Saptarshi Purkayastha, Ph.D. and Robert Quick from Indiana University talk about the digital divide and AI in a post-Covid context

In our aim to give our community an approach to the changemakers supported by Computing for Humanity, today we will provide the insights of two counterparts from Indiana University. They are  Saptarshi Purkayastha, PhD, Director of Health Informatics and Associate Professor of Data Science, Health Informatics from IUPUI (Indiana University and Purdue University), and Robert Quick, Director of the Cyberinfrastructure Integration Research Center at Indiana University.

Our first question is related to the moment when they decide to become researchers. Our conversation has been condensed and edited for clarity.

The moment that they decide to become researchers.

Robert Quick recalls his time as an undergraduate student and thinks that he became a researcher over time. Right after graduating in Physics, he was invited to work on computing issues surrounding CERN, the Large Hadron Collider and the Higgs Boson. The term CERN is derived from the acronym for the French "Conseil Européen pour la Recherche Nucléaire", or European Council for Nuclear Research, a provisional body founded in 1952 with the mandate of establishing a world-class fundamental physics research organization in Europe. The Large Hadron Collider (LHC) is the world’s largest and most powerful particle accelerator and the Higgs Boson is the fundamental particle associated with the Higgs field, a field that gives mass to other fundamental particles such as electrons and quarks. His involvement in this project was between the mid 2000s leading up to 2010. After that experience, he was involved with the Open Science Grid Project, and within this project is where he met Computing for Humanity’s founder and our mission. This chain of events makes him think that his path as a researcher was not a moment itself, but the decisions taken through his career experience.

On the other side, Saptarshi was a software architect working mainly in logistics and development for a large company. He was asked to install an electronic health record system in a clinic as part of his work. He did research about open-source software called OpenMRS which is a collaborative open-source project to develop software to support the delivery of health care in developing countries. And this was the moment when his mission found him, “I tried to implement tools and technology for the Government of India in the Ministry of Health, where one of the researchers from the University of Oslo met me and asked me if I wanted to get into academia and do a PhD. I then did research across the world on different sites such as in Tanzania, Malawi, Bangladesh, Nepal, and Bhutan. I also did some projects with the World Health Organization. Through this process of trying to build tools and technologies that would help low and middle-income countries, I became a researcher.”

The most significant challenge of being a researcher

Time is the most valuable treasure according to Rob Quick, finding the time to do everything. Professionally, as Director of the Cyber Infrastructure Integration Research Center, pursuing a Ph.D. and spending time with his wife, Rob manages to do it all.

For Saptarshi, the significant challenge has been twofold. One is finding funding to train or recruit students but also to be able to find more students with similar goals.


How Computing for Humanity crossed and impacted their lives

According to Rob, approximately two and a half years ago, they installed hardware provided by Computing for Humanity. However, he knew about the charity time before through the founder, Roy Chartier, in an event organized by the Open Science Grid. The pandemic delayed the installation, but this was successfully done, and Saptarshi was surprised that such a community of infrastructure existed and made available for their use.

Saptarshi added, “I think 60 plus students have created smaller VMs virtual machines on this cluster and they've been running the open-source EMR on it. They've been playing around with it and in that process, they install and learn how to manage the open-source electronic health record. Those electronic health record systems are primarily built for low and middle income countries and for places where affording the multi $1,000,000 electronic health record systems are not possible. That's why we've been able to train students on this and then on a couple of research projects we've been able to use the resource in the form of generating publications and a couple of dissertations done using this resource from master students. One of the papers is on building machine learning-based diagnosis system. We have some AI tools that look at all the patient records in the EMR and they're not able to make recommendations for appropriate diagnosis and treatment. That paper is still under review”.


To the cluster was added a DHIS 2, which is an open-source electronic health information system that can identify where resources should be allocated.

This data analysis led to an important finding, that Indiana is one of the most critical states in the United States in terms of outcomes from diabetic foot ulcers. According to the researcher, "What it means is when someone has diabetes, they have ulcers developed in different parts of the feet. When these are untreated or hard to heal because of diabetes, amputations are very common and the foot must be cut and about 8% of the patients who have diabetic foot ulcers, get cut”, Saptarshi concludes. He emphasizes that they have built a data warehouse on the donated infrastructure that is able to monitor patient records and genetic studies from all over Indiana. They are able to analyze risk factors and look at which patients are at high risk and nurses and midwives can target them sooner and ask them to come to the clinic for preventative measures and other treatments.


That is one of the projects where this infrastructure has been really useful and we are trying to apply for National Institute of Health funding to be able to support this in the long term. A lot of this experimental early work is possible because of the donated infrastructure from Computing for Humanity.


Long-term career goals

It's important for us to understand the motivations and goals of the researchers behind the projects we support. Our questions are like Tiny Desk moments, where we not only focus on the research itself, but also on gaining insights into what drives these researchers to pursue their work. During a conversation about their future career goals, Rob shared that he is currently pursuing his PhD and expects to complete it within the next two years. Retirement is another goal but general consulting and finding research opportunities are additional ideas he is passionate about.


On the other hand, Saptarshi thinks about training students who currently come with clinical backgrounds. He sees technology as an ally that allows practitioners to take better care of patients and find treatments and advancements that can ease disease. However, another challenge faced with technology is the difference between working in a lab setting and in an online environment. “It's very hard for people to evaluate how people are doing because you can't observe them live right in practice.  We have built some tools that can identify where and how the patients and the clinicians are accessing the patient information. This online setting is crucial because more people need to work remotely, and it is essential to identify and engage students in remote learning”, Saptarshi says.

Huge digital divide

As we close our interview, the topic of the accessibility of technology is broached.  Saptarshi comments that the digital divide or high-performance computing divide are actually much larger and significantly problematic for lower middle-income countries and low resource contexts.  He points to “A low resource context means slums or places where there are fewer resources or human resources trained even though it might be high income, but there are fewer people trained to be able to use technology and HPC to take care of patients in healthcare. That's what I mean by divide because of AI, you'd need a lot of computing to be able to do AI properly, and that I think is lacking in many places. The resources that Computing for Humanity provides may be helpful to bridge the divide”.

On the other hand, Rob Quick says, “making resources available is one thing and then making the resources available to bring the researchers in LMIC's (low or middle income countries) to a place where they can utilize those resources is another concern that is more of a training and workforce development”.

Some insights after COVID and AI

Rob is concerned about the balance of having technology for all interactions in a post COVID era. He says “A technology-heavy environment still has risks of dehumanizing in many ways. However, technology will never replace the human research mind. It might be able to process things infinitely faster and find correlations that may not come to us right away, but the deductive logic and reasoning behind any hypothesis being tested certainly come from the human mind”.


When this concern comes to the research field, Rob expresses that interoperability and the reusability of AI is something that is a major issue. He gives an example of someone who has a GitHub repository where the code lives that is developed in an environment that you then have to replicate or there is something where you have to pay a fee to get the environment they have created to actually use that algorithm. “I think that's a space where many more researchers can adopt things because a lot of them won't have the time, effort, or skill to develop their own AI algorithm to do their work”.

As we wind down and prepare to finish the interview, Saptarshi expresses his view, “AI truly is already very good in terms of quality, particularly in a context where there's a lack of specialists, in technology, be it in diagnostic medicine and on so forth. Also, the potential of it to be able to scale to the community health worker who is not very well trained but needs to deal with some very complex problems, so it'll be very useful for them. I think as a society we need to think of what happens to the jobs that get replaced and should be replaced for productivity purposes”.

Previous
Previous

Ari Ginsparg: Moving one step closer to making a difference in the lives of those affected by conditions like narcolepsy

Next
Next

James Green: Research is being challenged by access to High-Performance Computing, data sets, and new technologies