Learnings from Cypher 2019 from Sanjay Parikh, EVP & Co-Founder of Indigene - by Kay S

“You Don’t Have To Be Great To Start, But You Have To Start To Be Great.” –Zig Ziglar

This seems to be something one should remember before we embark on learning something we do not know.  When we sometimes find the process of learning long and tedious, the most important thing to remember is that time will pass no matter what, so might as well spend it learning something useful and upgrading your knowledge.  These are thoughts that went through my overwhelmed brain as I entered the Cypher Conference 2019 very recently during my stint at REVA University.  There were a few sessions I was able to catch up on when I was there and I would like to summarize my experience in the coming days as I felt that all learnings should be shared be it small or large as you never know who might find it relevant to their lives.  I would like to serve this as an attempt to document the learning I had regarding the upcoming trends and breakthroughs in the world of IoT, Artificial Intelligence, Machine Learning, and Analytics.

Some of the most important points that I gathered from the sessions that I did attend was:

1.     In order to make any AI/ML successful, one should always pay attention to the business needs and then ensure that the end customer needs are accounted for.

2.    One of the differentiating factors of this learning was that to think that AI/ML will replace human beings.  This could never be more farther than the truth.  The truth of the fact is that AI/ML are tools or programs that have been invented by human beings and hence they will always be in control of these.  One of the most apt statements by renowned Neurosurgeon, Dr. Ajay Bakshi did indeed drive the seriousness of responsibility and accountability in AI.  He stated, “Do not make artificial intelligence so intelligent that it starts controlling Cognitive intelligence by thinking for itself which can be disastrous.”  AI can only become a tool never a replacement to a human being.

One of the first sessions, I attended is summarized here below and currently, as I have joined IPL, an organization geared towards building product leaders, this seemed all the more relevant to the cohorts who are pursuing their career in product leadership.

Sanjay Parikh, EVP & Co-Founder of Indigene on product commercialization in Life Sciences in the age of Big Data

Mr. Parikh shed light on the importance of data and the path taken for production commercialization with the amount of huge data that is available today.  The emergence of data he stated starts with Research & development, regulatory affairs, Market access, Medical affairs, and lastly sales and marketing.  The various forms of data that exist in the above areas are biological data, clinical data, regulatory data, medical data, real world data, commercial data, payor data, social media, and data from wearables.  The volume and velocity of data amounts to 463 EB/Day from IoT and connected devices and 1100+ TB per lifetime from Genetic/Clinical and exogenous factors which come in both structured and unstructured form as per IDC data.  He spoke about advanced analytics in that the potential business impact in Lifesciences. As per the Mckinsey report, the potential EBITDA impact of advanced analytics on these departments are as below:

Research and early development -               12 - 25%

Development, regulatory & Safety -               13 - 20%                                              

Manufacturing and Supply Chain -                  4 - 8%

Market access, Commercial, and medical -  15 – 23%

However, he stated most pharma executives are unsatisfied with the analytics maturity of their companies; 54% believe that their current data and analytics capability is not meeting the needs of the organization, and 60% cite lack of access to the right data as the biggest barrier to increasing use of real world evidence.  He explained the traditional approach of analytics acceleration and automation versus Next Gen approach as below:

Analytics in the Life Sciences – A wide Spectrum of Opportunities

DISEASE INSIGHTS – UNLOCKING VALUE FROM REAL WORLD DATA

• Timely diagnosis of some potentially serious conditions is critical but frequently proves difficult as the symptoms are rarely indicative of the disease
• Early diagnostic determinants of the disease are key to improving patient’s prognoses.

He then spoke on customer segmentation using AI for Insights on Analytics ready data sets for exploratory data analysis with variable selection and data summary.  After this, he moved on to target variable creation for Brand Adoption and Patient Adherence.  Then comes Predication Score Approach comprising of training set, Mt. Algorithm selection and validation set which in turn trains the model by model diagnosis and performance evaluation.  Insights and recommendations will be given as below:

Confusion Matrix	Predicted Non-adherence	Predicted Adherence
Actual Non-adherence	78%	22%
Actual Adherence	31%	69%

He ended the session by insights on Customer segmentation using predictive scoring. He demonstrated the journey depicting adoption behaviour of the HCP through decision tree.

Hope you enjoyed this post and get some insights from it.  Something more from my desk from a different speaker tomorrow.  So long folks.

- By Kay S

**All views are of the speaker mentioned above.