Timothy S. Yeh

Outcome of pediatric patients with multiple organ system failure.

The association of multiple organ system failure (MOSF) with mortality was investigated in 831 consecutive admissions to a pediatric ICU. The incidence of MOSF (at least two organ system failures, OSF) was 27%. Of the 62 nonsurvivors, 60 (97%) had MOSF. The mortality for patients with MOSF was 54%, compared to a mortality of 0.3% for patients without MOSF. Mortality increased directly with increasing number of OSF (p <.0001). The mortality was 1% for one OSF, 11% for two OSF, 50% for three OSF, and 75% for four OSF. Comparison of these results with data from adult ICU patients indicates that the mortality and clinical course of MOSF in children is distinct from adults. MOSF is significantly associated with mortality in pediatric patients; however, it is not sufficiently discriminating to determine continuation or withdrawal of ICU support.

Variation in therapy and outcome for pediatric head trauma patients.

ObjectiveThis study was undertaken to examine variation in therapies and outcome for pediatric head trauma patients by patient characteristics and by pediatric intensive care unit. Specifically, the study was designed to examine severity of illness on admission to the pediatric intensive care unit, the therapies used during the pediatric intensive care unit stay, and patient outcomes. Data Sources and Setting Consecutive admissions from three pediatric intensive care units were recorded prospectively (n = 5,749). For this study, all patients with an admitting diagnosis of head trauma were included (n = 477). Data collection occurred during an 18-month period beginning in June 1996. All of the pediatric intensive care units were located in children’s hospitals, had residency and fellowship training programs, and were headed by a pediatric intensivist. MethodsAdmission severity was measured as the worst recorded physiological derangement during the period ≤6 hrs before pediatric intensive care unit admission. Therapies and resource use were based on the Therapeutic Intervention Scoring System with adaptations for pediatrics. The use of intracranial pressure monitoring was recorded on admission to the unit (within 1 hr) and at any time during the pediatric intensive care unit stay. Outcomes were measured at the time of pediatric intensive care unit discharge by the Pediatric Overall Performance Category scale. Risk factors for mortality were examined by using bivariate analyses with significant predictors as candidate variables in a logistic regression to predict expected mortality. Intracranial pressure monitoring and other therapies were added to the mortality prediction model to test for protective effects. Finally, race and insurance status were added to the model to test for differences in the quality of care. ResultsThe overall mortality rate for the entire sample was 7.8%. Mortality rates for children ≤1 yr old were significantly higher than for children >1 yr old (16.1% vs. 6.1%;p = .002). Comparisons by insurance status indicated that observed mortality rates were highest for self-paying patients. However, patient characteristics were not associated with use of therapies or standardized mortality rates after adjustment for patient severity. There was significant variation in the use of paralytic agents, seizure medications, induced hypothermia, and intracranial pressure monitoring on admission across the three pediatric intensive care units. In multivariate models, only the use of seizure medications was associated significantly with reduced mortality risk (odds ratio = 0.17; 95% confidence interval = 0.04–0.70;p = .014). ConclusionsTherapies and outcomes vary across pediatric intensive care units that care for children with head injuries. Increased use of seizure medications may be warranted based on data from this observational study. Large randomized controlled trials of seizure prophylaxis in children with head injury have not been conducted and are needed to confirm the findings presented here.

Validation of a Physiologic Stability Index for Use in Critically Ill Infants and Children

Summary: We developed a physiology-based scoring system, the Physiologic Stability Index (PSI) to assess severity of acute illness in the total population of pediatric Intensive Care Unit (ICU) patients. Thirty-four variables from seven physiologic systems were chosen, and the degree of abnormality of each variable was assigned a score reflecting the clinical importance of the derangements. Validity was demonstrated by comparing PSI to hospital mortality and to two other methods that reflect severity of illness, the Clinical Classification System (CCS) and the Therapeutic Intervention Scoring System (TISS). Four hundred and twenty three consecutive admissions to a multidisciplinary ICU were followed daily. Patients classified into higher CCS classes had significantly higher PSI scores (P < 0.001), and there was a highly significant correlation (P < 0.001) between PSI and TISS scores. The linear-logistic regression of observed mortality versus PSI was highly significant (P < 0.0001) and provided an excellent fit. Highly significant differences between survivors and nonsurvivors were observed for PSI scores (P < 0.001), as well as for composite slopes of the regression of PSI scores versus days of care (P < 0.001). These data demonstrate validity of the PSI scoring system.

Assessment of pediatric intensive care--application of the Therapeutic Intervention Scoring System.

There are few reports analyzing the results of intensive care for children. We evaluated quantitatively the amount of care required in our multidisciplinary pediatric ICU using the Therapeutic Intervention Scoring System (TISS) and assessed qualitatively the severity of illness using the Clinical Classification System (CCS). Over a 6-month period, there were 323 patients (99 CCS Class II, 83 Class III, 141 Class IV) whose overall mortality at 1-month follow-up was 10% (Class II, 0%; Class III 2%; Class IV, 23%). A strong association was obtained between CCS and TISS admission scores (Class II-TISS, 11 +/- 0.6; Class III-TISS, 20 +/- 0.8; Class IV-TISS, 38 +/- 1.0). Class IV patients had a highly significant difference between survivors (S) and nonsurvivors (NS) for admission TISS (S = 36, NS = 47, p less than 0.001) and highest TISS (S = 38, NS = 54, p less than 0.001), as well as slopes of the regression of TISS points versus days of care (S = -4.2 vs. NS = +2.3). The mortality of our Class IV patients was lower than a comparable adult population with similar TISS scores; however, the TISS regression slopes for Class IV patients were similar. We conclude that CCS and TISS are both useful for describing the pediatric intensive care patient population. TISS is particularly helpful in assessing the amount of care received as well as providing a means of evaluating severity of illness.

Evaluation of pediatric intensive care.

A total of 294 Clinical Classification System (CCS) Classes III and IV patients in a pediatric ICU (PICU) were evaluated in terms of severity of illness and quantity of care. The group was comprised of patients from 3 services: medicine, cardiovascular surgery, and other surgery. Severity of illness was measured by the Physiologic Stability Index (PSI) and quantity of care was measured by the Therapeutic Intervention Scoring System (TISS). Comparisons were made between survivors and nonsurvivors and among the 3 services. Nonsurvivors had significantly higher (p < .01) PSI and TISS scores than survivors. Medical patients had the highest PSI scores while cardiovascular surgery patients had the highest TISS scores. Analysis of 7-day regression slopes for all survivor groups and medicine and other surgery nonsurvivor groups demonstrated slopes consistent with the expected clinical course. Cardiovascular surgery nonsurvivor slopes were unique and demonstrated increasing stability with stable amounts of care. The PSI/TISS ratio was used to relate levels of physiologic instability to the amount of therapy. Medical patients had the highest ratios and cardiovascular surgery patients had the lowest ratios. Comparisons of survivors and nonsurvivors for the PSI/TISS ratios and regression slopes demonstrated differences that were not evident through comparison of PSI and TISS scores alone.

Factors explaining variability among caregivers in the intent to restrict life-support interventions in a pediatric intensive care unit.

OBJECTIVE To explore patient-related factors which influence the decisions of pediatric intensive care unit (ICU) caregivers to restrict life-support interventions. DESIGN Cross-sectional survey. SETTING A university-affiliated pediatric ICU. SUBJECTS All physicians and nurses caring for oncology or cardiology ICU patients. INTERVENTIONS Seven patient characteristics were systematically presented in 16 theoretical case scenarios. MEASUREMENTS AND MAIN RESULTS Individual linear regression models were constructed for each participant by calculating the importance caregivers placed on seven patient characteristics when deciding about starting intravenous vasopressors, performing chest compressions, and withdrawing life support. We compared the numerical and descriptive (very low, low, moderate, high) probability of survival. We surveyed 86 caregivers and 56 (65%) responded. The most important factors influencing decisions were family preferences (76% of decisions), followed by probability of survival (50%), and functional status (47%). There was marked variability among respondents in 38 (79%)/48 of the questions; 20% to 50% of caregivers chose opposing directions of patient management when they were asked to indicate the likelihood that they would perform a specific life-support intervention. The same term was never used by all respondents to describe the probability of survival for a scenario. CONCLUSION Critically ill children and their families could face markedly different attitudes about the restriction of life-support interventions, depending on which nurses and physicians are involved in their care.

205 MONITORING PATIENTS IN A PICU: COST CONTAINMENT POTENTIAL

ICUs are expensive and account for∼1% of the GNP. Many patients (pts) without life threatening dysfunction and not utilizing unique PICU therapies who receive only monitoring could be cared for in other hospital areas at a significant cost savings. Methods: All admissions to a 16 bed PICU for 13.5 months were evaluated. Daily assessemnts of types of care were accomplished with the Therapeutic Intervention Scoring System (TISS). Each of the TISS components were categorized as PICU therapies (e.g. mech ventilation), mentoring: personal intensive (e.g. hourly vital signs), monitoring: technology intensive (e.g. arterial catheter) and routine care. Daily assessments of severity of illness utilized the Physiologic Stability Index (PSI). Patients were divided into low (mortality risk<l%) and high risk groups by an average PSI cutoff of 9.Results: 822 admissions utilizing 3969 days of care were evaluated. 226 pts (27.5%) utilizing 297 days of care (7.5%) did not receive unique PICU therapies (monitoring pts). 94.2% (213/266) were at low risk. A single diagnosis or clinical service did not predominate. Most care modalities used by the monitoring pts were personnel intensive (e.g. hourly VS − 82.7%, accurate I/O - 56.6%, mult stat studies - 45.1%). 93.8% of monitoring pts stayed ≤ 2 days.Conclusions: 1) Significant numbers of PICU pts never recieve unique benefits from their PICU stay. 2) Most monitoring pts are at low risk. 3) Protocols for improving PICU efficiency should be directed at screening admissions rather than reducing stay and should concentrate on care needs, not diagnostic entities.

203 SEVERITY OF ILLNESS IN CHILDREN: MULTI NSTITUTIONAL VALIDATION OF THE PHYSIOLOGIC STABILITY INDEX (PSI)

Direct assessment of severity of illness would benefit many aspects of pediatric intensive care (e.q. cost containment, quality of care, therapeutic efficacies, institutional comparisons). The PSI, a direct assessment of severity of illness has been prospectively validated in 1 PICU. The PSI assesses risk of mortality by sampling 34 variables from 7 physiologic systems. However, assumptions in the PSIs development, especially the possibility of under or over measurements, require a multi-institutional validation. Participating centers collected daily PSI scores on≥200 consecutive PICU patients or for≥6 months. The observed mortality for ordered admission PSI scores was compared to the predicted mortality. The model to predict mortality was developed by logistic regression analysis predicting mortality from the organ system PSI scores and age as predictor variables (822 patients).Results: 4 centers have completed data collection. The average mortality rates ranged from 3.0% to 12.5% (p<.01). However, when adjusted for severity of illness by admission PSI scores, the observed mortality rates for the admission PSI intervals were statistically reliably predicted by the model (X2 (3) for each center = 0.84, 2.54, 4.15, 6.36).Conclusions: 1) The PSI is a statistically reliable predictor of mortality in all tested centers. 2) A single system is applicable to all tested centers.

OUTCOME OF CHILDREN WITH MULTIPLE SYSTEM ORGAN FAILURE

Multiple system organ failure (OSF) in infants and children has not been studied. We hypothesized that (1) severity of illness (mortality) is associated with increasing number of OSFs and (2) the mortality associated with multiple OSF in pediatric patients would differ significantly from adult results. Methods: Physiologic data (pertaining to the C-V, respiratory, neurologic, hematologic, and renal systems) in 831 consecutive admissions to a pediatric ICU were analyzed daily. Criteria for OSF were rigidly defined (e.g. renal failure = BUN > 100, creatinine > 2, dialysis). Results: 467 (56%) of patients had 1 or more OSFs. Mortality increased directly with increasing number of OSFs (p < .0001) as follows: 1 OSF (n = 241) = 1% mortality; 2 OSFs (n = 142) = 11% mortality; 3 OSFs (n = 72) = 50% mortality; 4 OSFs (n = 12) = 75% mortality. There were no significant differences in mortality among specific OSFs or combination of OSFs. For the first 10 days of OSF, mortality was not associated with duration of OSF. The independece of mortality and duration of OSF is significantly different from published adult results (p <.005). Discussion: Mortality is significantly associated with increasing numbers of OSFs but not the duration of OSF or specific OSF combinations. The mortality for multiple OSF in pediatric patients is significantly less than in adults. Results of studies in adults ICU patients do not necessarily apply to pediatric patients.