WOW!
What a GREAT conference the Pharmaceutical Marketing Science Association (PMSA) put on for 3 ½ days at the end of April in San Antonio, TX.
WHY?
Almost 300 pharmaceutical commercial analytics professionals discussed the “next generation” of analytics. Here are some key takeaways on that hot topic presented at the conference. PMSA generously shares all of the presentations on their conference website.
Telling the Data Story through Machine Learning
Machine Learning (ML) and Artificial Intelligence (AI) are proving to be the way pharma analytics is utilizing BIG DATA. ML and AI help make sense of the data that comes from all the different data sources and helps analysts tell THE story. Ashish Sharma and Kaiwen Zheng from Axtria gave a presentation on, “From Unstructured to Structured Data: Can Machine Learning Help Make Sense of EMR/HER Data?”
Analytics Drives Multi-Channel Dashboards
As Paul Rabideau from KMK Consulting and Julia Brodsky from Novartis discussed in their presentation on “The Strategic Use of Analytics for Business Impact: Multi-Channel Dashboard,” basic marketing mix models are paving the way for determining which promotion is selling your drug.
Artificial Intelligence (AI) Giving Pharma the Competitive Edge
In their talk on “AI Can Give Pharma Companies a Competitive Edge: Here’s How!” Dharmendra Sahay and Arun Shastri from ZS Associates give examples of how AI is being used in commercial pharma analytics today and how to get started on utilizing this game changing technique in your firm.
Machine Learning Identifies HCPs in Rare Disease
“Winning With Analytics When The Chips Are Stacked Against You: A Novel High-Dimensional Hybrid Machine Learning Approach Identifying High-Value Rare Disease Specialists“, presented by Jack Lin of Actelion and Rick Rosenthal and Tim Hare of Symphony Health, shared the effort they made identifying 8000 variables spanning diagnostic, prescribing and procedure data silos. This group then summarized the variables by physician and modeled by ensembles of learning agents robust to the high-dimensional data. Their effort resulted in the identification of thousands of new patients providing the opportunity to efficiently expand promotion efforts.
Machine Learning Helps HCPs Target Undiagnosed Patients
Orla Doyle and Steven Laux from IQVIA presented “Optimizing Physician Targeting to Find Undiagnosed Patients: An Application of Advanced Machine Learning Methods to Hepititis C.” With the new HCV drugs on the market today the under-diagnosed population for this infection can be found through early detection algorithms therby leading to substantial increases in the addressable market.
Machine Learning Provides Potential Patients for Physicians
In the last session of the conference, a team from IQVIA presented “Innovative Machine Learning Methods to Enhance Accuracy and Effectiveness of Physician Alert.” IQVIA used traditional logistic regression for their base line model but they found that recent breakthroughs in deep neural networks demonstrated superior model performance in cases of large volume, high dimensionality, missing records and complex data structure.
