Ask the Expert: Reviewing Proposals

Over the past four years, the Pittsburgh Health Data Alliance has received over 200 proposals for funding. UPMC Enterprises’ Zariel Johnson, PhD, takes us behind the scenes to explain where these proposals come from, how the Alliance reviews them, and trends that she’s seen.

We recently spoke with CMU’s Center for Machine Learning Director Joe Marks, PhD about what prospective entrepreneurs should take into consideration in order to have the best chances at securing funding from an investor. He shared best practices about knowing your goals, doing your research, and networking. But what do things look like on the other side?

What research proposal areas do the University of Pittsburgh’s Center for Commercial Applications of Healthcare Data (CCA) and Carnegie Mellon University’s Center for Machine Learning and Health (CMLH) see most often?

Both centers see proposals that touch on all areas of healthcare such as precision medicine fueled by healthcare quality improvement, -omics approaches, patient engagement and experience, and efficiencies and infrastructure. Due to the relationship between Pitt and UPMC, many of the teams submitting proposals to the CCA include teams working closely with clinicians and clinician-researchers. An example is the CCA project Coronary Artery Disease Intelligent Detection via Metabolic Expression (CADidME). Pitt Department of Biomedical Informatics Associate Professor Dr. Vanathi Gopalakrishnan, teamed up with Dr. Steven Reis, the founding director of Pitt’s Clinical and Translational Science Institute and a distinguished UPMC cardiologist. In their PHDA project, Dr. Gopalakrishnan is applying advanced algorithms to a rich dataset collected by Dr. Reis in an effort to identify patients at high risk for adverse cardiac events and optimize patient care.

 

What about AI and machine learning?

CMU is a world leader in machine learning, artificial intelligence, and computer science, and many of the proposals and active PHDA projects from the CMLH reflect that strength. We see researchers who are able to apply and adapt statistical and machine learning methods to overcome the unique challenges posed by healthcare data. Dr. Daniel Nagin, Professor in CMU’s Heinz College, is working with Dr. Jonathan Elmer, a Pitt researcher and UPMC Emergency Medicine physician, on their project to predict patient outcomes following cardiac arrest. Drs. Nagin and Elmer have had to process large, complex datasets pertaining to neurological activity in order to apply a statistical technique that is well-suited to support clinical prognostication.

 

What is the proposal process for applicants?

The majority of PHDA projects are solicited through request for proposal (RFP) cycles that are administered by each of the two PHDA centers: the Center for Machine Learning and Health (CMLH) at CMU and the Center for Commercial Applications of Healthcare Data (CCA) at Pitt. The proposal processes and frequency of RFPs differ between the two centers, but both begin with the opportunity for potential applicants to receive feedback and guidance as they write a full proposal. During this period, investigators are encouraged to work with the Centers to clarify their project scope, identify commercial potential, and seek additional collaborators if needed. Applicants then submit a full proposal that outlines the healthcare problem that their project will address, what they plan to accomplish with support from the PHDA, how their project could lead to a commercial product or service, and a requested budget. The relevant PHDA center and UPMC Enterprises teams then review proposals and select a set of finalist teams who are invited to deliver pitch presentations to an audience of stakeholders at UPMC Enterprises. As finalist teams prepare their pitch presentations and slides, they are again provided with feedback from the Centers and UPMC Enterprises teams.

 

What happens after the presentations?

Teams that are selected for PHDA support will finalize their project plans based on input gathered throughout the evaluation process before they begin work. Those teams that are not selected are provided with written feedback to help the team understand the rationale for the decision and potentially improve their proposal for another PHDA RFP cycle or for other funding mechanisms. We understand that investigators have limited time, and we aim to make this process valuable whether they are ultimately selected for support or not.

 

What do you expect to see in the next few years?

We have seen an increase in the number of proposals that we receive and their breadth across topics in healthcare as researchers have identified it as a challenging and rewarding area to apply their work. I hope to see continued and growing engagement with schools and departments across both universities and attraction of new perspectives and methods that are not routinely applied to healthcare. The PHDA is seeing more and more projects that bring together all three partners: CMU, Pitt, and UPMC, and I expect that more teams will adopt this model in order to do scientifically rigorous work that can lead to real benefits in the healthcare system.

 

Thanks, Zariel.

To apply for Pittsburgh Health Data Alliance funding, visit our partner pages here:

• Pitt’s Center for Commercial Applications of Healthcare Data
• CMU’s Center for Machine Learning and Health