Suman De, MD, MHA, had a dream of carrying on his physician father’s legacy to serve the community and become a doctor to help make an impact in the world. That dream came true, but it also led him in an unexpected and non-traditional direction — to NTT DATA Services — and making an even greater impact than he ever imagined.
Today, Dr. Suman is a Principal Solution Consultant with our Healthcare Data and Intelligence practice. He consults with clients looking to create an enterprise data and analytics strategy or design innovative solutions and execute advanced data science projects.
I recently talked with him about his journey from doctor to what he calls a “citizen-data scientist” and how data science technologies, like artificial intelligence, machine learning and deep learning, are bringing incredible change to healthcare analytics.
Q. Growing up in India, you dreamed of becoming a pediatrician. So how did you end up working with healthcare data instead of children?
A. I always had a passion for practicing community medicine and child health. Early in my career, I worked for the World Health Organization (WHO) in the National Polio Surveillance Project, a program to support India’s polio eradication efforts.
I was involved in early detection, diagnosis, investigation and immunization of children with paralytic polio in rural areas of West Bengal, India. I used to collect and analyze data to identify areas and child populations at risk. I saw firsthand how data-driven knowledge could make work more efficient and actionable. And that could translate into tailored strategies to proactively identify, immunize children with low vaccination coverage in the areas with high polio prevalence.
With this understanding and experience, I decided to change my career path to find ways to make healthcare data work in the best interest of patients and community health. I thought I could help more kids this way than if I treated them one at a time because no matter how good the training and clinical infrastructure is, to be the most effective, with your clinical outcomes, you need the data and the insights from the data.
Q. As part of the NTT DATA Services Healthcare team, you help health plans and healthcare providers find insights in their data that can make a difference in patient/members’ lives. Tell me about how you used advanced data science techniques like machine learning in a recent analytics project with them to make a difference?
A. Our client was seeking ways to proactively identify members who are high cost, high risk and would require immediate care management interventions to stay healthy and stay away from a future ER/IP visit. So, they asked us to be part of a population health analytics challenger project, where we competed with their in-house analytics team and other vendors to create predictive models to identify these risky members.In the challenge, we moved away from the typical analytics solution to take an innovative approach in building the predictive models using machine learning. We used an automated machine learning tool that could use both structured and unstructured data for building predictive models at a faster speed.
A typical approach would have been for our data scientists to manually program the predictive models using one or two algorithms, which could have taken weeks or months. With the machine learning technique, we created predictive models using 45 different statistical algorithms and auto-picked the one that performed the best. In total, five different predictive models were created in just four business days.
We identified the top 10% of high-risk members by cost and service utilizations and flagged 180 potential case management candidates. The “WOW” factors for the client were:
- Speedy development of the predictive models (as other vendors couldn’t match our delivery timeline, even the client’s data scientists took nearly six weeks to create just one predictive model) and,
- Revealing the Top 10 Data Attributes — reason factors that influenced their member risk.
Also, the enrichment of our predictive models using external data, i.e., the social determinants of health risk scores (exercise and diet risk, addiction risk, transportation risk, support system risk, care engagement risk, etc.) was icing on the cake. It helped their care/case managers get a more accurate understanding of the member socio-behavioral, attitudinal aspects to frame tailored intervention strategies, which was not possible by just predicting on their existing internal systems data.
This was a true game-changing project. It’s a whole new approach to data analytics that helps lower the cost by vastly reducing the go-to-market time and increasing the speed to value. Our solution was so innovative that clinicians, like me, can now become what I call a “citizen-data scientist” and get started on building automated predictive models without specialized data science training.
For me, it was the gateway into providing more game-changing efforts for our clients. Today, I’m harnessing the power of advanced data science technologies like AI, Machine Learning and more.
Blending these technologies with my clinical knowledge, I’m now creating innovative solutions that can truly transform today’s “Sick Care” to real “Health Care.” I can use data and technology to impact even more patients and realize the World Health Organization’s vision of true Health for All.
Watch Suman’s Game Changers video here.
Get to know NTT DATA
Accelerate your digital healthcare journey. Learn more about NTT DATA’s solutions for Healthcare Providers.
Post Date: 19/06/2019