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.

BACKGROUND: Infectious Diseases Society of America guidelines recommend that key antimicrobial stewardship program (ASP) personnel include an infectious disease (ID) physician leader and dedicated ID-trained clinical pharmacist. Limited resources prompted development of an alternative model by using ID physicians and service-based clinical pharmacists at a pediatric hospital. The aim of this study was to analyze the effectiveness and impact of this alternative ASP model. METHODS: The collaborative ASP model incorporated key strategies of education, antimicrobial restriction, day 3 audits, and practice guidelines. High-use and/or high-cost antimicrobial agents were chosen with audits targeting vancomycin, caspofungin, and meropenem. The electronic medical record was used to identify patients requiring day 3 audits and to communicate ASP recommendations. Segmented regression analyses were used to analyze quarterly antimicrobial agent prescription data for the institution and selected services over time. RESULTS: Initiation of ASP and day 3 auditing was associated with blunting of a preexisting increasing trend for caspofungin drug starts and use and a significant downward trend for vancomycin drug starts (relative change –12%) and use (–25%), with the largest reduction in critical care areas. Although meropenem use was already low due to preexisting requirements for preauthorization, a decline in drug use (–31%, P = .021) and a nonsignificant decline in drug starts (–21%, P = .067) were noted. A 3-month review of acceptance of ASP recommendations found rates of 90%, 93%, and 100% for vancomycin, caspofungin, and meropenem, respectively. CONCLUSIONS: This nontraditional ASP model significantly reduced targeted drug usage demonstrating acceptance of integration of service-based clinical pharmacists and ID consultants.

OBJECTIVEEvidence indicates that users incur significant physical and cognitive costs in the use of order sets, a core feature of computerized provider order entry systems. This paper develops data-driven approaches for automating the construction of order sets that match closely with user preferences and workflow while minimizing physical and cognitive workload.MATERIALS AND METHODSWe developed and tested optimization-based models embedded with clustering techniques using physical and cognitive click cost criteria. By judiciously learning from users' actual actions, our methods identify items for constituting order sets that are relevant according to historical ordering data and grouped on the basis of order similarity and ordering time. We evaluated performance of the methods using 47,099 orders from the year 2011 for asthma, appendectomy and pneumonia management in a pediatric inpatient setting.RESULTSIn comparison with existing order sets, those developed using the new approach significantly reduce the physical and cognitive workload associated with usage by 14-52%. This approach is also capable of accommodating variations in clinical conditions that affect order set usage and development.DISCUSSIONThere is a critical need to investigate the cognitive complexity imposed on users by complex clinical information systems, and to design their features according to 'human factors' best practices. Optimizing order set generation using cognitive cost criteria introduces a new approach that can potentially improve ordering efficiency, reduce unintended variations in order placement, and enhance patient safety.CONCLUSIONSWe demonstrate that data-driven methods offer a promising approach for designing order sets that are generalizable, data-driven, condition-based, and up to date with current best practices.

This study examines a new approach of using the Design Structure Matrix (DSM) modeling technique to improve the design of Electronic Medical Record (EMR) user interfaces. The usability of an EMR medication dosage calculator used for placing orders in an academic hospital setting was investigated. The proposed method captures and analyzes the interactions between user interface elements of the EMR system and groups elements based on information exchange, spatial adjacency, and similarity to improve screen density and time-on-task. Medication dose adjustment task time was recorded for the existing and new designs using a cognitive simulation model that predicts user performance. We estimate that the design improvement could reduce time-on-task by saving an average of 21 hours of hospital physicians' time over the course of a month. The study suggests that the application of DSM can improve the usability of an EMR user interface.

Order sets as part of computerized provider order entry (CPOE) have the potential to improve care delivery by making it faster and easier for physicians to enter orders and by guiding care according to known best practices. Currently, order sets are not utilized to their full extent due to factors such as user inexperience, lack of updated content with evolving best practices, and inability to modify an order set to include relevant items. This exploratory study uses order data from Asthma and Appendectomy patients at a large pediatric healthcare institution to examine the optimization of current ordering patterns using direct and cognitive click-through costs as evaluation criteria. We examine four models where modifications to current ordering practices are analyzed: improving order set usage through removal of inexperienced-user effect, changed default setting based on scientific evidence, and newly designed order sets through K-means clustering. While improving current ordering practice was found to reduce cost across all diagnoses and severity levels, the most significant decrease in cost was realized when clustering individual items into new order sets, pointing to a promising new approach for order set optimization.

Higher cognitive workload due to poor usability is a significant, unanticipated consequence of healthcare information technology (IT), resulting in new types of medical errors. An important example of this can be observed in the use of order sets, which allow safe and efficient provider order entry guided by known best practices. This paper aims to improve IT-enabled order entry by re-designing order sets using data-driven approaches to develop new order sets that match current usage and workflow, while incurring minimum cognitive workload. Applying optimization models embedded with clustering techniques, our methods identify items for constituting order sets that are relevant based on historical ordering data wherein items for a single patient are often placed together or in close temporal proximity during hospital stay. Results indicate that the new approaches dominate current solutions, significantly reducing cognitive workload, and improving order set content. Data driven methods thus offer a promising approach for designing order sets that are generalizable, evidence-based and up-to-date with current best practices.

Computerized physician order entry (CPOE) systems can create unintended consequences. These include medication errors and adverse drug events. We look at a less understood error; patient misidentification. First, two email surveys were used to establish potential risk factors for this error. Next, an automated detection trigger was designed and validated with inpatient medication orders at a large pediatric hospital. The incidence was 0.064% per medication ordered. Finally, a case-control study identified the following as significant risk factors on multivariate analysis: patient age, last name spelling, bed proximity, medical service, time/date of order, and ordering intensity. These results can be used to improve patient safety by increasing awareness of high risk situations and guiding future research.

Integrating clinical data with administrative data across disparate electronic medical record systems will help improve the internal and external validity of comparative effectiveness research. The Pediatric Health Information System (PHIS) currently collects administrative information from 43 pediatric hospital members of the Child Health Corporation of America (CHCA). Members of the Pediatric Research in Inpatient Settings (PRIS) network have partnered with CHCA and the University of Utah Biomedical Informatics Core to create an enhanced version of PHIS that includes clinical data. A specialized version of a data federation architecture from the University of Utah ("FURTHeR") is being developed to integrate the clinical data from the member hospitals into a common repository ("PHIS+") that is joined with the existing administrative data. We report here on our process for the first phase of federating lab data, and present initial results.

Although endemic measles transmission has been interrupted in the United States, importations of this highly infectious virus continue. On March 28, 2009, a physician notified the Pennsylvania Department of Health (PADOH) of a measles case involving an unvaccinated child. Within 5 days, four additional cases were reported to PADOH and the Allegheny County Health Department. All five infected persons had been in the same hospital emergency department (ED) on March 10; one of them was a physician who worked in the ED. To find the source patient, PADOH reviewed electronic records of patients evaluated in the ED on March 10 for fever and rash. This identified a child who arrived recently from India, was treated for viral exanthema, and discharged. On April 3, PADOH obtained serum from this child and confirmed a diagnosis of measles. After an extensive regional search and investigation of the six patients' 4,000 contacts, no additional cases were identified. The hospital reviewed employee health records to identify any exposed personnel who did not have serologic evidence of measles immunity. Among 168 potentially exposed employees, 72 (43%) had no documented measles immunity, thus requiring serologic testing and subsequent vaccination if they lacked serologic evidence of immunity. This outbreak highlights the potential for measles transmission in health-care settings. To decrease transmission, clinicians should know the signs and symptoms of measles, request travel histories of patients suspected of any infectious disease, and isolate potentially infectious patients. Hospital employees should have documented immunity to measles, and employees without evidence of measles immunity should be offered vaccination in accordance with Advisory Committee on Immunization Practices (ACIP) and Hospital Infection Control Practices Advisory Committee (HICPAC) recommendations.

Order sets as part of the Computerized Provider Order Entry (CPOE) system can improve care delivery through allowing faster and easier physician order entry guided by known best practices. This study examines current utilization patterns of order sets and "a la carte" orders in a pediatric environment with a preliminary investigation of methods to automate the creation and modification of order sets using historical ordering data. We examine the current usage of order sets associated with Asthma Minor and Appendectomy Minor patients to understand how physicians are utilizing order sets, and how order set usage is associated with the time of ordering and characteristics of order sets. K-means clustering was applied to orders to generate evidence-based order sets that are learned from historical hospital data. We demonstrate that coverage rate of order sets and ordering efficiency can be increased through modifications of existing sets and creation of new sets.

Background: This study describes differences in the values of cerebrospinal fluid (CSF) white blood cell (WBC), glucose, and protein counts in infants less than 60 days of age with fever who were not proven to have viral or bacterial meningitis. Methods: Three independent retrospective medical record reviews were conducted using a population of infants less than 60 days of age who presented to the Emergency Department with fever. Full-term infants were included if a lumbar puncture was performed within 24 hours of admittance and bacterial or viral meningitis was not identified as the cause of fever. Results: A total of 1091 infants were included and grouped by week of age. Significant trends were found for CSF WBC and CSF protein with the highest values observed during the first week of life. Mean for CSF WBC was 8.63 cells/mm3 for infants aged 0 to 1 week and decreased for each age group ending with infants 8 weeks of age having a mean of 2.22 cell/mm3. For CSF protein, a similar trend was observed. No significant differences were found for CSF glucose. Conclusions: Significant differences exist for infants by week of age for CSF WBC and CSF protein. These values can be used to assist in interpreting laboratory findings and making management decisions for infants less than 60 days of age.

OBJECTIVES:The objective was to describe the emergency department (ED) resource burden of the spring 2009 H1N1 influenza pandemic at U.S. children's hospitals by quantifying observed-to-expected utilization.METHODS:The authors performed an ecologic analysis for April through July 2009 using data from 23 EDs in the Pediatric Health Information System (PHIS), an administrative database of widely distributed U.S. children's hospitals. All ED visits during the study period were included, and data from the 5 prior years were used for establishing expected values. Primary outcome measures included observed-to-expected ratios for ED visits for all reasons and for influenza-related illness (IRI).RESULTS:Overall, 390,983 visits, and 88,885 visits for IRI, were included for Calendar Weeks 16 through 29, when 2009 H1N1 influenza was circulating. The subset of 106,330 visits and 31,703 IRI visits made to the 14 hospitals experiencing the authors' definition of ED surge during Weeks 16 to 29 was also studied. During surge weeks, the 14 EDs experienced 29% more total visits and 51% more IRI visits than expected (p < 0.01 for both comparisons). Of ED IRI visits during surge weeks, only 4.8% were admitted to non-intensive care beds (70% of expected, p < 0.01), 0.19% were admitted to intensive care units (44% of expected, p < 0.01), and 0.01% received mechanical ventilation (5.0% of expected, p < 0.01). Factors associated with more-than-expected visits included ages 2-17 years, payer type, and asthma. No factors were associated with more-than-expected hospitalizations from the ED.CONCLUSIONS:During the spring 2009 H1N1 influenza pandemic, pediatric EDs nationwide experienced a marked increase in visits, with far fewer than expected requiring nonintensive or intensive care hospitalization. The data in this study can be used for future pandemic planning.

OBJECTIVETo determine the comparative effectiveness of common pleural drainage procedures for treatment of pneumonia complicated by parapneumonic effusion (ie, complicated pneumonia).DESIGNMulticenter retrospective cohort study.SETTINGForty children's hospitals contributing data to the Pediatric Health Information System.PARTICIPANTSChildren with complicated pneumonia requiring pleural drainage.MAIN EXPOSURESInitial drainage procedures were categorized as chest tube without fibrinolysis, chest tube with fibrinolysis, video-assisted thoracoscopic surgery (VATS), and thoracotomy.MAIN OUTCOME MEASURESLength of stay (LOS), additional drainage procedures, readmission within 14 days of discharge, and hospital costs.RESULTSInitial procedures among 3500 patients included chest tube without fibrinolysis (n = 1762), chest tube with fibrinolysis (n = 623), VATS (n = 408), and thoracotomy (n = 797). Median age was 4.1 years. Overall, 716 (20.5%) patients received an additional drainage procedure (range, 6.8-44.8% across individual hospitals). The median LOS was 10 days (range, 7-14 days across individual hospitals). The median readmission rate was 3.8% (range, 0.8%-33.3%). In multivariable analysis, differences in LOS by initial procedure type were not significant. Patients undergoing initial chest tube placement with or without fibrinolysis were more likely to require additional drainage procedures. However, initial chest tube without fibrinolysis was the least costly strategy.CONCLUSIONThere is variability in the treatment and outcomes of children with complicated pneumonia. Outcomes were similar in patients undergoing initial chest tube placement with or without fibrinolysis. Those undergoing VATS received fewer additional drainage procedures but had no differences in LOS compared with other strategies.

Objectives To assess the relationship between children's hospital readmission and the performance of child health systems in the states in which hospitals are located.Study design We conducted a retrospective cohort study of 197 744 patients 2 to 18 years old from 39 children's hospitals located in 24 states in the United States in 2005. Subjects were observed for a year after discharge for readmission to the same hospital. The odds of readmission were modeled on the basis of patient-level characteristics and state child health system performance as ranked by the Commonwealth Fund.Results A total of 1.8% of patients were readmitted within a week, 4.8% within a month, and 16.3% within 365 days. After adjustment for patient-level characteristics, the probability of readmission varied significantly between states (P=.001), and the likelihood of readmission during the ensuing year increased as the states' health system performance ranking improved. States in the best ranking quartile had a 2.03% higher readmission rate than states in the lowest quartile (P=.02); the same directional relationship was observed for readmission intervals from 1 to 365 days after discharge.Conclusions Hospital readmission rates are significantly related to the performance of the surrounding health care system. (J Pediatr 2010; 157: 98-102).

It is accepted that intravenous fluid (IVF) therapy can result in hospital-acquired dysnatremias in pediatric patients, with associated morbidity and mortality. There is interest in improving IVF therapy to prevent dysnatremias, but the optimal approach is controversial. In this study, we develop Natremia Deviation and Intravenous Renderer (NaDIR), a tool that preprocesses large volumes of electronic medical record data obtained from an academic pediatric hospital in order to analyze (1) IVF therapy, (2) the epidemiology of dysnatremias, and (3) the impact of IVFs on changes in serum sodium (ΔS(Na)). We then applied NaDIR to 3,256 inpatient records over a 3 month period, which revealed (1) a 19.9% incidence of dysnatremias, (2) a significant increase in lengths of stay associated with dysnatremias, and (3) a novel linear relationship between ΔS(Na) and IVF tonicity. This demonstrates that EMR data that can be readily analyzed to discover epidemiologic and predictive knowledge.

BACKGROUND. Children with complex chronic conditions depend on both their families and systems of pediatric health care, social services, and financing. Investigations into the workings of this ecology of care would be advanced by more accurate methods of population-level predictions of the likelihood for future hospitalization. METHODS. This was a retrospective cohort study. Hospital administrative data were collected from 38 children's hospitals in the United States for the years 2003–2005. Participants included patients between 2 and 18 years of age discharged from an index hospitalization during 2004. Patient characteristics documented during the index hospitalization or any previous hospitalization during the preceding 365 days were included. The main outcome measure was readmission to the hospital during the 365 days after discharge from the index admission. RESULTS. Among the cohort composed of 186856 patients discharged from the participating hospitals during 2004, the mean age was 9.2 years, with 54.4% male and 52.9% identified as non-Hispanic white. A total of 17.4% were admitted during the previous 365 days, and among those discharged alive (0.6% died during the admission), 16.7% were readmitted during the ensuing 365 days. The final readmission model exhibited a c statistic of 0.81 across all hospitals, with a range from 0.76 to 0.84 for each hospital. Bootstrap-based assessments demonstrated the stability of the final model. CONCLUSIONS. Accurate population-level prediction of hospital readmissions is possible, and the resulting predicted probability of hospital readmission may prove useful for health services research and planning.