Meet the NEJM Sprint Challenge Winners

Published

NEJM is hosting a summit on Aligning Incentives for Sharing Clinical Data on April 3-4, 2017. Leading up to the summit, the SPRINT Data Analysis Challenge was held between November 2016 and March 2017 to explore the potential benefits of sharing data from clinical trials. Individuals and groups all over the world were invited to analyze the dataset underlying the SPRINT article, which was published in 2015. A group of medical students known as team ‘Renality Check’ from Boston University School of Medicine earned second place out of 200 qualifying teams of researchers from around the world in the recent NEJM SPRINT Challenge. The team members are: Rahul Aggarwal, Nicholas Chiu, Sang Myung Han, Haares Mirzan, Jason Park, Ben Petrie, and Jackson Steinkamp. NEJM Editorial Fellow Ramya Ramaswami asked them about this impressive accomplishment.

How did you find out about the NEJM SPRINT Challenge?

While perusing the New England Journal of Medicine, we stumbled upon a link for the SPRINT challenge. As a group of friends with a variety of backgrounds in data analysis and research, we thought it would be a unique experience to look through the data set and see what we could find. We saw the challenge as an opportunity to undertake an independent project as students and to hone our skills in data analysis and literature evaluation. In particular, hypothesis generation is a part of the scientific process that medical students do not often have the opportunity to tackle independently, since we frequently join existing projects with faculty members and mentors who already have questions of their own. Also, the preclinical years of medical school largely focus on memorization of facts, so we were excited to have an opportunity to use the more logical and creative parts of our brains in this contest. We did not initially think we would be able to place in this competition — it was a very welcome surprise that we ultimately did!

Your submission for the SPRINT Challenge was about chronic kidney disease. How did you decide on your question for the challenge?

In short, we let the data guide us. Our group was comprised of two subteams: the literature review team and the statistical analysis team. Early on in the challenge, the statistics team conducted exploratory analyses on the data set and presented their findings to the group. We discussed the results and generated new, related questions that would help expand on the results, incorporating the findings of the literature review team. This back-and-forth process was repeated three or four times before we ultimately converged on our final question. Throughout the process, we were looking for results that could be condensed into one-line clinical summaries -- something that a practicing physician could read and immediately take into consideration during clinical decision-making.

With the competition rapidly coming to a close, we ended up finding two significant results related to the CKD subpopulation, and decided that there was a clinically impactful story about this population that was not fully told in the original study. From there, we started focusing our subsequent analyses on CKD in order to fully explore this story.

How do you hope your results will inform the medical community?

First, we hope that our results will lead to more cautious use of intensive blood pressure management in patients with CKD, as well as contribute to a consensus in the medical community on appropriate blood pressure management in the CKD population.

Additionally, we hope that the massive amount of results that have this contest has generated will demonstrate the power of clinical trial data sharing -- not only to verify results, but also to drive new discoveries and meta-analyses. We think many individuals have the time and ability to find novel results within existing datasets that would positively affect clinical practice, but for a multitude of reasons they do not have access to the data. Of course, the existing incentive structures are complicated, but that does not mean that we should back down from the challenge of making data access more open.

What would you say to medical school colleagues aspiring to do research?

One of our biggest difficulties was that our team was composed entirely of students. Without a faculty mentor, we faced problems accessing the dataset, delaying us from even starting the analysis.

Our advice to students is to have confidence in your abilities and not shy away from situations where it may seem that you are underqualified. In addition, any skill can be acquired if you invest the time. Do not ever feel like you cannot learn something just because you do not occupy the right place in the institutional hierarchy or do not have someone explicitly teaching you. For instance, there are hundreds of online courses, tutorials, and public datasets to use to help you hone your data analysis abilities.

Although we encountered some hurdles because of our relative youth and inexperience, we feel that embracing our identity as students was crucial to our success. The research process can be arduous, but as a team of friends, we always looked forward to spending a Friday night exploring the dataset. We knew each other well, so we understood each team member's strengths and could communicate effectively. As a result, the process was not only fluid, but also enjoyable. We want to emphasize just how powerful and meaningful friendships can be when working together to accomplish a shared goal.

Interested in knowing more about Data Sharing and the other Challenge winners? Join us for the live webcast of the summit — register here for the event which will begin on April 3.

View a video interview with Team Renality Check »