Outcomes in Pediatric Central Line-associated Bloodstream Infections Treated With Antimicrobial Locks: A 14-Year Retrospective Analysis.

BACKGROUND: Central line-associated bloodstream infections (CLABSI) are the most common healthcare-associated infection in children. Antimicrobial lock therapy (ALT) is a recommended component of CLABSI treatment; however, studies characterizing pediatric ALT-based CLABSI therapy are few. METHODS: All CLABSI treated with ALT at a single tertiary-care pediatric hospital for years 2004-2018 were identified by retrospective chart review and demographic, clinical, microbiological and pharmacy data were extracted. Three clinical outcomes were defined: CLABSI clearance, central venous catheter (CVC) removal and infection recurrence. Factors associated with these outcomes were examined and patterns in pathogen occurrence were described. RESULTS: During the study period, 1188 CLABSI treated with ALT were identified. In all, 969 (85%) CLABSI initially cleared. In the CLABSI with initial clearance, CVCs were retained a median of 63 days after infection onset, and 20% recurred. Skin-colonizing microbes ( Staphylococcus aureus , coagulase-negative Staphylococci) were more likely to produce monomicrobial infections, whereas bowel-colonizing microbes ( Escherichia coli , Enterococci) were more associated with polymicrobial infections. In a multivariate model, infection history, resistant pathogens and CVC type were strongly correlated with infection recurrence, whereas bone marrow transplant status, neutrophil count and line type were correlated with CVC removal. CONCLUSIONS: Optimal therapeutic approaches to pediatric CLABSI remain uncertain, including the utility of ALT. Prior reports of ALT use in children have been smaller in size and largely described hematology-oncology or hemodialysis populations. We describe trends in pediatric CLABSI, including efficacy among solid organ transplant recipients and children with short-gut syndrome. These data provide guidance on selecting CLABSI for optimal ALT use.

Association of Bacteremia with Vaccination Status in Children Aged 2 to 36 Months.

OBJECTIVE: To determine the association between bacteremia and vaccination status in children aged 2-36 months presenting to a pediatric emergency department. STUDY DESIGN: Retrospective cohort study of children aged 2-36 months with blood cultures obtained in the pediatric emergency department between January 2013 and December 2017. The exposure of interest was immunization status, defined as number of Haemophilus influenzae type B (Hib) and Streptococcus pneumoniae vaccinations, and the main outcome positive blood culture. Subjects with high-risk medical conditions were excluded. RESULTS: Of 5534 encounters, 4742 met inclusion criteria. The incidence of bacteremia was 1.5%. The incidence of contaminated blood culture was 5.0%. The relative risk of bacteremia was 0.79 (95% CI 0.39-1.59) for unvaccinated and 1.20 (95% CI 0.52-2.75) for undervaccinated children relative to those who had received age-appropriate vaccines. Five children were found to have S pneumoniae bacteremia and 1 child had Hib bacteremia; all of these subjects had at least 3 sets of vaccinations. No vaccine preventable pathogens were isolated from blood cultures of unvaccinated children. We found no S pneumoniae or Hib in children 2-6 months of age who were not fully vaccinated due to age (95% CI 0-0.13%) and the contamination rate in this group was high compared with children 7-36 months (6.6% vs 3.7%). CONCLUSIONS: Bacteremia in young children is an uncommon event. Contaminated blood cultures were more common than pathogens. Bacteremia from S pneumoniae or Hib is uncommon and, in this cohort, was independent of vaccine status.

Predicting outcomes in central venous catheter salvage in pediatric central line-associated bloodstream infection.

OBJECTIVE: Central line-associated bloodstream infections (CLABSIs) are a common, costly, and hazardous healthcare-associated infection in children. In children in whom continued access is critical, salvage of infected central venous catheters (CVCs) with antimicrobial lock therapy is an alternative to removal and replacement of the CVC. However, the success of CVC salvage is uncertain, and when it fails the catheter has to be removed and replaced. We describe a machine learning approach to predict individual outcomes in CVC salvage that can aid the clinician in the decision to attempt salvage. MATERIALS AND METHODS: Over a 14-year period, 969 pediatric CLABSIs were identified in electronic health records. We used 164 potential predictors to derive 4 types of machine learning models to predict 2 failed salvage outcomes, infection recurrence and CVC removal, at 10 time points between 7 days and 1 year from infection onset. RESULTS: The area under the receiver-operating characteristic curve varied from 0.56 to 0.83, and key predictors varied over time. The infection recurrence model performed better than the CVC removal model did. CONCLUSIONS: Machine learning-based outcome prediction can inform clinical decision making for children. We developed and evaluated several models to predict clinically relevant outcomes in the context of CVC salvage in pediatric CLABSI and illustrate the variability of predictors over time.

Use of C-Reactive Protein and Ferritin Biomarkers in Daily Pediatric Practice.

Recent pediatric clinical research has begun to focus on risk stratification tools using multibiomarker models. C-reactive protein (CRP) and ferriti biomarkers are widely available and used to varying degrees in daily practice, but there is no single source examining the evidence behind their use.We set out to summarize the evidence behind the use of CRP and ferritin biomarkers in pediatric practice and to begin development of a consensus for their future use for pediatricians.All the literature involving CRP and ferritin in pediatrics available on PubMed was surveyed. Research applicable to daily pediatric practice was summarized in the body of the article. Pediatric clinicians of various subspecialties contributed to the summary of the use of CRP and ferritin biomarkers in clinical practice in various disease processes. A clinical decision pathway is described, and evidence is summarized.CRP and ferritin biomarkers have diverse uses with various cutoff values in the literature, making their use in daily practice difficult. Elevation of these markers coincides with their significant elevation in uncontrolled inflammation.CRP and ferritin biomarkers are widely used in pediatrics. This review provides a resource summarizing evidence into a single source. There is sufficient evidence to indicate that these biomarkers of inflammation can be useful in guiding clinical decision making in specific clinical scenarios; however, further work is needed to improve their use in clinical practice.

A Decision Tree Using Patient Characteristics to Predict Resistance to Commonly Used Broad-Spectrum Antibiotics in Children With Gram-Negative Bloodstream Infections.

BACKGROUND: As rates of multidrug-resistant gram-negative infections rise, it is critical to recognize children at high risk of bloodstream infections with organisms resistant to commonly used empiric broad-spectrum antibiotics. The objective of the current study was to develop a user-friendly clinical decision aid to predict the risk of resistance to commonly prescribed broad-spectrum empiric antibiotics for children with gram-negative bloodstream infections. METHODS: This was a longitudinal retrospective cohort study of children with gram-negative bacteria cared for at a tertiary care pediatric hospital from June 2009 to June 2015. The primary outcome was a bloodstream infection due to bacteria resistant to broad-spectrum antibiotics (ie, cefepime, piperacillin-tazobactam, meropenem, or imipenem-cilastatin). Recursive partitioning was used to develop the decision tree. RESULTS: Of 689 episodes of gram-negative bloodstream infections included, 31% were resistant to broad-spectrum antibiotics. The decision tree stratified patients into high- or low-risk groups based on prior carbapenem treatment, a previous culture with a broad-spectrum antibiotic resistant gram-negative organism in the preceding 6 months, intestinal transplantation, age ≥3 years, and ≥7 prior episodes of gram-negative bloodstream infections. The sensitivity for classifying high-risk patients was 46%, and the specificity was 91%. CONCLUSION: A decision tree offers a novel approach to individualize patients' risk of gram-negative bloodstream infections resistant to broad-spectrum antibiotics, distinguishing children who may warrant even broader antibiotic therapy (eg, combination therapy, newer ß-lactam agents) from those for whom standard empiric antibiotic therapy is appropriate. The constructed tree needs to be validated more widely before incorporation into clinical practice.

Serious Bacterial Infections in Neonates Presenting Afebrile With History of Fever.

BACKGROUND: Infants ≤28 days of age with fever are frequently hospitalized while undergoing infectious evaluation. We assessed differences in rates of serious bacterial infection (SBI; bacteremia, bacterial meningitis, urinary tract infection) and invasive bacterial infection (IBI; bacteremia, bacterial meningitis) among the following neonates: (1) febrile at presentation (FP), (2) afebrile with history of fever without subsequent fever during hospitalization, and (3) afebrile with history of fever with subsequent fever during hospitalization. METHODS: We performed a single-center retrospective study of neonates evaluated for SBI during emergency department evaluation between January 1, 2006, and December 31, 2017. Patients were categorized into FP, afebrile with no subsequent fever (ANF), and afebrile with subsequent fever (ASF) groups. We compared rates of SBI and IBI between groups using logistic regression and assessed time to fever development using time-to-event analysis. RESULTS: Of 931 neonates, 278 (29.9%) were in the ANF group, 93 (10.0%) were in the ASF group, and 560 (60.2%) were in the FP group. Odds of SBI in neonates ANF were 0.42 (95% confidence interval [CI] 0.23-0.79) compared with infants FP, although differences in IBI were not statistically significant (0.52, 95% CI 0.19-1.51). In infants ASF, median time to fever was 5.6 hours (interquartile range, 3.1-11.4). Infants ASF had higher odds of SBI compared to infants FP (odds ratio 1.93, 95% CI 1.07-3.50). CONCLUSIONS: Neonates with history of fever who remain afebrile during hospitalization may have lower odds for SBI and be candidates for early discharge after an observation period.

Immature neutrophils in young febrile infants.

BACKGROUND: To describe the diagnostic value of the absolute band count (ABC) and ratio of immature to total neutrophils (I:T) for invasive bacterial infections (IBIs; bacterial meningitis and bacteraemia) among young febrile infants. METHODS: We performed a cross-sectional study in a paediatric emergency department of febrile infants ≤60 days over 12 years to evaluate the accuracy of the ABC and I:T for IBI. RESULTS: Of 2930 included patients, 75 (2.6%) had IBIs. The area under the curve (AUC; 95% CI) for ABC was 0.69 (0.62 to 0.76) with sensitivity 0.27 (0.17 to 0.38) and specificity 0.94 (0.93 to 0.95) at cutoff ≥1500 cells/µL. The AUC for I:T was 0.65 (0.59 to 0.72) with sensitivity 0.29 (0.19 to 0.41) and specificity 0.88 (0.87 to 0.89) at cutoff ≥0.2. Only the ABC in infants 29-60 days was minimally accurate. CONCLUSION: The ABC and I:T were generally inaccurate for detecting IBI in febrile infants. Guidelines without these parameters may be better for risk assessment.

Factors associated with serious bacterial infections in infants ≤60 days with hypothermia in the emergency department.

BACKGROUND: We sought to investigate risk factors for serious bacterial infection (SBI: bacterial meningitis, bacteremia, and urinary tract infection [UTI]) among infants ≤60 days of age presenting to the emergency department (ED) with hypothermia (temperature < 36 °C). METHODS: We performed a single center study over a 12-year period including all patients ≤60 days old with hypothermia, excluding patients who did not receive a blood culture and patients who received antibiotics prior to culture acquisition. The primary outcome was SBI. Secondary outcomes were mortality and herpes simplex infection. We performed multivariable logistic regression to identify risk factors for primary outcomes reporting adjusted odds ratios with 95% confidence intervals (aOR, 95% CI). RESULTS: 360 infants were identified. 10/360 (2.8%) had an SBI. All episodes of SBI occurred in infants ≤28 days of age. Two patients had meningitis, two had meningitis with bacteremia, one had isolated bacteremia, and five had UTI. Associated diagnoses included prematurity (46.9%), hyperbilirubinemia (28.3%) and dehydration (14.7%). In multivariable analysis, presentation at 15-28 days (7.60, 1.81-31.86; p = 0.005) compared to 0-14 days, higher absolute neutrophil count (1.25, 1.04-1.50; p = 0.015) and lower platelet count (0.99, 0.99-1.00; p = 0.046) were associated with SBI. Three patients without SBI died during or soon after their hospitalization. One patient had positive testing for herpes simplex. CONCLUSION: In this cohort of hypothermic infants, 2.8% had a SBI. Age of presentation, ANC, and lower platelet count were associated with serious infections. Hypothermic infants presenting to the ED carry significant morbidity and require prospective study to better risk-stratify this population.

Should Children With Constipation Undergo Blood Lead Level Screening?

This study aimed to compare the prevalence of elevated blood lead level in children with constipation to the population prevalence. We reviewed the charts of 441 children who were screened with a blood lead level on presenting to the gastroenterology clinic at UPMC Children's Hospital of Pittsburgh for evaluation of constipation. The prevalence of blood lead level greater than 5 µg/dL was 1.36% (6/441; 95% confidence interval = 1.23% to 1.49%), which is significantly lower than the 4.01% prevalence in the population reported by the Center for Disease Control and Prevention. No patients had a blood lead level greater than 10 µg/dL. Age under 5 years old showed an increased odds of lead level greater than 5 µg/dL (odds ratio = 7.5; 95% confidence interval = 1.2 to 47.3, P < .05). We concluded that children seen in the gastroenterology clinic for constipation are unlikely to have an elevated blood lead level on routine screening.

Insights From the Geographic Spread of the Lyme Disease Epidemic.

Background: Lyme disease is the most common reportable zoonotic infection in the United States. Recent data suggest spread of the Ixodes tick vector and increasing incidence of Lyme disease in several states, including Pennsylvania. We sought to determine the clinical presentation and healthcare use patterns for pediatric Lyme disease in western Pennsylvania. Methods: The electronic medical records of all patients with an International Classification of Disease, Ninth Revision, diagnosis of Lyme disease between 2003 and 2013 at Children's Hospital of Pittsburgh were individually reviewed to identify confirmed cases of Lyme disease. The records of 773 patients meeting these criteria were retrospectively analyzed for patient demographics, disease manifestations, and healthcare use. Results: An Lyme disease increased exponentially in the pediatric population of western Pennsylvania. There was a southwestward migration of Lyme disease cases, with a shift from rural to nonrural zip codes. Healthcare provider involvement evolved from subspecialists to primary care pediatricians and emergency departments (EDs). Patients from nonrural zip codes more commonly presented to the ED, while patients from rural zip codes used primary care pediatricians and EDs equally. Conclusions: The current study details the conversion of western Pennsylvania from a Lyme-naive to a Lyme-epidemic area, highlighting changes in clinical presentation and healthcare use over time. Presenting symptoms and provider type differed between those from rural and nonrural zip codes. By elucidating the temporospatial epidemiology and healthcare use for pediatric Lyme disease, the current study may inform public health measures regionally while serving as an archetype for other areas at-risk for Lyme disease epidemics.

Oral Management for Pediatric Lyme Meningitis.

BACKGROUND: Guidelines for pediatric Lyme meningitis recommend treatment with parenteral therapy [1, 2]. Adult studies suggest that Lyme meningitis can be successfully treated with oral therapy. Our objective was to evaluate the clinical response, side effects and outcome of oral therapy for Lyme meningitis in the pediatric population compared with parenteral therapy in an area endemic for Lyme disease. METHODS: We conducted a case series chart review from January 2012 to May 2017 of pediatrics patient diagnosed and treated for Lyme meningitis. We recorded clinical presentation, laboratory values, antimicrobial therapy and follow up after therapy to compare the efficacy of oral versus parenteral route of therapy. RESULTS: We identified 38 patients diagnosed with Lyme meningitis. Thirty-two patients were discharge with exclusively oral therapy with: doxycycline and amoxicillin. We had only 2 patients developed potential adverse effects from oral doxycycline therapy. All patients treated with oral antibiotics had resolution of symptoms on follow up appointments. CONCLUSIONS: Oral therapy for Lyme meningitis yields no serious adverse events, was well tolerated and showed resolution of symptoms.

Sternoclavicular Arthritis as a Clinical Presentation for Lyme Disease.

Lyme disease is caused by Borrelia burgdorferi and can lead to dermatologic, neurologic, cardiac, and musculoskeletal manifestations. The arthritis of Lyme disease is typically monoarticular, with the knee being most commonly involved. Lyme arthritis of small joints has not previously been well described. We report 3 children who presented with sternoclavicular joint swelling and who were found to have Lyme disease based on enzyme-linked immunosorbent assay and Western blot. This description of sternoclavicular Lyme arthritis highlights the importance of considering Lyme disease in the differential and diagnostic workup of new onset, small joint arthritis in patients presenting from or with travel to Lyme endemic regions.

The effects of natural language processing on cross-institutional portability of influenza case detection for disease surveillance.

OBJECTIVES: This study evaluates the accuracy and portability of a natural language processing (NLP) tool for extracting clinical findings of influenza from clinical notes across two large healthcare systems. Effectiveness is evaluated on how well NLP supports downstream influenza case-detection for disease surveillance. METHODS: We independently developed two NLP parsers, one at Intermountain Healthcare (IH) in Utah and the other at University of Pittsburgh Medical Center (UPMC) using local clinical notes from emergency department (ED) encounters of influenza. We measured NLP parser performance for the presence and absence of 70 clinical findings indicative of influenza. We then developed Bayesian network models from NLP processed reports and tested their ability to discriminate among cases of (1) influenza, (2) non-influenza influenza-like illness (NI-ILI), and (3) 'other' diagnosis. RESULTS: On Intermountain Healthcare reports, recall and precision of the IH NLP parser were 0.71 and 0.75, respectively, and UPMC NLP parser, 0.67 and 0.79. On University of Pittsburgh Medical Center reports, recall and precision of the UPMC NLP parser were 0.73 and 0.80, respectively, and IH NLP parser, 0.53 and 0.80. Bayesian case-detection performance measured by AUROC for influenza versus non-influenza on Intermountain Healthcare cases was 0.93 (using IH NLP parser) and 0.93 (using UPMC NLP parser). Case-detection on University of Pittsburgh Medical Center cases was 0.95 (using UPMC NLP parser) and 0.83 (using IH NLP parser). For influenza versus NI-ILI on Intermountain Healthcare cases performance was 0.70 (using IH NLP parser) and 0.76 (using UPMC NLP parser). On University of Pisstburgh Medical Center cases, 0.76 (using UPMC NLP parser) and 0.65 (using IH NLP parser). CONCLUSION: In all but one instance (influenza versus NI-ILI using IH cases), local parsers were more effective at supporting case-detection although performances of non-local parsers were reasonable.

A study of the transferability of influenza case detection systems between two large healthcare systems.

OBJECTIVES: This study evaluates the accuracy and transferability of Bayesian case detection systems (BCD) that use clinical notes from emergency department (ED) to detect influenza cases. METHODS: A BCD uses natural language processing (NLP) to infer the presence or absence of clinical findings from ED notes, which are fed into a Bayesain network classifier (BN) to infer patients' diagnoses. We developed BCDs at the University of Pittsburgh Medical Center (BCDUPMC) and Intermountain Healthcare in Utah (BCDIH). At each site, we manually built a rule-based NLP and trained a Bayesain network classifier from over 40,000 ED encounters between Jan. 2008 and May. 2010 using feature selection, machine learning, and expert debiasing approach. Transferability of a BCD in this study may be impacted by seven factors: development (source) institution, development parser, application (target) institution, application parser, NLP transfer, BN transfer, and classification task. We employed an ANOVA analysis to study their impacts on BCD performance. RESULTS: Both BCDs discriminated well between influenza and non-influenza on local test cases (AUCs > 0.92). When tested for transferability using the other institution's cases, BCDUPMC discriminations declined minimally (AUC decreased from 0.95 to 0.94, p<0.01), and BCDIH discriminations declined more (from 0.93 to 0.87, p<0.0001). We attributed the BCDIH decline to the lower recall of the IH parser on UPMC notes. The ANOVA analysis showed five significant factors: development parser, application institution, application parser, BN transfer, and classification task. CONCLUSION: We demonstrated high influenza case detection performance in two large healthcare systems in two geographically separated regions, providing evidentiary support for the use of automated case detection from routinely collected electronic clinical notes in national influenza surveillance. The transferability could be improved by training Bayesian network classifier locally and increasing the accuracy of the NLP parser.

Development and validation of a continuously age-adjusted measure of patient condition for hospitalized children using the electronic medical record.

Awareness of a patient's clinical status during hospitalization is a primary responsibility for hospital providers. One tool to assess status is the Rothman Index (RI), a validated measure of patient condition for adults, based on empirically derived relationships between 1-year post-discharge mortality and each of 26 clinical measurements available in the electronic medical record. However, such an approach cannot be used for pediatrics, where the relationships between risk and clinical variables are distinct functions of patient age, and sufficient 1-year mortality data for each age group simply do not exist. We report the development and validation of a new methodology to use adult mortality data to generate continuously age-adjusted acuity scores for pediatrics. Clinical data were extracted from EMRs at three pediatric hospitals covering 105,470 inpatient visits over a 3-year period. The RI input variable set was used as a starting point for the development of the pediatric Rothman Index (pRI). Age-dependence of continuous variables was determined by plotting mean values versus age. For variables determined to be age-dependent, polynomial functions of mean value and mean standard deviation versus age were constructed. Mean values and standard deviations for adult RI excess risk curves were separately estimated. Based on the "find the center of the channel" hypothesis, univariate pediatric risk was then computed by applying a z-score transform to adult mean and standard deviation values based on polynomial pediatric mean and standard deviation functions. Multivariate pediatric risk is estimated as the sum of univariate risk. Other age adjustments for categorical variables were also employed. Age-specific pediatric excess risk functions were compared to age-specific expert-derived functions and to in-hospital mortality. AUC for 24-h mortality and pRI scores prior to unplanned ICU transfers were computed. Age-adjusted risk functions correlated well with similar functions in Bedside PEWS and PAWS. Pediatric nursing data correlated well with risk as measured by mortality odds ratios. AUC for pRI for 24-h mortality was 0.93 (0.92, 0.94), 0.93 (0.93, 0.93) and 0.95 (0.95, 0.95) at the three pediatric hospitals. Unplanned ICU transfers correlated with lower pRI scores. Moreover, pRI scores declined prior to such events. A new methodology to continuously age-adjust patient acuity provides a tool to facilitate timely identification of physiologic deterioration in hospitalized children.

Antibiotic Use in Primary Palatoplasty: A Survey of Practice Patterns, Assessment of Efficacy, and Proposed Guidelines for Use.

BACKGROUND: The literature provides no guidelines for antibiotic use in palatoplasty. The authors sought to ascertain practice patterns; review a large, single-surgeon experience, and propose guidelines for antibiotic use in primary palatoplasty. METHODS: A six-question survey was e-mailed to all surgeons of the American Cleft Palate-Craniofacial Association. A retrospective study was also conducted of the senior author's 10-year primary palatoplasty series, and two groups were studied. Group 1 received no antibiotics. Group 2 received preoperative and/or postoperative antibiotics. RESULTS: Three hundred twelve of 1115 surgeons (28 percent) responded to the survey. Eighty-five percent administered prophylactic antibiotics, including 26 percent who used a single preoperative dose. A further 23 percent gave 24 hours of postoperative therapy; 12 percent used 25 to 72 hours, 16 percent used 4 to 5 days, and 12 percent used 6 to 10 days. Five percent of surgeons administered penicillin, 64 percent administered a first-generation cephalosporin, 13 percent administered ampicillin/sulbactam, and 8 percent gave clindamycin. The authors reviewed 311 patients; 173 receive antibiotics and 138 did not. Delayed healing and fistula rates did not differ between groups: 16.8 percent versus 15.2 percent (p = 0.71) and 2.9 percent versus 1.4 percent (p = 0.47), respectively. A single patient treated without antibiotics developed a postoperative bacteremia. This case did not meet the Centers for Disease Control definition of a surgical site infection, but the patient developed a palatal fistula. CONCLUSIONS: Antibiotic use in primary palatoplasty varies widely. The authors' data support a clinician's choice to forego antibiotic use; however, given the significance of palatal fistulae and the single case of postoperative streptococcal bacteremia, the study group recommends a single preoperative dose of ampicillin/sulbactam. Current evidence cannot justify the use of protracted antibiotic regimens. CLINICAL QUESTION/LEVEL OF EVIDENCE: Therapeutic, III.

Evaluation of Electronic Medical Record Vital Sign Data Versus a Commercially Available Acuity Score in Predicting Need for Critical Intervention at a Tertiary Children's Hospital.

OBJECTIVES: Evaluate the ability of vital sign data versus a commercially available acuity score adapted for children (pediatric Rothman Index) to predict need for critical intervention in hospitalized pediatric patients to form the foundation for an automated early warning system. DESIGN: Retrospective review of electronic medical record data. SETTING: Academic children's hospital. PATIENTS: A total of 220 hospitalized children 6.7 ± 6.7 years old experiencing a cardiopulmonary arrest (condition A) and/or requiring urgent intervention with transfer (condition C) to the ICU between January 2006 and July 2011. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Physiologic data 24 hours preceding the event were extracted from the electronic medical record. Vital sign predictors were constructed using combinations of age-adjusted abnormalities in heart rate, systolic and diastolic blood pressures, respiratory rate, and peripheral oxygen saturation to predict impending deterioration. Sensitivity and specificity were determined for vital sign-based predictors by using 1:1 age-matched and sex-matched non-ICU control patients. Sensitivity and specificity for a model consisting of any two vital sign measurements simultaneously outside of age-adjusted normal ranges for condition A, condition C, and condition A or C were 64% and 54%, 57% and 53%, and 59% and 54%, respectively. The pediatric Rothman Index (added to the electronic medical record in April 2009) was evaluated in a subset of these patients (n = 131) and 16,138 hospitalized unmatched non-ICU control patients for the ability to predict condition A or C, and receiver operating characteristic curves were generated. Sensitivity and specificity for a pediatric Rothman Index cutoff of 40 for condition A, condition C, and condition A or C were 56% and 99%, 13% and 99%, and 28% and 99%, respectively. CONCLUSIONS: A model consisting of simultaneous vital sign abnormalities and the pediatric Rothman Index predict condition A or C in the 24-hour period prior to the event. Vital sign only prediction models have higher sensitivity than the pediatric Rothman Index but are associated with a high false-positive rate. The high specificity of the pediatric Rothman Index merits prospective evaluation as an electronic adjunct to human-triggered early warning systems.