Blog: October 2, 2018
AI in Healthcare Marketing: A New Data Frontier
By Sara Aby, PhD, Senior Data Scientist, Crossix
Over the past few years, artificial intelligence (AI), fueled by machine learning, has taken the healthcare industry by storm. By “teaching” computer algorithms to learn from data, a door has opened for new technologies to emerge that can act and adapt without human intervention.
Current Potential of AI and Machine Learning in Healthcare Marketing
Promising clinical applications of AI and machine learning in healthcare have made global headlines, including for wellness initiatives, personalized treatment plans, diagnostics, clinical research, and risk management. Behind the scenes, data scientists are also tackling operational issues like how to optimize pharma budgets and, ultimately, improve patient care.
Using machine learning is one way marketers are creating efficiencies and optimizing their budgets. Pharma brands are moving away from traditional marketing methods. Instead, they are partnering with technology companies, like Crossix, to use data and machine learning to enhance their promotional strategies. Equipped with detailed, person-level insights, these companies are able to create consumer segments that precisely identify more of the brand’s target audience. Now, marketers can spend their media dollars more efficiently and achieve a greater return on their advertising investment by engaging a smaller population of high-value patients.
At Crossix, we use machine learning across hundreds of millions of data points to generate audience segments in a completely unique and privacy-safe way. To illustrate how this works, let’s look at the process we used to help a diabetes brand reach a more targeted audience.
- Our technology was developed with privacy at its core, which allowed us to combine anonymized person-level healthcare data (like medical claims) with thousands of non-healthcare data points. No other company has the capability to connect this volume of data at scale, while still preserving privacy and HIPAA requirements.
- We applied machine learning techniques that sort through data sets to determine which non-healthcare attributes are most predictive of a diabetic and calculate weight for each of the attributes.
- Our machine learning technology developed hundreds of models on selected attributes and measured the performance of each model. It then identified the model that showed the most promising performance for implementation.
- Once this was determined, each person was scored and then ranked by how likely they are to be a diabetic. The entire U.S. population was then sorted by rank into 10 deciles (each one contains 24 million people).
- These deciles were then overlaid with media audiences (e.g. online device pools, cable subscriber lists, etc.) for more precise targeting.
After applying Crossix segments, the diabetes brand saw significant lift compared to contextual targeting across key campaign metrics. In this case, the diabetes brand was able to reach 3.5 million more target consumers at a fraction of the cost. Additionally, Crossix segments drove significantly more specialist visits and new patient starts than contextual targeting at a lower rate.
This process can be applied across a wide range of conditions, allowing marketers to reach the right target audiences across different media channels.
Marketing campaigns are getting even more unified with the help of AI and machine learning. Crossix recently developed a new generation of audience targeting technology—Crossix Audience Fusion™. This solution goes beyond traditional data to create hyper-targeted campaigns by fully integrating DTC and healthcare professional (HCP) strategies. Marketers can now target highly qualified consumers who are the patients of target HCPs even more precisely. Recent results have shown that Crossix Audience Fusion is 2.5x more precise than other high performing tactics.
It’s clear that the technological advancements generated through AI have transformed the healthcare marketing industry. Without these capabilities, marketers would still be restricted to less precise targeting tactics, ultimately resulting in media waste. With the AI door open, I’m excited to see the ways the industry will continue to innovate.