Lab Users

Below are the researchers in the IMEDS lab and information about the research being completed. Click on the researchers name/organization to find out more about their work.

  • Auburn University, Richard Hansen: This project focuses on how to use machine learning techniques on training datasets created for acute kidney injury, acute liver injury, and myocardial infarction.
  • Beth Israel Deaconess Medical Center, Trevor E. Davis: This research aims to: (1) identify diseases comorbid with Juvenile Idiopathic Arthritis (JIA), and (2) study the potential effects of medications on the development of comorbidities.
  • Brown University, Nina Joyce: This research aims to: compare the use of the retrospective cohort to the NCC for the study of side effects associated with a rare drug exposure in a pediatric population.
  • Columbia University, Carol Friedman: Project objective is to develop and evaluate automated methods for detecting unknown adverse drug reactions using heterogeneous data, knowledge-based and statistical-based methods.
  • Columbia University, Mandev Gill: Project objective is to develop statistical methods for predicting health outcomes from high dimensional longitudinal health histories. Development of computationally efficient algorithms for estimation.
  • Columbia University, Thomas Nyberg: The goal of this work is to understand the kinds of inputs needed to verify the validity of implementations of statistical algorithms under the restriction that there is no known comparable reference implementation.
  • Columbia University, Zach Shahn: Methods development for predicting health outcomes from high dimensional longitudinal health histories and for performing causal inference on time series in the presence of unmeasured confounders using graphical models.
  • Columbia University, Chunhua Weng: Methods development for profiling patient populations from longitudinal observational data and to inform clinical trial patient selection with population summaries.
  • Center for Drug Evaluation and Research, FDA: Conducting the following training exercises: (1) gain familiarity with the OMOP common data model and with different database sources and (2) gain familiarity with observational studies methods applied by OMOP (e.g., reproduce results of the OMOP 2011/2012 experiment regarding Acute Myocardial Infarction endpoint.
  • Evidation Health, Foschini Group: By modeling the diffusion process of drug prescription and procedures to identify common patterns of diffusion and the factors that drive these patterns. This will help drug developers and health care providers to predict and optimize the update and spread of new treatments.
  • GlaxoSmithKline R&D: Use of benzodiazepines and the risk of hip/femur fracture: A methodological comparison across data sources, common data model approaches, and epidemiological designs. This research will evaluate the performance of Mini-Sentinel (MS) and Observational Medical Outcomes Partnership (OMOP) analytical tools using the example of Benzodiazepines (BZD) and Hip/Femur Fractures.
  • IBM; Kush Varshney, Dmitry Malioutov, Amin Emad: This research will develop a method for learning interpretable clinical prediction rules using sparse signal representation techniques.
  • Massachusetts Institute of Technology; Cynthia Rudin: This research will develop statistical learning methods to learn scoring systems from data, without any manual construction. Applying the method to the massive observational data at IMEDS will be useful for determining a simple, sparse set of correlative factors that underlie an adverse event, and will provide a suite of accurate scoring systems that will be used by physicians to make better treatment decisions.
  • Mini-Sentinel Operations Center, FDA: Mini-Sentinel is a pilot project sponsored by the FDA to create an active surveillance system to monitor the safety of FDA-regulated medical products. Mini-Sentinel uses existing administrative and electronic health care data from multiple collaborating and Data Partner institutions around the country.
  • Montana State University/University of New Mexico; Aurelien Mazurie and Christophe Lambert: This research will map medical events as reported in EHRs/claims data to specific patient-­centered outcomes such as frequency and length of stay in the hospital,change of physicians, and management of chronic, somatic  pathologies; and develop a scalable software framework to quickly and automatically cluster patients based on similarity of medical history, capable of processing records for 100M+ patients.
  • NIH Investigators: This research project will focus on analyzing and measuring the comprehensiveness of Electronic Health Record (EHR) in a given Integrated Data Repository (IDR) (e.g., datasets in the IMEDS Lab).
  • NIH; Ferdinand Dhombres: The research objective is to assess the potential risk in drug prescriptions during pregnancy, with respect to the new FDA standard.
  • OHDSI Investigator Group: This project focuses on software and infrastructure to support post-marketing evidence generation by creating summary statistics for associations between all combinations of medical products and outcomes, to be examined in the context of relevant evidence such as the literature and spontaneous reporting data to improve our understanding of how to interpret the safety and comparative effectiveness data derived from observational data.
  • Outcomes Insights, Inc: This project will run SQL queries using the datasets in the IMEDS lab as a testing framework to gather results for validation and benchmark tests.
  • RAND Corporation: The goal of this project is to discover discriminative sequential treatment patterns between patients with better than expected and worse than expected outcomes at given health status at incident diagnosis with a chronic condition.
  • Sanofi Investigators: This research will focus on evaluating one Mini-Sentinel protocol-based assessment using the IMEDS Research Lab.
  • Stanford University, Andrew Radin: This project will provide a set of tools to allow Parkinson’s researchers to conduct drug repurposing studies using clinical data.
  • Temple University SOM, Mark Weiner: This work will assess the data content and tools in the IMEDS Research Lab for use in a newly-developed educational curriculum designed to provide practical instruction on working with clinical data for research purposes.
  • University College London, Mijung Park: The goal of this research is to develop methods for improving the performance of personalised predictions of disease risk using statistical and machine learning tools.
  • University of Miami, Ramin Moghaddass and Skordilis Erotokritos: The primary objective of this research project is to provide a new statistical tool, which can help us leverage the dependencies within drugs/treatments and health outcomes to better understand causal effects. Our ultimate goal is to develop new statistical methods, which be used as a drug surveillance and control tool over the course of a treatment(s) period.
  • University of Pittsburgh, Translational Informatics Applied to Drug Safety (TRIADs) Investigator Group:
    • SOW#1 - This research will prioritize potential drug-drug interactions (PDDIs) involving warfarin, statins, and psychotropics by the need for pharmacoepidemiologic investigation. It will investigate the feasibility of using the large-scale observational data available in the IMEDS laboratory to 1) establishing the risk of exposure to PDDIs, and 2) develop population-specific adverse event prediction models that include PDDI exposure as a risk factor.
    • SOW #2 - The objective of this study is to derive reliable data that government payers and large healthcare organizations can combine with emerging data on adverse event risks and costs to conduct cost-effectiveness analyses for pre-emptive pharmacogenomics testing.
  • University of Texas-Austin, Ghosh Group:To develop methods which will extract medical concepts from the EHR data with minimal human supervision. The resulting phenotypes can be used for proactive patient management and prediction of progress of patients.
  • University of Wisconsin-Madison, Page Group: Aim 1: Make available as an OMOP method the approach in our AAAI-12 paper (Page et al., 2012). Aim 2: Test whether a Markov network approach can improve ROC area in recovery of adverse drug events beyond the performance achieved in our AAAI-12 paper (Page et al., 2012). Aim 3: Test whether continuous-time Bayesian networks (CTBNs) can outperform the approaches of Aims 1 and 2.
  • UCLA David Geffen School of Medicine, Department of Biomathematics: This project focuses on the integration of Bayesian methods and high performance computing to achieve better identification of drug risk.
  • Ivan Zorych, PhD: Investigate the usage of developed methods (disproportionality analysis, case-control, bayesian self-controlled case series, cohort, and bayesian logistic regression) for the sequential monitoring.

  • created at TagCrowd.com