Please see open positions here http://www.tsuilab.com/opportunities
Please see open positions here http://www.tsuilab.com/opportunities
Dr. Tsui’s upcoming keynote speech on “Predictive Modeling and its Applications in Healthcare” at 2017 Artificial Intelligence Conference, San Diego, CA, June 28-29. (http://artificialintelligence.conferenceseries.com)
Dr. Tsui in the news release for the project on infant mortality prediction and reduction in Allegheny County! (see news release here)
Congrats to Prof. Tsui for the award of the C-WIN project! (see news release from UPMC)
Congrats to Victor Ruiz for his conference abstract, entitled “Prediction of 30-Day All-Cause Pediatric Hospital Readmission from Medications and Healthcare Utilization“, accepted by the Biomedical and Healthcare Informatics conference, Orlando, FL, Feb 16-19, 2017
Congrats to Utkars Jain for his conference abstract, entitled "Constructing Eigenvalue/Coefficient Feature Vectors for Heart Attack Classification", accepted by the Biomedical and Healthcare Informatics conference, Orlando, FL, Feb 16-19, 2017
Tsui FR, Posada J, Shi L, Draper AJ, Xue D, Ruiz V, Su H, Kuan P, Mi F, Ye Y, and Ryan N, Predictive modeling for classification of positive valance system severity from initial psychiatric evaluation notes, 2016 CEGS NGRID Shared-Tasks and Workshop on Challenges in Natural Language Processing for Clinical Data, Chicago, IL, Nov. 11, 2016.
Jose D. Posada, Sergio Castro, Neal Ryan, Henk Harkema, and Fuchiang (Rich) Tsui, Social-context sentence annotation from psychiatric clinical reports, Science 2016, University of Pittsburgh, Pittsburgh, PA, Oct. 20, 2016
Adriana Johnson and Fuchiang (Rich) Tsui, Identifying Risk Factors for Morbidity and Mortality in the Pediatric Cardiac Intensive Care Unit: A Review of 634 Articles, Science 2016, University of Pittsburgh, Pittsburgh, PA, Oct. 20, 2016
Congrats to Dr. Rich Tsui for receiving 2016 Pitt Innovator Award, April 25, 2016.
Victor Ruiz's poster "Use of Diagnosis Related Groups to Predict All-Cause pediatric Hospital Readmission Within 30 days After Discharge" won the distinguished poster award at AMIA, Nov. 2015. Great work!
Ye Ye has just passed her oral comprehensive exam on November 6, 2015, sincere congratulations to her!
Jose Posada Aguilar's poster was accepted by science2015(http://www.science2015.pitt.edu/)!Poster Title:Inpatient 30-Day Readmission Prediction using cTAKES Poster abstract: Hospital readmissions are costive and potentially preventable. Approximately 3.3 million of adult patients were readmitted to hospitals within 30 days of discharge in 2011 with an estimated cost for the hospitals of 3.3 billion dollars according to the Center for Medicare and Medicaid Services. Up to 79% of readmissions can be potentially preventable. Most current tools for predicting 30-day readmissions only rely on structured data such as the number of inpatient visits in last 6 months, demographic information, etc. However, most detailed patient symptoms, findings and diagnosis are locked in clinicians’ reports that are in free-text form. We have developed the System for Hospital Adaptive Readmission Prediction and Management system (SHARP) that employs cTAKES, an open-source natural language processing (NLP) platform, to extract clinical findings from clinicians’ reports, and applies machine-learning algorithms to identify patients with risk of 30-day readmissions. We compared the performance of SHARP using cTAKES and MedLEE, a state-of-the-art NLP tool, by applying them to freetext clinician reports of heart failure (HF) patients at UPMC.
Amie gave outstanding talk at the NLM talk NLM Informatics Training Conference, the title of the talk is "Using Laboratory Data for Prediction of 30-Day Hospital Readmission: A Retrospective Study. Time: 2015 Jun 23 - 24. Bethesda, MD."Abstract: Prior work in readmission risk prediction has under-utilized laboratory data, which may provide valuable information about a patient’s condition. We aim to assess the contribution of laboratory data in predicting readmission risk. Preliminary work has focused on pediatric seizure, which has the highest volume of pediatric readmissions but no identified readmission risk factors. We used discharge diagnosis ICD-9 codes to identify seizure-specific visits to Children’s Hospital of Pittsburgh of UPMC during 2007-2012. Patients were considered readmitted if they returned to the hospital within 30-days post-discharge. We extracted features to summarize laboratory data for each patient. We used a training dataset (2007-2011) to rank features with information gain ratio and added features to a baseline model in order of rank. We kept features that improved prediction accuracy under 10-fold cross validation. A test (held-out) dataset (2012) was used to compare the AUROCs of the baseline model and the model with the added laboratory features. The addition of laboratory features significantly improved the prediction ability of the model, which suggests that laboratory data may be useful in identifying patients at risk of readmission. Ongoing work includes examining the contribution of laboratory data to readmission risk in adult heart failure patients.
Amie's poster "Using Laboratory Data for Prediction of 30-Day Hospital Readmission of Pediatric Seizure Patients" presented at AMIA 2014 Annual Symposium, Washington D.C. won the AMIA 2014 Distinguished Poster Award!Abstract: Laboratory data may provide valuable clinical information related to a patient’s risk of readmission. This study assessed the contribution of laboratory data in predicting readmission risk for pediatric seizure patients. We extracted basic summary features from laboratory data and selectively incorporated them into a baseline prediction model to determine if readmission prediction accuracy could be improved. We found that the added features from laboratory data increased the model’s AUROC from 0.55 to 0.63 with borderline significance (p-value=0.023).
Ye Ye got the best Student Paper 2nd Place in 2014 Biomedical Informatics Training Program Retreat.Ye Y, Tsui F-U, Wagner MM, Espino JU, Li Q. Influenza detection from emergency department reports using natural language processing and Bayesian network classifiers. JAMIA 2014 doi:10.1136/amiajnl-2013-001934. PMID: 24406261
Prof. Tsui interviewed by Pittsburgh Business Times for his research on hospital readmission reduction (http://www.bizjournals.com/pittsburgh/print-edition/2013/11/22/analytic-tool-helps-docs-make-better.html)
Jialan Que and Prof. Tsui received the Scientific Achievement award for Outstanding Research Articles in Biosurveillance, 2012 Annual International Society for Disease Surveillance (ISDS): Expanding Collaborations to Chart a New Course in Public Health Surveillance (December 4-5, 2012, San Diego, California)