

METHODS

All consecutive patients over a six month period were included in the study except patients with a PICU stay of less than 2 hours, those transferred to other PICUs, pediatric surgical cases, trauma patients and those dying within 24 hours of admission. The PRISM and PIM scores of all patients included in the study were computed and the outcome was noted in terms of survival or non-survival. Mortality discrimination was quantified by calculating the area under the receiver operating characteristic (ROC) curve. Hosmer and Lemeshow goodness-of-fit test was used to calibrate the scores.

RESULTS

Two hundred and thirty patients were enrolled with mean age of 40.6 months and male to female ratio of 1.2:1. There were 56 deaths (mortality rate 24.3%). The mortality in infants was higher (37.8 %) as compared to non-infants (16.2 %) (p = 0.011). The predicted deaths with PRISM score was 24.3%. The area under the ROC curve was 0.851 (95% CI 0.790-0.912). The Hosmer and Lemeshow goodness-of-fit test showed good calibration (p = 0.627, chi square = 1.75, degree of freedom = 3). The predicted deaths with the PIM score was 7.38%. The area under the ROC curve for PIM score was 0.838 (95 % CI 0.776-0.899). The Hosmer and Lemeshow goodness-of-fit showed a poor calibration for PIM score (p = 0.0281, chi-square = 10.866, degree of freedom = 4).

CONCLUSION

Both PRISM and PIM scores have a good discriminatory performance. The calibration with PRISM score is good but the PIM score displays poor calibration.

Just, in time: ethical implications of serial predictions of death and morbidity for ventilated premature infants

OBJECTIVES

For a cohort of extremely premature, ventilated, newborn infants, we determined the power of either serial caretaker intuitions of "die before discharge" or serial illness severity scores to predict the outcomes of death in the NICU or neurologic performance at corrected age of 2 years.

METHODS

We identified 268 premature infants who were admitted to our NICU in 1999-2004 and required mechanical ventilation. For each infant on each day of mechanical ventilation, we asked nurses, residents, fellows, and attending physicians the following question: "Do you think this child is going to live to go home or die before hospital discharge?" In addition, we calculated illness severity scores until either death or extubation.

RESULTS

A total of 17,066 intuition profiles were obtained on 5609 days of mechanical ventilation in the NICU. One hundred (37%) of 268 profiled infants had > or = 1 intuition of die before discharge. Only 33 infants (33%) with an intuition of die actually died in the NICU. Of 48 infants with even 1 day of corroborated intuition of die in the NICU, only 7 (14%) were alive with both Mental Developmental Index and Psychomotor Developmental Index scores of > 69, and only 2 (4%) were alive with both Mental Developmental Index and Psychomotor Developmental Index Scores of > 79 at corrected age of 2 years. On day of life 1, the Score for Neonatal Acute Physiology II value for nonsurvivors (38.2 +/- 18.1) was significantly higher than that for survivors (26.3 +/- 12.7). However, this difference decreased steadily over time as scores improved for both groups.

CONCLUSIONS

Illness severity scores become progressively less helpful over time in distinguishing infants who will either die in the NICU or survive with low Mental Developmental Index/Psychomotor Developmental Index scores. Serial caretaker intuitions of die before discharge also fail to identify prospective nonsurviving infants. However, corroborated intuitions of die before discharge identify a subset of infants whose likelihood of surviving to 2 years with both MDI and PDI > 80 is approximately 4%.

Review of the acuity scoring systems for the pediatric intensive care unit and their use in quality improvement

Acuity scoring systems quantitate the severity of clinical conditions and stratify patients according to presenting patient condition. In the pediatric intensive care unit, the complexity and number of clinical scoring systems are increasing as their applications for clinicians, health services researches, and quality improvement broaden. This article is a review of acuity scoring systems for the pediatric intensive care unit, including examples of scoring systems available, the methods used in assessing these tools, the ways in which these systems are used, and the utility of acuity scoring systems in accurate benchmarking. It is anticipated that with increasing health care costs and competition and increased focus on medical error reduction and quality improvement, the demands for risk-adjusted outcomes and institutional benchmarking will increase; therefore, as clinicians, academicians, and administrators, it is imperative that we be knowledgeable of the methods and applications of these acuity scoring systems to ensure their quality and appropriate use.

Simplified PRISM III score and outcome in the pediatric intensive care unit

BACKGROUND

To evaluate the association of the PRISM III (pediatric risk of mortality) score with the infant outcome in the pediatric intensive care unit (PICU), and to determine if this score could be simplified.

METHODS

A prospective cohort study was carried out with 170 infants who were consecutively admitted to the PICU. The PRISM III score with 17 physiologic variables was performed during the first 8 h of admission to the unit. Statistical analysis was done with logistic regression, odds ratios (OR) with 95% confidence intervals (95% CI), and receiver operating curve. The Alfa value was set at 0.05.

RESULTS

There were 42 deaths (24.7%). The two main causes of death were septic shock (28.6%) and head trauma (16.7%). The PRISM III score had a sensitivity of 0.71, and a specificity of 0.64 as a mortality predictor. Out of the 17 physiologic variables only four of them were significant: abnormal pupillary reflexes OR 9.9 (95% CI, 3.5-28.4), acidosis OR 3.1 (95% CI, 2.0-4.9), blood urea nitrogen concentration OR 1.03 (95% CI, 1.01-1.04), and white blood cell count OR 1.02 (95% CI, 1.01-1.03). The whole logistic regression model had a coefficient of determination R(2) = 0.219, P < 0.001.

CONCLUSIONS

In this setting, the PRISM III score had good sensitivity and specificity to predict mortality. This score could be simplified using only the four variables that were significant in this study. This modified PRISM III score could reduce the cost of patient care especially in developing countries PICU.

Acute severe asthma: Outcome and Medicaid insurance

OBJECTIVE: Economically disadvantaged children receive less preventive asthma care and more inpatient care. Studies have not evaluated the association of insurance status on children with severe exacerbations. We evaluated differences in severity of illness, resource use, and outcome associated with Medicaid insurance among children receiving intensive care for asthma. DESIGN: Retrospective cohort study. SETTING: Fourteen American pediatric intensive care units participating in the Pediatric Intensive Care Evaluations database. Methods: Patients with a primary diagnosis of asthma treated from May 1995 to February 2000 were identified. Demographic information and clinical data were evaluated to determine whether there was an association between Medicaid insurance, severity of illness, and length of stay. RESULTS: Twenty-six percent of the children had Medicaid insurance; 22% of children with Medicaid insurance received mechanical ventilation compared with 15% of those with commercial insurance and 16% in a health maintenance organization. After adjustment for severity of illness (Pediatric Risk of Mortality III and use of invasive therapies), Medicaid insurance was significantly associated with increased length of stay in the intensive care unit and hospital. Among children who received mechanical ventilation, patients with Medicaid also received ventilator support significantly longer. CONCLUSIONS: Asthmatic children receiving Medicaid had longer pediatric intensive care unit and hospital stays and an increased risk of mechanical ventilation compared with asthmatic children with commercial or health maintenance organization insurance. Further studies are needed to evaluate differences in outcome and resource utilization for economically disadvantaged asthmatic children.

Serial assessment of mortality in the neonatal intensive care unit by algorithm and intuition: certainty, uncertainty, and informed consent

OBJECTIVES

Does predictive power for outcomes of neonatal intensive care unit (NICU) patients get better with time? Or does it get worse? We determined the predictive power of Score for Neonatal Acute Physiology (SNAP) scores and clinical intuitions as a function of day of life (DOL) for newborn infants admitted to our NICU.

METHODS

We identified 369 infants admitted to our NICU during 1996-1997 who required mechanical ventilation. We calculated SNAP scores on DOL 1, 3, 4, 5, 7, 10, 14, 21, 28, and weekly thereafter until either death or extubation. We also asked nurses, residents, fellows, and attendings on each day of mechanical ventilation: "Do you think this child is going to live to go home to their family, or die before hospital discharge?"

RESULTS

Two thousand twenty-eight SNAP scores were calculated for 285 infants. On DOL 1, SNAP for nonsurvivors (24 +/- 8.7 [standard deviation]) was significantly higher than SNAP for survivors (13 +/- 6.1). However, this difference diminished steadily and by DOL 10 was no longer statistically significant (12.7 +/- 4.9 vs 10.0 +/- 4.8). On each NICU day, at all ranges of SNAP scores, there were at least as many infants who would ultimately survive as would die. Consequently, the positive predictive value of any SNAP value for subsequent mortality was <0.5 on all NICU days. Prediction profiles were obtained for 230 ventilated infants reflecting over 11 000 intuitions obtained on 2867 patient days. One hundred fifty-seven (81%) of 192 survivor profiles displayed consistent accurate prediction profiles-at least 90% of their NICU ventilation days were characterized by 100% prediction of survival. Twenty-five (13%) of 192 surviving infants survived somewhat unexpectedly; that is, after at least 1 day characterized by at least 1 estimate of "death." Thirty-three (60%) of the 55 nonsurvivors died before DOL 10. Eighty-two percent of the prediction profiles for these early dying infants were homogeneous, dismal, and accurate. Twenty-two (40%) of the 55 nonsurvivors died after DOL 10. Seventeen (78%) of these 22 late-dying infants were predicted to live by many observers on many hospital days. Sixty-one (30%) of 230 profiled patients had at least 1 NICU day characterized by at least 1 prediction of death; 26/61 (43%) of these patients were incorrectly predicted; that is, they survived. Seventeen infants who were predicted to die during but survived nonetheless were assessed neurologically at 1 year. Fourteen (82%) of these 17 were not neurologically normal-8 were clearly abnormal, 1 suspicious, and 5 had died.

CONCLUSIONS

If absolute certainty about mortality is the only criterion that can justify a decision to withhold or withdraw life-sustaining treatment in the NICU, these data would make such decisions difficult on the first day of life, and increasingly problematic thereafter. However, if we acknowledge that medicine is inevitably an inexact science and that clinical predictions can never be perfect, we can ask the more interesting question of whether good but less-than-perfect predictions of imprecise but ethically relevant clinical outcomes can still be useful. We think that they can-and that they must.

Acute severe asthma: differences in therapies and outcomes among pediatric intensive care units

OBJECTIVE

To determine differences in therapies and outcomes among pediatric intensive care units for patients with acute severe asthma.

DESIGN

Retrospective cohort study.

SETTING

Eleven pediatric intensive care units participating in the Pediatric Intensive Care Evaluations.

PATIENTS

Patients were 1528 children with a primary diagnosis of asthma.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We studied severity of illness, length of stay, and use of invasive interventions. The patients at the centers had similar median physiologic measures of illness and Pediatric Risk of Mortality III scores. The patients received a wide range of invasive interventions depending on institution, including mechanical ventilation (3% to 47%), arterial catheter placement (4% to 46%), central venous catheter (2% to 51%), and determination of a blood gas (24% to 70%). At institutions where mechanical ventilation was used more commonly (>20%, high use), intensive care and hospital stays were longer for asthmatic patients regardless of mechanical ventilation requirement compared with centers with lower use of mechanical ventilation. The status of "high-use center" was an independent predictor for intensive care stay (p = .005) and hospital length of stay (p = .017) as well as duration of mechanical ventilation (p = .014) after adjustment for age, degree of hypercarbia, maximal respiratory rate, use of an arterial catheter, and Pediatric Risk of Mortality III scores among ventilated children.

CONCLUSIONS

We found that use of invasive interventions including mechanical ventilation and vascular monitoring varied greatly by institution. Centers with higher use of mechanical ventilation had longer median intensive care stay and hospital stays. Pediatric asthma management for acute severe asthma may be improved by clear elucidation of the institutional practices where fewer invasive interventions were used to achieve better outcomes.

Serial evaluation of the oncological pediatric risk of mortality (O-PRISM) score following allogeneic bone marrow transplantation in children

The O-PRISM score was introduced for risk assessment in children transferred to intensive care following BMT. The aim of this study is to determine the prognostic value of a serial evaluation of the O-PRISM score. Ninety-three children, 58 allogeneic-related and 35 unrelated BMT, were evaluated. At weekly intervals, the O-PRISM was calculated based on the standard PRISM score and the three additional variables CRP, GVHD and hemorrhage. Overall survival was 0.51 +/- 0.05 (48/93 patients). Seventeen children died of recurrent disease and 28 of BMT-related complications. High O-PRISM scores significantly correlated with adverse outcome. The relative risks of DOC of patients with scores > or =10 compared to patients with lower scores were: day 0: 3.9 (95% confidence-interval: 1.1-13.7, P = 0.02), day 7: 2.0 (0.7-6.2, P = 0.20), day 14: 5.2 (1.9-14.0, P = 0.001), day 21: 5.6 (1.9-16.5, P = 0.001), day 28: 11.5 (3.8-100.9, P < 0.001), day 35: 7.3 (1.9-27.7, P = 0.001). As early as day 0, children with scores > or =10 points showed a higher cumulative incidence of DOC than patients with lower scores (0.69 +/- 0.15 vs 0.27 +/- 0.05, P = 0.02). The O-PRISM score represents a useful clinical parameter for serial risk assessment following BMT. As it indicates fatal events early, it may be helpful for parent information and even more for the early establishment of intensified supportive treatment. The O-PRISM score may therefore be a valuable parameter for the evaluation of different strategies for BMT and supportive treatment.

Bed rationing and allocation in the intensive care unit

That the resources available for intensive care cannot be infinite is self-evident. Parallel increases in medical capability, cost, and community expectations have forced intensivists to confront the reality of resource limitation. Traditional bioethical structures cope poorly with this focus beyond the traditional patient-doctor relationship. Allocation of funds for intensive care may be case-based, historically based, per diem, or capitation-based but is always heavily influenced by political and economic considerations. Attempts have been made to relate costs to severity or intervention scores, but all these techniques are limited by the high fixed costs of intensive care. Methods available to help the physician faced with patient-selection dilemmas include cost-effectiveness and cost-utility analysis. These techniques involve assessment of the quality of life with the help of several well-validated quantitative approaches. Choosing between competing patients for intensive care beds is often more a theoretical issue than a practical one, because alternative arrangements can almost always be made. Physicians have an ethical and social responsibility to further develop the tools to inform community debate on these issues.

Complications in pediatric tracheostomies

OBJECTIVE

To examine complications of pediatric tracheostomy.

STUDY DESIGN

Retrospective.

METHODS

Chart review of children undergoing tracheotomy or laryngeal diversion between 1990 and 1999.

RESULTS

Charts of 142 children were examined. Average age was 2.64 years (standard deviation [SD], 4.73 y) at surgery. Duration of tracheostomy was 2.08 years (SD, 1.72 y) for those decannulated, 3.12 years (SD, 2.5 y) for those still with a stoma, and length of follow-up for the whole group was 4.14 years (SD, 8.69 y). At last follow-up, 56% had a tracheostomy, 29% had none, and 15% had died; one death was tracheostomy-related. Three percent had intraoperative complications, 11% had complications before the first tracheostomy tube change, and 63% had complications after the first tube change. Thirty-four percent had a trial of decannulation; 85% of these were successful. Fifty-four percent of those decannulated had complications. Number of complications was not related to duration of follow-up. In-hospital mortality was congruent to mortality predicted by PRISM (Pediatric Rate of Mortality) scores.

CONCLUSIONS

Forty-three percent had serious complications involving loss of the tracheostomy airway (tube occlusion or accidental decannulation) or requiring a separate surgical procedure. Deaths directly attributable to tracheostomy complications occurred in 0.7%.

Combining physician's subjective and physiology-based objective mortality risk predictions

OBJECTIVE

None of the currently available physiology-based mortality risk prediction models incorporate subjective judgements of healthcare professionals, a source of additional information that could improve predictor performance and make such systems more acceptable to healthcare professionals. This study compared the performance of subjective mortality estimates by physicians and nurses with a physiology-based method, the Pediatric Risk of Mortality (PRISM) III. Then, healthcare provider estimates were combined with PRISM III estimates using Bayesian statistics. The performance of the Bayesian model was then compared with the original two predictions.

DESIGN

Concurrent cohort study.

SETTING

A tertiary pediatric intensive care unit at a university affiliated children's hospital.

PATIENTS

Consecutive admissions to the pediatric intensive care unit.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

For each of the 642 consecutive eligible patients, an exact mortality estimate and the degree of certainty (continuous scale from 1 to 5) associated with the estimate was collected from the attending, fellow, resident, and nurse responsible for the patient's care. Bayesian statistics were used to combine the PRISM III and certainty weighted subjective predictions to create a third Bayesian estimate of mortality. PRISM III discriminated survivors from nonsurvivors very well (area under curve [AUC], 0.924) as did the physicians and nurses (AUCs attendings, 0.953; fellows, 0.870; residents, 0.923; nurses, 0.935). Although the AUCs of the healthcare providers were not significantly different from the AUCs of PRISM III, the Bayesian AUCs were higher than both the healthcare providers' AUCs (p < or = .09 for all) and PRISM III AUCs. Similarly, the calibration statistics for the Bayesian estimates were superior to the calibration statistics for both the healthcare providers and PRISM III models.

CONCLUSIONS

The results of this study demonstrated that healthcare providers' subjective mortality predictions and PRISM III mortality predictions perform equally well. The Bayesian model that combined provider and PRISM III mortality predictions was more accurate than either provider or PRISM III alone and may be more acceptable to physicians. A methodology using subjective outcome predictions could be more relevant to individual patient decision support.

Review of the methodologies and applications of scoring systems in neonatal and pediatric intensive care

Scoring systems and risk prediction rules quantitate the severity of clinical conditions and stratify patients according to a specified outcome. In intensive care medicine, the complexity and number of clinical scoring systems is increasing as their utility in both health services research and clinical medicine broadens. We anticipate that with increasing healthcare costs and competition, the demand for risk adjusted outcomes and institutional benchmarking will increase. As academicians and clinicians, it is vital to be knowledgeable regarding the methodologies and applications of these scoring and risk prediction systems to ensure their quality and appropriate utilization.

Introduction of the oncological pediatric risk of mortality score (O-PRISM) for ICU support following stem cell transplantation in children

Prognostic scoring systems based on physiological parameters have been established in order to predict the outcome of ICU patients. It has been demonstrated that the predictive value of these scores is limited in patients following hematopoietic stem cell transplantation (HSCT). Therefore, we evaluated patients from the Düsseldorf pediatric stem cell transplantation center with regard to predisposing factors and prognostic variables for ICU treatment and outcome. Between January 1989 and December 1998, 180 HSCT have been performed. The clinical, laboratory and HSCT-related parameters such as conditioning treatment, engraftment, GVHD, infections and HSCT toxicity were prospectively recorded and retrospectively analyzed. Established pediatric scoring systems (PRISM, TISS, P-TISS) were applied. Twenty-eight patients required intensive care (16 male, 12 female, median age: 10.9 years (range: 0.4 to 18.9 years), five autologous, 13 allogeneic-related and 10 unrelated transplanted patients). Ventilator-dependent respiratory failure was the most frequent cause of admission to the ICU (n = 23). Fourteen of 28 patients were discharged from ICU, and six of 28 patients achieved a long-term survival (110 to 396 weeks). At admission to the ICU, impaired cardiovascular status, high CRP levels and presence of macroscopic bleeding were each associated with fatal outcome (P < 0.05). The Pediatric Risk of Mortality (PRISM) score was not prognostically significant at the 0.05 level. Long-term survival after discharge from the ICU correlated with HSCT-related parameters such as the type of transplant and severity of GVHD (P = 0.002). By introduction of HSCT related parameters such as severity of GVHD (grade 2: 2 points; grade >2: 4 points), CRP-level (>10 mg/dl: 4 points), and presence of macroscopic bleeding (4 points) into the PRISM score a new oncological PRISM ('O-PRISM') score was established. This score significantly correlated with the risk of mortality in the ICU (P = 0.01). In conclusion, the new O-PRISM score accurately characterizes the clinical situation of children requiring ICU treatment following HSCT. It distinguishes more appropriately between success and failure of ICU treatment following HSCT than the standard prognostic scores. It needs to be evaluated in future prospective studies of critically ill children after HSCT. Bone Marrow Transplantation (2000).