Prediction of patient disposition: comparison of computer and human approaches and a proposed synthesis.

OBJECTIVE: To compare the accuracy of computer versus physician predictions of hospitalization and to explore the potential synergies of hybrid physician-computer models. MATERIALS AND METHODS: A single-center prospective observational study in a tertiary pediatric hospital in Boston, Massachusetts, United States. Nine emergency department (ED) attending physicians participated in the study. Physicians predicted the likelihood of admission for patients in the ED whose hospitalization disposition had not yet been decided. In parallel, a random-forest computer model was developed to predict hospitalizations from the ED, based on data available within the first hour of the ED encounter. The model was tested on the same cohort of patients evaluated by the participating physicians. RESULTS: 198 pediatric patients were considered for inclusion. Six patients were excluded due to incomplete or erroneous physician forms. Of the 192 included patients, 54 (28%) were admitted and 138 (72%) were discharged. The positive predictive value for the prediction of admission was 66% for the clinicians, 73% for the computer model, and 86% for a hybrid model combining the two. To predict admission, physicians relied more heavily on the clinical appearance of the patient, while the computer model relied more heavily on technical data-driven features, such as the rate of prior admissions or distance traveled to hospital. DISCUSSION: Computer-generated predictions of patient disposition were more accurate than clinician-generated predictions. A hybrid prediction model improved accuracy over both individual predictions, highlighting the complementary and synergistic effects of both approaches. CONCLUSION: The integration of computer and clinician predictions can yield improved predictive performance.

Effect of a Sepsis Screening Algorithm on Care of Children with False-Positive Sepsis Alerts.

OBJECTIVES: To determine if implementation of an automated sepsis screening algorithm with low positive predictive value led to inappropriate resource utilization in emergency department (ED) patients as evidenced by an increased proportion of children with false-positive sepsis screens receiving intravenous (IV) antibiotics. STUDY DESIGN: Retrospective cohort study comparing children <18 years of age presenting to an ED who triggered a false-positive sepsis alert during 2 different 5-month time periods: a silent alert period when alerts were generated but not visible to clinicians and an active alert period when alerts were visible. Primary outcome was the proportion of patients who received IV antibiotics. Secondary outcomes included proportion receiving IV fluid boluses, proportion admitted to the hospital, and ED length of stay (LOS). RESULTS: Of 1457 patients, 1277 triggered a false-positive sepsis alert in the silent and active alert periods, respectively. In multivariable models, there were no changes in the proportion administered IV antibiotics (27.0% vs 27.6%, aOR 1.1 [0.9,1.3]) or IV fluid boluses (29.7% vs 29.1%, aOR 1.0 [0.8,1.2]). Differences in ED LOS and proportion admitted to the hospital were not significant when controlling for similar changes seen across all ED encounters. CONCLUSIONS: An automated sepsis screening algorithm did not lead to changes in the proportion receiving IV antibiotics or IV fluid boluses, department LOS, or the proportion admitted to the hospital for patients with false-positive sepsis alerts.

The effect of an electronic health record-based tool on abnormal pediatric blood pressure recognition.

BACKGROUND: Recognition of high blood pressure (BP) in children is poor, partly due to the need to compute age-sex-height referenced percentiles. This study examined the change in abnormal BP recognition before versus after the introduction of an electronic health record (EHR) app designed to calculate BP percentiles with a training lecture. METHODS AND RESULTS: Clinical data were extracted on all ambulatory, non-urgent encounters for children 3-18 years old seen in primary care, endocrinology, cardiology, or nephrology clinics at an urban, academic hospital in the year before and the year after app introduction. Outpatients with at least 1 BP above the age-gender-height referenced 90th percentile were included. Abnormal BP recognition was defined as a BP related ICD-9 code, referral to nephrology or cardiology, an echocardiogram or renal ultrasound to evaluate BP concern, or a follow-up primary care visit for BP monitoring. Multivariable adjusted logistic regression compared odds of recognition before and after app introduction. Of 78 768 clinical encounters, 3521 had abnormal BP in the pre- and 3358 in the post-app period. App use occurred in 13% of elevated BP visits. Overall, abnormal BP was recognized in 4.9% pre-app period visits and 7.1% of visits post-app (P < .0001). Recognition was significantly higher when the app was actually used (adjusted OR 3.17 95% CI 2.29-4.41, P < .001). Without app use recognition was not different. CONCLUSIONS: BP app advent modestly increased abnormal BP recognition in the entire cohort, but actual app use was associated with significantly higher recognition. Predictors of abnormal BP recognition deserve further scrutiny.

Reforming procedural skills training for pediatric residents: a randomized, interventional trial.

BACKGROUND: Pediatric housestaff are required to learn basic procedural skills and demonstrate competence during training. To our knowledge, an evidenced-based procedural skills curriculum does not exist. OBJECTIVE: To create, implement, and evaluate a modular procedural skills curriculum for pediatric residents. METHODS: A randomized, controlled trial was performed. Thirty-eight interns in the Boston Combined Residency Program who began their training in 2005 were enrolled and randomly assigned. Modules were created to teach residents bag-mask ventilation, venipuncture, peripheral intravenous catheter (PIV) insertion, and lumbar puncture skills. The curriculum was administered to participants in the intervention group during intern orientation. Interns in the control group learned procedural skills by usual methods. Subjects were evaluated by using a structured objective assessment on simulators immediately after the intervention and 7 months later. Success in performing live-patient procedures was self-reported by subjects. The primary outcome was successful performance of the procedure on the initial assessment. Secondary outcomes included checklist and knowledge examination scores, live-patient success, and qualitative assessment of the curriculum. RESULTS: Participants in the intervention group performed PIV placement more successfully than controls (79% vs 35%) and scored significantly higher on the checklist for PIV placement (81% vs 61%) and lumbar puncture (77% vs 68%) at the initial assessment. There were no differences between groups at month 7, and both groups demonstrated declining skills. There were no statistically significant differences in success on live-patient procedures. Those in the intervention group scored significantly higher on knowledge examinations. CONCLUSIONS: Participants in the intervention group were more successful performing certain simulated procedures than controls when tested immediately after receiving the curriculum but demonstrated declining skills thereafter. Future efforts must emphasize retraining, and residents must have sufficient opportunities to practice skills learned in a formal curriculum.

Utility of lumbar puncture for first simple febrile seizure among children 6 to 18 months of age.

OBJECTIVES: American Academy of Pediatrics consensus statement recommendations are to consider strongly for infants 6 to 12 months of age with a first simple febrile seizure and to consider for children 12 to 18 months of age with a first simple febrile seizure lumbar puncture for cerebrospinal fluid analysis. Our aims were to determine compliance with these recommendations and to assess the rate of bacterial meningitis detected among these children. METHODS: A retrospective cohort review was performed for patients 6 to 18 months of age who were evaluated for first simple febrile seizure in a pediatric emergency department between October 1995 and October 2006. RESULTS: First simple febrile seizure accounted for 1% of all emergency department visits for children of this age, with 704 cases among 71 234 eligible visits during the study period. Twenty-seven percent (n = 188) of first simple febrile seizure visits were for infants 6 to 12 months of age, and 73% (n = 516) were for infants 12 to 18 months of age. Lumbar puncture was performed for 38% of the children (n = 271). Samples were available for 70% of children 6 to 12 months of age (131 of 188 children) and 25% of children 12 to 18 months of age (129 of 516 children). Rates of lumbar puncture decreased significantly over time in both age groups. The cerebrospinal fluid white blood cell count was elevated in 10 cases (3.8%). No pathogen was identified in cerebrospinal fluid cultures. Ten cultures (3.8%) yielded a contaminant. No patient was diagnosed as having bacterial meningitis. CONCLUSIONS: The risk of bacterial meningitis presenting as first simple febrile seizure at ages 6 to 18 months is very low. Current American Academy of Pediatrics recommendations should be reconsidered.

Focal motor seizure in a wrestler with a sport-related concussion.

UNLABELLED: Although sport-related concussion is a common injury, it is infrequently associated with seizure. While concussive convulsions, consisting of brief, generalized myoclonic activity while an athlete is unconscious have been described, the authors are aware of no published cases of concussion complicated by focal motor seizures. The authors describe the case of a 16-year-old male wrestler who sustained a sport-related concussion complicated by a focal motor seizure. The acute assessment and management of his injury, as well as follow-up until resolution of his symptoms, is presented. A brief review of the association between convulsive activity and mild traumatic brain injury follows. KEYWORDS: concussion; mild traumatic brain injury; seizure.

The use of routine laboratory studies as screening tools in pediatric abdominal trauma.

BACKGROUND: Standard laboratory panels were shown to play an important role in the evaluation of pediatric blunt abdominal trauma before the routine use of computed tomography (CT) scan. Recently, only a few relatively limited studies have evaluated the use of these "trauma panels." OBJECTIVE: To evaluate the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of routine "trauma panels" for evaluating intra-abdominal injury in pediatric blunt trauma patients. METHOD: We undertook a retrospective medical record review of all children with potential major blunt abdominal trauma who entered the Children's Hospital (Boston, MA) trauma registry from July 1996 to August 1999. Routine laboratory tests during those years included sodium, glucose, white blood cell count, hematocrit, platelets, prothrombin time, activated partial thromboplastin time, aspartate aminotransferase (AST), alanine aminotransferase, amylase, lipase, and urinalysis. Individual findings were considered abnormal if they fell out of the laboratory's respective reference range. We determined sensitivity, specificity, PPV, NPV, and the 95% confidence interval for each test, using abdominal pathology identified by CT scan as the gold standard. RESULTS: Three hundred eighty-two patients were included. Of that, 68% were men. Median age was 115 months (intraquartile range, 60-159 months). In total, 241 of the patients (63%) had an abdominal CT scan performed, 83 of which (33%) had abnormal findings. Abnormal values for glucose, AST, urinalysis, and white blood cell count were the most frequently observed abnormalities (67%, 47%, 43%, and 43%, respectively). Among the 83 patients with abdominal pathology, glucose and AST had the highest sensitivity (75% and 63%, respectively). Lipase had the highest PPV at 75%, and AST had the highest negative predictive at 71%. No routine laboratory test had excellent sensitivity, specificity, PPV, and NPV. CONCLUSIONS: Routine "trauma panels" should not be obtained as a screening tool in children with blunt trauma being evaluated for intra-abdominal injury.