Using Big Data as a Diagnostic Tool

Applied Data

Applied Data Sciences Center

Researchers in the Applied Data Sciences Center (ADSC) take vast amounts of health and related data for “cleaning and integration” and make it ready for analysis and research by industry, academic and government partners.

Data analytics and machine learning then enable the better identification and prediction of diseases.

The year 2018 saw significant growth and progress as the Applied Data Sciences Center (ADSC) addressed its long-term goal to drive insights into cardio-metabolic health and nutrition through multi-organizational and trans-disciplinary collaborations that integrate and analyze disparate data sources.

Five-Year Strategy

The ADSC’s five-year strategy is to become a nationally-recognized collaboration hub located in Indianapolis driving multi-organizational and trans-disciplinary projects, which integrates and analyzes disparate data sources to provide insights into cardio-metabolic health, including community health, nutrition and related problems. Strategic priorities, include:

  • Rely on key technology platforms or partnerships for data access, wrangling and integration, advanced analytics, and capabilities for the direct capture of both digital and biological patient data.
  • Progress towards a broader center strategy with multiple principal investigators.
  • Ensure the financial sustainability of the center.
  • Increase collaborations for data access or problem insights.
  • Initiate a new public health collaboration addressing cardio-metabolic health issues

Ways we use health data to improve health

The center grew from two to five staff members, providing additional capacity for its multiple ongoing collaborations and pilots. The number of active collaborations remained at five throughout most of the year, but further pilots with other organizations to explore additional data sets or technology platforms are complete or in process. 

Research on type 2 diabetes

Collaboration at the ADSC

From left: Meeta Pradhan, Ph.D., senior data scientist, and Dan Robertson, Ph.D., director, Applied Data Sciences Center (ADSC), discuss a graph depicting progression of type 2 diabetes to associated comorbidities, such as kidney disease, generated in collaboration with IBRI’s research partners from analysis of electronic health records.

Partners: Eli Lilly and Company and Roche Diabetes Care GmbH

The type 2 diabetes real-world data collaboration project to explore disease progression, patient stratification, digital diagnostics and new therapeutic pathways was renewed with funding from partners Eli Lilly and Company and Roche Diabetes Care GmbH.

A paper describing the data cleaning process applied to electronic health records (EHRs) was submitted to the Journal of the American Medical Informatics Association (JAMIA). An article by researchers from Roche, IBM, the IBRI, Lilly and Regenstrief Institute describing how real-world patient data can better predict diabetes-related kidney disease in patients with the chronic disease was published in the January 2019 issue of Nature Medicine.

This project will continue in 2019 and will expand to include additional use cases and, potentially, other datasets. This collaboration is expected to support the new Lilly Diabetes Center of Excellence at the IBRI, as well as the collection of patient samples through the Indiana BioBank, a panel of biological samples linked to the corresponding EHRs.

Work continued to extract an updated version of the patient dataset, making it available for further research. This project was showcased at multiple forums, including:

  • IU School of Medicine Fourth Annual Diabetes Symposium
  • Lilly’s Grand Rounds Employee Event
  • Regenstrief Institute’s Work in Progress lecture series
  • BioCrossroads' Indiana Life Sciences Summit
Nature Medicine, January 2019

Nature Medicine, January 2019

More peer reviewed news and research

Journal of the American Medical Informatics Association

Forthcoming: A paper has been prepared and submitted to the publication that describes the data cleaning process applied to electronic health records.

Nature Medicine

January 2019: The journal published an article written by our researchers on how real-world patient data can help predict diabetes-related chronic kidney disease.

Exhibition on toxicological safety predictions

Partners: Corteva AgriScience and Eli Lilly and Company

The collaboration with Corteva AgriScience and Lilly to improve toxicological safety predictions for newly discovered molecules increased in visibility. Following the 18-month renewal last year, the team focused on platform enhancements, including new scientific visualizations and the promotion of external visibility by presenting an exhibition at the Society of Toxicology Annual Conference in San Antonio, Tex.

This meeting and subsequent conversations led an additional pharmaceutical company to analyze one of their internal datasets using this platform. Further conversations are ongoing with other pharmaceutical and agri-chemical companies where this capability may be useful.

A paper describing the platform and the underlying science was submitted to Toxicological Sciences, and this collaboration won “Best Industry Collaboration” at the international Agrow Awards 2018.

Our platforms for collecting health data

Partners: MavenSphere, Indiana University School of Medicine and Stamford Health

The ADSC collaboration with the Indianapolis technology company MavenSphere, to develop a digital biomarker platform that collects patient data via smartphones, moved forward last year with the collection of new patient data from both Indiana University School of Medicine in Indianapolis and Stamford Health in Connecticut. Based on these successes, the platform was nominated for a Clinical and Translational Sciences Institute (CTSI) award from IU School of Medicine.

This project was also the runner-up in BioCrossroads’ inaugural Inject Tech Challenge with an award of $6,000, and it was highlighted in a talk at the 2018 BioIT World & Expo Conference in Boston. Proposals for further research applications that will generate additional revenue are in process.

Daniel Robertson, Ph.D., Director, Applied Data Sciences Center

Daniel Robertson, Ph.D.
Director, Applied Data Sciences Center

New data tech and public health initiatives

Partners: CAPriCORN, LifeOmic, Virtusa and Cardinal Health

Multiple pilots targeting data access or the assessment of strategic technology partnerships were initiated, including access to data from CAPriCORN (the Chicago Area Patient-Centered Outcomes Research Network) for enhanced patient diversity data, an assessment of Indianapolis-based LifeOmic’s precision health platform to support research efforts, and a joint collaboration with Massachusetts-based Virtusa and Fuse by Cardinal Health in Ohio, to assess both Virtusa’s technology platform and Fuse’s synthetic EHR data.

Conversations are ongoing for additional public health collaborations, including joint efforts with the Indiana Primary Health Care Association (IPHCA) for access to low-income and diverse patient data, and with the State of Indiana’s Management Performance Hub (MPH) aiming to connect patient clinical data with socio-economic and claims data.

Every gift makes an impact.

Our research is made possible by your generous gift and it will help lead to solutions to major health challenges affecting almost one out of every three people globally.

donate now

Subscribe to our email list

Interested in: