Duke AI Health and +Data Science Launch Successful Proposal Studio Series

By Rabail Baig, MA

Keeping up with the pace of research in health data science is challenging at the best of times, and the COVID-19 pandemic has not made things any easier. For this reason, Duke AI Health and the Duke +Data Science (+DS) program worked together this spring to launch the Proposal Studio Virtual Learning Experiences (vLE). The Proposal Studios sessions were designed to help investigators develop effective, successful proposals for research projects involving health data science.

Series host Michael Pencina, PhD, Vice Dean for Data Science and Information Technology, observes that the series was a departure from more customary approaches.

“Instead of introducing biostatistical ideas one at a time, what we did with Proposal Studios was a little bit different and more real-world,” notes Pencina. “We found that discussing actual clinical research questions with data science experts and considering multiple methods and the nuances of the best fit of methods to problems often resulted in a dialog that went beyond the basic concepts taught in the classroom.”

Co-hosted by Shelley Rusincovitch, MMCi, Associate Director of Informatics at Duke AI Health, the Proposal Studios were formatted as a series of one-hour sessions, featuring small teams of Duke investigators who discussed their concepts with data science experts from across the University. The sessions, which Pencina describes as comprising a practical combination of education and research, also provided participants with a forum to share their experiences with the broader Duke community.

“Instead of having the hammer working for the nails, we attempted to address real problems through approaches designed to best answer these real-world questions,” Pencina said. “Our goal was to help ensure that the proposals our clinical investigators put forward will be one step closer to a competitive grant application.”

From March through April of 2021, +DS held four successful proposal studios, assisting 13 investigators to develop scientific proposals. Open to anyone within the Duke community, the series attracted a total of 129 attendees and averaged 32 audience members per vLE.  

We spoke with series hosts Michael Pencina and Shelley Rusincovitch about the success of the Proposal Studios and what is next for the program.

Origin of the Proposal Studio concept

The AI Health Proposal Studios aim to provide a structured opportunity for investigators to engage with Duke’s top data science expertise and thought leadership around the scientific components of their proposals. It was conceptualized in the second half of 2020 with the overarching goal of fostering high-impact, rigorous, and competitive proposals for scientific awards.

The Proposal Studio concept builds on success from our experience in the Forge Design Workshops (although in a necessarily virtual format to begin). In this photo from 2018, Michael Pencina and Shelley Rusincovitch stand in front of the output from a design workshop team.

In late 2020, we issued a call for applications for the AI Health Proposal Studios. Shortly after the announcement, we received an astounding 26 applications, which far surpassed our expectations and revealed a high level of interest in proposal development and early-stage project ideation. It was after the overwhelmingly positive response that we decided to frame some of these engagements as virtual learning experiences (vLEs) in order to share this activity with a larger community of learners at Duke.

Duke +DS’s collaboration with Duke AI Health

The overarching mission of Duke AI Health is to harness the power of artificial intelligence, machine learning, and related quantitative fields for the benefit of medical research, healthcare delivery, and the health of individuals and communities. The Duke +Data Science (+DS) program extends this mission to an educational focus, with two key individuals within AI Health – Ricardo Henao, PhD (associate director of AI Health) and Shelley Rusincovitch (associate director of informatics for AI Health) – providing scientific and operational leadership for +DS.

Plus Data Science is a Duke-wide program that operates in partnership with departments, schools, and institutes to enable faculty, students, and staff to employ data science at a level tailored to their needs, level of expertise, and interests. Since its launch in 2018, +DS has supported engagement and learning within the Duke community through 115 co-curricular learning experiences, eight machine learning schools, and 72 students in substantive applied projects. The +DS Coursera course “Introduction to Machine Learning” has reached more than 129,000 learners.

Series rationale and goals

Through the Proposal Studios, our goal is to expand access to state-of-the-art data science resources and national leadership and expertise for Duke investigators. We’re proud of the success this spring in building a pathway from small, internally-supported analyses into larger external opportunities and empowering Duke investigators to translate small, limited research pilots into robust, higher-impact proposals and analyses.

We’ve seen a strong response to the concept and initial set of learning experiences, and we plan to continue building on this success. We plan to hold two new proposal studio vLEs in June with Maciej Mazurowski, PhD, Associate Professor of Radiology, Electrical and Computer Engineering, and Biostatistics and Bioinformatics at Duke, contributing his data science expertise.

The Spring Proposal Studio vLE Series

 We are delighted by the strong response to the Proposal Studios, and we especially thank the investigators and quantitative experts for sharing their time and expertise.

The first Proposal Studio vLE, Part 1 of Structured Data Analyses, was launched on March 29, 2021, with quantitative experts Ben Goldstein, PhD, and Matthew Engelhard, PhD, joining Dr. Pencina as host. The session featured the Department of Surgery’s João Ricardo Vissoci, PhD. who discussed his proposal concept for traumatic brain injury in Tanzania; the School of Nursing’s Michael Cary, PhD, RN, who discussed hip fracture readmissions; and the Duke Global Health Institute’s Osondu Ogbuoji, MBBS, MPH, who talked about claims analysis for Ghana and Nigeria.

The subsequent Proposal Studio vLE on Omics Analyses was held on April 5, 2021. The Duke Molecular Physiology Institute’s Allison Ashley-Koch, PhD, and William E. Kraus, MD,  were joined by Duke Cancer Institute’s Anthony D. Sung, MD, to discuss their respective proposal concepts on genomic analysis related to sickle cell anemia, lifestyle intervention adherence, and transplant optimization with quantitative experts David Page, PhD, Chair of Biostatistics and Bioinformatics, and Svati Shah, MD, MHS, Associate Dean of Genomics, joining Dr. Pencina.

The Proposal Studio on Novel Data Sources and Platform Development took place on April 7, 2021, with quantitative expertise provided by Drs. Henao and Pencina. The session featured presentation by the Global Health Institute’s Catherine Stanton, MD, MS, who discussed her proposal concept for an injury repository; Duke Anesthesiology’s Marie-Louise Meng, MD, who shed light on cardiovascular risk in pregnancy; and Noa Segall, PhD, also from the Department of Anesthesiology, who drew attention to inpatient hypoglycemia.

The concluding Proposal Studio vLE, Part 2 of Structured Data Analyses on April 21, 2021, featured proposal concepts on AI-based pulmonary function testing, sickle cell severity, long-term follow up for diabetic retinopathy, and ovarian cancer care recommender, presented by Duke Health pulmonologist Yuh-Chin Huang, MD, MHS, the School of Nursing’s Mariam Kayle, PhD, RN, the Department of Ophthalmology’s Durga Borkar, MD, and the Duke Cancer Institute’s Haley A. Moss, MD, MBA, respectively. The session’s quantitative expertise was contributed by the Department of Statistical Science’s Samuel Berchuck, PhD, and Dr. Pencina.