Shital Shah

Comparing the accuracy of syndrome surveillance systems in detecting influenza-like illness: GUARDIAN vs. RODS vs. electronic medical record reports

BACKGROUND A highly sensitive real-time syndrome surveillance system is critical to detect, monitor, and control infectious disease outbreaks, such as influenza. Direct comparisons of diagnostic accuracy of various surveillance systems are scarce. OBJECTIVE To statistically compare sensitivity and specificity of multiple proprietary and open source syndrome surveillance systems to detect influenza-like illness (ILI). METHODS A retrospective, cross-sectional study was conducted utilizing data from 1122 patients seen during November 1–7, 2009 in the emergency department of a single urban academic medical center. The study compared the Geographic Utilization of Artificial Intelligence in Real-time for Disease Identification and Alert Notification (GUARDIAN) system to the Complaint Coder (CoCo) of the Real-time Outbreak Detection System (RODS), the Symptom Coder (SyCo) of RODS, and to a standardized report generated via a proprietary electronic medical record (EMR) system. Sensitivity, specificity, and accuracy of each classifiers ability to identify ILI cases were calculated and compared to a manual review by a board-certified emergency physician. Chi-square and McNemars tests were used to evaluate the statistical difference between the various surveillance systems.ResultsThe performance of GUARDIAN in detecting ILI in terms of sensitivity, specificity, and accuracy, as compared to a physician chart review, was 95.5%, 97.6%, and 97.1%, respectively. The EMR-generated reports were the next best system at identifying disease activity with a sensitivity, specificity, and accuracy of 36.7%, 99.3%, and 83.2%, respectively. RODS (CoCo and SyCo) had similar sensitivity (35.3%) but slightly different specificity (CoCo = 98.9%; SyCo = 99.3%). The GUARDIAN surveillance system with its multiple data sources performed significantly better compared to CoCo (χ2 = 130.6, p < 0.05), SyCo (χ2 = 125.2, p < 0.05), and EMR-based reports (χ2 = 121.3, p < 0.05). In addition, similar significant improvements in the accuracy (>12%) and sensitivity (>47%) were observed for GUARDIAN with only chief complaint data as compared to RODS (CoCo and SyCo) and EMR-based reports. CONCLUSION In our study population, the GUARDIAN surveillance system, with its ability to utilize multiple data sources from patient encounters and real-time automaticity, demonstrated a more robust performance when compared to standard EMR-based reports and the RODS systems in detecting ILI. More large-scale studies are needed to validate the study findings, and to compare the performance of GUARDIAN in detecting other infectious diseases.

Data mining based decision-making approach for predicting survival of kidney dialysis patients

Abstract Dialysis care is particularly complex and multiple factors may influence patient survival. The cost of such treatment for end stage kidney disease is high and needs attention for reducing it. Individual patient survival may depend on an intricate interrelationship between various demographic and clinical variables, medications, medical interventions and the dialysis treatment prescription. In this research, a data mining approach is used to extract knowledge regarding the interactions between the features and the outcome. There exist a complex and contradictory relationships among data mining rules that are difficult to interpret and implement. To resolve these conflicts a decision-making algorithm is developed using sixteen different classifiers. The decision-making algorithm employs simple and weighted voting schemes. Thus in this paper, a hybrid data mining enhanced decision making approach is used for predictions of an individual patient surviving beyond the median survival time. The concepts introduced in this research have been applied and tested using data collected at four dialysis sites.

Comparison of Static versus Dynamic Ultrasound for the Detection of Endotracheal Intubation

Introduction In the emergency department setting, it is essential to rapidly and accurately confirm correct endotracheal tube (ETT) placement. Ultrasound is an increasingly studied modality for identifying ETT location. However, there has been significant variation in techniques between studies, with some using the dynamic technique, while others use a static approach. This study compared the static and dynamic techniques to determine which was more accurate for ETT identification. Methods We performed this study in a cadaver lab using three different cadavers to represent variations in neck circumference. Cadavers were randomized to either tracheal or esophageal intubation in equal proportions. Blinded sonographers then assessed the location of the ETT using either static or dynamic sonography. We assessed accuracy of sonographer identification of ETT location, time to identification, and operator confidence. Results A total of 120 intubations were performed: 62 tracheal intubations and 58 esophageal intubations. The static technique was 93.6% (95% confidence interval [CI] [84.3% to 98.2%]) sensitive and 98.3% specific (95% CI [90.8% to 99.9%]). The dynamic technique was 92.1% (95% CI [82.4% to 97.4%]) sensitive and 91.2% specific (95% CI [80.7% to 97.1%]). The mean time to identification was 6.72 seconds (95% CI [5.53 to 7.9] seconds) in the static technique and 6.4 seconds (95% CI [5.65 to 7.16] seconds) in the dynamic technique. Operator confidence was 4.9/5.0 (95% CI [4.83 to 4.97]) in the static technique and 4.86/5.0 (95% CI [4.78 to 4.94]) in the dynamic technique. There was no statistically significant difference between groups for any of the outcomes. Conclusion This study demonstrated that both the static and dynamic sonography approaches were rapid and accurate for confirming ETT location with no statistically significant difference between modalities. Further studies are recommended to compare these techniques in ED patients and with more novice sonographers.

A Comparison of Insulin Doses for the Treatment of Hyperkalemia in Patients with Renal Insufficiency

To compare the safety and efficacy of 5 units versus 10 units of insulin for the treatment of hyperkalemia in patients with renal insufficiency.

Medicaid beneficiaries who continue to use the ED: a focus on the Illinois Medical Home Network

OBJECTIVES Frequent, nonurgent emergency department use continues to plague the American health care system through ineffective disease management and unnecessary costs. In 2012, the Illinois Medical Home Network (MHN) was implemented to, in part, reduce an overreliance on already stressed emergency departments through better care coordination and access to primary care. The purpose of this study is to characterize MHN patients and compare them with non-MHN patients for a preliminary understanding of MHN patients who visit the emergency department. Variables of interest include (1) frequency of emergency department use during the previous 12 months, (2) demographic characteristics, (3) acuity, (4) disposition, and (5) comorbidities. METHODS We performed a retrospective data analysis of all emergency department visits at a large, urban academic medical center in 2013. Binary logistic regression analyses and analysis of variance were used to analyze data. RESULTS Medical Home Network patients visited the emergency department more often than did non-MHN patients. Medical Home Network patients were more likely to be African American, Hispanic/Latino, female, and minors when compared with non-MHN patients. Greater proportions of MHN patients visiting the emergency department had asthma diagnoses. Medical Home Network patients possessed higher acuity but were more likely to be discharged from the emergency department compared with non-MHN patients. CONCLUSIONS This research may assist with developing and evaluating intervention strategies targeting the reduction of health disparities through decreased use of emergency department services in these traditionally underserved populations.

Relabeling algorithm for retrieval of noisy instances and improving prediction quality

A relabeling algorithm for retrieval of noisy instances with binary outcomes is presented. The relabeling algorithm iteratively retrieves, selects, and re-labels data instances (i.e., transforms a decision space) to improve prediction quality. It emphasizes knowledge generalization and confidence rather than classification accuracy. A confidence index incorporating classification accuracy, prediction error, impurities in the relabeled dataset, and cluster purities was designed. The proposed approach is illustrated with a binary outcome dataset and was successfully tested on the standard benchmark four UCI repository dataset as well as bladder cancer immunotherapy data. A subset of the most stable instances (i.e., 7% to 51% of the sample) with high confidence (i.e., between 64%-99.44%) was identified for each application along with most noisy instances. The domain experts and the extracted knowledge validated the relabeled instances and corresponding confidence indexes. The relabeling algorithm with some modifications can be applied to other medical, industrial, and service domains.

Comparison of color flow with standard ultrasound for the detection of endotracheal intubation

Introduction: Intubation is a frequently performed procedure in emergency medicine that is associated with significant morbidity and mortality when unrecognized esophageal intubation occurs. However, it may be difficult to visualize the endotracheal tube (ETT) in some patients. This study assessed whether the addition of color Doppler was able to improve the ability to visualize the ETT location. Methods: This study was performed in a cadaver lab using three different cadavers chosen to represent varying neck circumference. Cadavers were randomized to tracheal or esophageal intubation. Blinded sonographers then assessed the location of the ETT using either grayscale or color Doppler imaging. Accuracy of sonographer identification of ETT location, time to identification, and operator confidence were assessed. Results: One hundred and fifty intubations were performed and each was assessed by both standard and color Doppler techniques. There were 78 tracheal intubations and 72 esophageal intubations. The standard technique was 99.3% (95% CI 96.3 to 99.9%) accurate. The color flow technique was also 99.3% (95% CI 96.3 to 99.9%) accurate. The mean operator time to identification was 3.24 s (95% CI 2.97 to 3.51 s) in the standard approach and 5.75 s (95% CI 5.16 to 6.33 s) in the color flow technique. The mean operator confidence was 4.99/5.00 (95% CI 4.98 to 5.00) in the standard approach and 4.94/5.00 (95% CI 4.90 to 4.98) in the color flow technique. Conclusion: When added to standard ultrasound imaging, color flow did not improve accuracy or operator confidence for identifying ETT location and resulted in a longer examination time.

Variation in the accuracy of ultrasound for the detection of intubation by endotracheal tube size

Introduction: Rapid and accurate confirmation of endotracheal tube (ETT) placement is a fundamental step in definitive airway management. Multiple techniques with different limitations have been reported. Recent studies have evaluated the accuracy, time to performance, and physician confidence for ultrasound in both cadaveric models and live patients. However, no study to date has measured the effect of ETT size. Our study is the first to measure the accuracy of ultrasound for ETT confirmation based on ETT size. Methods: This study was performed in a cadaver lab using three different cadavers chosen to represent varying neck circumferences. Cadavers were intubated in a random sequence with respect to both the location of intubation (i.e., tracheal vs esophageal) and sizes of ETT. Three ETT sizes were utilized: 6.0‐, 7.0‐, and 8.0‐mm. Blinded sonographers assessed the location of the ETT using the static technique. Accuracy of sonographer identification, time to identification, and operator confidence were assessed. Results: 453 assessments were performed. Overall, ultrasound was 99.1% (95% CI 97.8% to 99.7%) accurate in identification of correct location of intubation. The mean time to placement was 6.45 s (95% CI 5.62 to 7.28). The mean operator confidence level was 4.72/5.0 (95% CI 4.65 to 4.78). There was no significant difference between ETT sizes with respect to any of the outcomes. Conclusion: The diagnostic accuracy of ultrasound for ETT confirmation did not vary with the use of different ETT sizes. Further studies are needed to determine if the accuracy would change with more novice providers or in specific patient populations.

The Impact of Weather on Influenza-like Illness Rates in Chicago

Description of a statistical model to account for weather variation in influenza-like illness surveillance.