Predicting three-year kidney graft survival in recipients with systemic lupus erythematosus.

Predicting the outcome of kidney transplantation is important in optimizing transplantation parameters and modifying factors related to the recipient, donor, and transplant procedure. As patients with end-stage renal disease (ESRD) secondary to lupus nephropathy are generally younger than the typical ESRD patients and also seem to have inferior transplant outcome, developing an outcome prediction model in this patient category has high clinical relevance. The goal of this study was to compare methods of building prediction models of kidney transplant outcome that potentially can be useful for clinical decision support. We applied three well-known data mining methods (classification trees, logistic regression, and artificial neural networks) to the data describing recipients with systemic lupus erythematosus (SLE) in the US Renal Data System (USRDS) database. The 95% confidence interval (CI) of the area under the receiver-operator characteristic curves (AUC) was used to measure the discrimination ability of the prediction models. Two groups of predictors were selected to build the prediction models. Using input variables based on Weka (a open source machine learning software) supplemented with additional variables of known clinical relevance (38 total predictors), the logistic regression performed the best overall (AUC: 0.74, 95% CI: 0.72-0.77)-significantly better (p < 0.05) than the classification trees (AUC: 0.70, 95% CI: 0.67-0.72) but not significantly better (p = 0.218) than the artificial neural networks (AUC: 0.71, 95% CI: 0.69-0.73). The performance of the artificial neural networks was not significantly better than that of the classification trees (p = 0.693). Using the more parsimonious subset of variables (six variables), the logistic regression (AUC: 0.73, 95% CI: 0.71-0.75) did not perform significantly better than either the classification tree (AUC: 0.70, 95% CI: 0.68-0.73) or the artificial neural network (AUC: 0.73, 95% CI: 0.70-0.75) models. We generated several models predicting 3-year allograft survival in kidney transplant recipients with SLE that potentially can be used in practice. The performance of logistic regression and classification tree was not inferior to more complex artificial neural network. Prediction models may be used in clinical practice to identify patients at risk.

Nursing informatics competencies required of nurses in Taiwan.

PURPOSE: In today's workplace, nurses are highly skilled professionals possessing expertise in both information technology and nursing. Nursing informatics competencies are recognized as an important capability of nurses. No established guidelines existed for nurses in Asia. This study focused on identifying the nursing informatics competencies required of nurses in Taiwan. METHODS: A modified Web-based Delphi method was used for two expert groups in nursing, educators and administrators. Experts responded to 323 items on the Nursing Informatics Competencies Questionnaire, modified from the initial work of Staggers, Gassert and Curran to include 45 additional items. Three Web-based Delphi rounds were conducted. Analysis included detailed item analysis. Competencies that met 60% or greater agreement of item importance and appropriate level of nursing practice were included. RESULTS: N=32 experts agreed to participate in Round 1, 23 nursing educators and 9 administrators. The participation rates for Rounds 2 and 3=68.8%. By Round 3, 318 of 323 nursing informatics competencies achieved required consensus levels. Of the new competencies, 42 of 45 were validated. A high degree of agreement existed for specific nursing informatics competencies required for nurses in Taiwan (97.8%). CONCLUSIONS: This study provides a current master list of nursing informatics competency requirements for nurses at four levels in the U.S. and Taiwan. The results are very similar to the original work of Staggers et al. The results have international relevance because of the global importance of information technology for the nursing profession.

Predicting the start week of respiratory syncytial virus outbreaks using real time weather variables.

BACKGROUND: Respiratory Syncytial Virus (RSV), a major cause of bronchiolitis, has a large impact on the census of pediatric hospitals during outbreak seasons. Reliable prediction of the week these outbreaks will start, based on readily available data, could help pediatric hospitals better prepare for large outbreaks. METHODS: Naïve Bayes (NB) classifier models were constructed using weather data from 1985-2008 considering only variables that are available in real time and that could be used to forecast the week in which an RSV outbreak will occur in Salt Lake County, Utah. Outbreak start dates were determined by a panel of experts using 32,509 records with ICD-9 coded RSV and bronchiolitis diagnoses from Intermountain Healthcare hospitals and clinics for the RSV seasons from 1985 to 2008. RESULTS: NB models predicted RSV outbreaks up to 3 weeks in advance with an estimated sensitivity of up to 67% and estimated specificities as high as 94% to 100%. Temperature and wind speed were the best overall predictors, but other weather variables also showed relevance depending on how far in advance the predictions were made. The weather conditions predictive of an RSV outbreak in our study were similar to those that lead to temperature inversions in the Salt Lake Valley. CONCLUSIONS: We demonstrate that Naïve Bayes (NB) classifier models based on weather data available in real time have the potential to be used as effective predictive models. These models may be able to predict the week that an RSV outbreak will occur with clinical relevance. Their clinical usefulness will be field tested during the next five years.

Specialist discrimination of toxic exposure severity at a poison control center.

BACKGROUND: Perceived severity has been shown to affect decision-making processes in telephone triage. However, the accuracy of specialists in poison information's (SPIs') perceptions of severity of poison exposures is unknown. OBJECTIVE: The purpose of this study was to describe the ability of SPIs to predict severity of medical outcome on the basis of the information obtained during the initial poison control center's phone call. METHODS: This study analyzed 22,576 cases of human poison exposure in one regional poison control center. At the time of the initial call, SPIs assigned a predicted severity rating. SPIs then assigned a medical outcome rating when closing each case. Animal exposures not coded, not followed, and confirmed nonexposures were excluded. RESULTS: For overall SPI's discrimination of more severe versus less severe cases, A(z) = 0.94 with asymmetric 95% confidence intervals (0.87, 0.97), indicating excellent discrimination. The sensitivity of SPIs in discriminating a major effect from any other effect was 0.62. The false-negative rate for discrimination of a moderate, major, or fatal effect from a minor effect or no effect was 0.32, with sensitivity = 0.68. CONCLUSIONS: The overall ability of the SPIs to predict exposure severity is excellent but less accurate with less frequently encountered, more severe cases. A better understanding of SPI's decision-making processes, including the relationship between perceived severity and decision-making strategies, is necessary for the development of educational strategies and decision support technologies.

Building knowledge for poison control: the novel pairing of communication analysis with data mining methods.

As information systems become increasingly integrated with health care delivery, vast amounts of clinical data are stored. Knowledge discovery and data mining methods are potentially powerful for the induction of knowledge models from this data relevant to nursing outcomes. However, an important barrier to the widespread application of these methods for induction of nursing knowledge models is that important concepts relevant to nursing outcomes are often unrepresented in clinical data. For instance, communication approaches are not necessarily consciously chosen by nurses, yet they are known to impact multiple clinical outcomes including satisfaction, pain and symptom response, recovery, physiological change (e.g., blood pressure), and adherence. Decisions about communication behaviors are likely intuitive and instantaneously made in response to cues offered by the patient. For this reason, among others, important choices and actions of nurses are not routinely documented. And so for many clinical outcomes relevant to nursing, important concepts such as communication are not represented in clinical data repositories. In studying poison control center outcomes, it is important to consider not only routinely documented clinical data, but the communication processes and verbal cues of both patient and SPI. In a novel approach, our current study of poison control center outcomes pairs a qualitative study of the communication patterns of SPIs and callers to a regional poison control center, with predictive modeling of poison control center outcomes using knowledge discovery and data mining methods. This three year study, currently in progress, pairs SPI-caller communication analysis with predictive models resulting from the application of knowledge discovery and data mining methods to three years' of archived clinical data. The results will form a hybrid model and the basis for future decision support interventions that leverage knowledge about both implicit and explicit factors that contribute to poison control center outcomes.

Patient-level analysis of outcomes using structured labor and delivery data.

This paper presents methods for identifying and analyzing associations among nursing care processes, patient attributes, and patient outcomes using unit-level and patient-level representations of care derived from computerized nurse documentation. The retrospective, descriptive analysis included documented nursing events for 900 Labor and Delivery patients at three hospitals over the 2-month period of January and February 2006. Two models were used to produce quantified measurements of nursing care received by each patient. The first model considered only the hourly census of nurses and patients. The second model considered the size of nurses' patient loads as represented by computerized nurse-entered documentation. Significant relationships were identified between durations of labor and nursing care scores generated by the second model. In addition to the clinical associations identified, the study demonstrated an approach with global application for representing the amount of nursing care received at the individual patient level in analyses of patient outcomes.

Modeling the distribution of Nursing Effort using structured Labor and Delivery documentation.

Our study objectives included the development and evaluation of models for representing the distribution of shared unit-wide nursing care resources among individual Labor and Delivery patients using quantified measurements of nursing care, referred to as Nursing Effort. The models were intended to enable discrimination between the amounts of care delivered to patient subsets defined by attributes such as patient acuity. For each of five proposed models, scores were generated using an analysis set of 686,402 computerized nurse-documented events associated with 1093 patients at three hospitals during January and February 2006. Significant differences were detected in Nursing Effort scores according to patient acuity, care facility, and in scores generated during shift change versus non-shift change hours. The development of nursing care quantification strategies proposed in this study supports outcomes analysis by establishing a foundation for measuring the effect of patient-level nursing care on individual patient outcomes.

Nurse residency program implementation: the Utah experience.

Recently, nurse residency programs have been shown to improve satisfaction and enhance the retention of new graduate nurses, offering one solution for hospital executives, administrators, and managers searching for innovative ways to address nursing staff shortages. This article identifies crucial lessons that will assist leaders in designing and implementing a nurse residency program in their own institutions. The lessons are drawn from the experience of the successful University of Utah program. Four important practical components of such programs are described: an adaptive curriculum, promotion of autonomy, mentoring, and meeting the needs of participants with associate degrees. Although the lessons are based on the perspective of one nurse residency program, they hold import for the design of nurse residency programs in diverse settings.

Forecasting hospital census at a tertiary care children's hospital.

Developing a forecasting tool for patient census allows for improved staffing, better resource utilization and mobilization, and improved timing of educational campaigns around the disease control process. Using a neural network approach we evaluated several different models and variables for predicting patient census prospectively. These initial studies enabled selection of a subset of predictor variables and show that different network models, and variables must be used based on the season.

Classification of smoking cessation status with a backpropagation neural network.

This study examined the ability of a backpropagation neural network (BPNN) classifier to distinguish between current and former smokers in the 2000 National Health Interview Survey (NHIS) sample adult file. The BPNN classifier performance exceeded that of random chance, with asymmetric 95% confidence intervals for A(z) (area under receiver operating characteristic curve)=(0.7532, 0.7790). Separation of current and former smokers was imperfect, as illustrated by the receiver operating characteristic (ROC) curve. Additionally, performance did not exceed that of a comparison classifier created using logistic regression. Attribute subset selection identified three novel attributes related to smoking cessation status. This study establishes the ability of backpropagation neural networks to classify a complex health behavior, smoking cessation. It also illustrates the hypothesis-generating capacity of data mining methods when applied to large population-based health survey data. Ultimately, BPNN classifiers of smoking cessation status may be useful in decision support systems for smoking cessation interventions.