The cutting-edge technology developed by Medivo, which was founded in 2010 by Jason and Sundeep Bhan along with Destry Sulkes vouches to use artificial intelligence techniques to synthesize lab results and identify them in real time. The machine learning algorithm created by the Medivo Data Science team combines state-of-the-art statistical modelling deployed on a scalable technology platform that can process millions of test records in seconds.
The idea behind launching Medivo was to deliver on the promise of precision medicine by providing decision support on the use of targeted therapeutics. Medivo’s data analytics services have the potential to impact the health of millions of patients.
The company is growing at an unprecedented 200% per year and over a year ago they became global by opening another location in Bangalore. Since then they have expanded to Payer and Labs verticals within the US (in addition to the Life Science vertical) and are building amazing machine learning based products to help their clients and partners across the board.
In a conversation with Tatiana Sorokina, Head of Data Science at Medivo, we tried to explore the importance of analytics in healthcare space, how is Medivo utilizing it and more, as in the excerpts below:[divider top=”no” size=”1″]
[dropcap size=”2″]AIM[/dropcap] Analytics India Magazine: Tell us about your journey in the analytics industry.
[dropcap size=”2″]TS[/dropcap] Tatiana Sorkina: While working in the educational consulting industry during my graduation from Moscow State University, I observed the process of supporting clients who were pursuing higher education abroad to be very manual and slow. To address this, I gathered the data collected by the agency I worked at and created an algorithm that helped significantly increase the throughput of applications we were processing on a regular basis.
It was during a graduate program specializing in Analytics and Data Science at Columbia Business School in New York City that I got introduced to Medivo, a healthcare tech start-up in New York by a friend and was asked to join it as an intern. Following a confirmation with the company in 2012, I was asked to start my own market research group after which I shifted my focus to secondary market research, or what we now call Data Science. [quote] Realizing the potential of lab data and how valuable it was for Life Science companies, we developed an algorithm in collaboration with one of our clients, that could identify patients diagnosed with a rare form of cancer just based on lab data! It’s gaining popularity assured that we are onto something! Soon the team of analysts expanded who worked with clinical team to uncover new insights in more than 25 therapeutic conditions (we will be in 50+ by end of 2016). [/quote] [divider top=”no” size=”1″]
AIM: How important is analytics in the Healthcare space?
TS: Analytics and data are becoming increasingly important in healthcare, especially in the US. Everyone, from patients to large hospitals and research centres, are sitting on a pile of data that is not being utilizing to its full potential. This is not because they don’t want to or don’t realize its importance, but because they don’t have the resources and expertise to build an infrastructure to collect, process, clean, harmonize, and mine the data.
For years, data aggregators have been collecting data and feeding it back to customers. With the rise of these data aggregators we saw a big wave of analytics consulting companies ready to analyse the data. However, very quickly data became commoditized and too messy and big to buy by itself. That’s why we are witnessing a paradigm shift where it has become vendor’s responsibility to not only collect and process big data in healthcare, but also apply analytics and machine learning that helps draw intelligence. This is where Medivo truly excels.[divider top=”no” size=”1″]
AIM: Would you like to highlight the benefits of using data science at Medivo?
TS: [quote] Using data science at Medivo is a true necessity. Without artificial intelligence algorithms, we cannot clean and harmonize lab data in a time frame that is acceptable to our clients. We cannot analyse massive volumes of unstructured data manually, and we most definitely cannot manually interpret complex reports written by pathologists in a manual fashion. [/quote]
I know that it is somewhat of a trend to use data science in companies now because that’s what the “cool kids” are doing, but at Medivo we use the power of data science not because of the “coolness” factor (although, who wouldn’t want that?) but because it’s what our whole business is built upon.
We hire best in class data scientists not because we want to be known for hiring the best talent, but simply because being a mediocre data scientist is just not good enough for us to keep up with our exponential growth. Best in class data scientists for us are as essential as seasoned sales reps for a pharmaceutical brand.[divider top=”no” size=”1″]
AIM: Can you highlight a specific use case of data science strategy that has brought significant value to Medivo?
TS: I think the biggest win for us was using data science for increasing our operational efficiencies. Don’t get me wrong, we didn’t build an algorithm that made people functions obsolete. On the contrary, we made people jobs so much more important by giving them tools they could use to be more efficient. Think about a doctor who, instead of reading through a handwritten pathology report and guessing whether a pathologist wrote “malignant tumour” or “benign neoplasia”, or typing up correct diagnosis and entering into a database, already has a tool capable of recognizing the right diagnosis and can specify business rules to train a machine learning algorithm that helps predict the diagnosis![divider top=”no” size=”1″]
AIM: Tell us about some important contemporary trends that you see emerging in the present analytics space across the globe.
TS: Because the analytics space itself is a place for learning and innovation, doing business in this space brings along the need to constantly evolve and improve. One of the most obvious trends is the search for simplicity. More and more, end users of data analytics are seeking answers in the most simplistic form possible. Another trend along those lines is speed. Healthcare decision-makers want everything in real-time, and the analytics industry must strive to conduct analyses as quickly and efficiently as possible. Lastly, the industry is also trending towards predictive analytics. Real-time data sometimes isn’t even quick enough, and consumers want to predict outcomes before they happen. This is Medivo’s most recent exploration and something that we are very proud to offer.[divider top=”no” size=”1″]
AIM: What are the most significant challenges you see in the Analytics space?
TS: Lack of talent. I don’t mean talent as an aptitude to be a data scientist, but rather the ability to constantly learn and apply relevant skills to solve problems. Being a Data Scientist is a unique skill in the sense that it is not a trade that you can learn once and then keep using it repeatedly. New technologies emerge every year, making older technologies virtually outdated.
In this constantly changing world of innovation in data science, even walking fast is not enough to keep up with the pace – you need to be able to run. Not everyone is willing to dedicate themselves to constantly learning and striving to become better at what they do. Only a fraction of the people we interview are true scientists, who never give up and push themselves to learn more.