Influence Health provides the healthcare industry’s only integrated digital consumer engagement and activation platform. The Influence Health platform enables providers, employers, and payers to positively influence consumer decision making and health behaviors well beyond the physical care setting through personalized and interactive multichannel engagement. Since 1996, the Birmingham, Alabama-based company has helped more than 1,100 provider organizations influence consumers in a way that is transformative to financial and quality outcomes. Healthcare is a personal business. Each patient’s needs are different and require an individual response. On the other hand—as the cost of providing healthcare services continues to rise—hospitals and health systems increasingly need to harness economies of scale by catering to larger and larger populations. The challenge then becomes providing a personalized approach while operating on a large scale. Influence Health specializes in helping its healthcare sector clients solve this challenge by getting to know their existing and potential patients better and targeting each individual with the right health services at the right time. Advanced predictive analytics technology from IBM allows Influence Health to help its clients discover the factors that have the most influence on patients’ healthcare decisions. By assessing the propensity of hundreds of millions of prospects to require specific healthcare services, Influence Health is able to boost revenues and response rates for healthcare campaigns, improving outcomes for its clients and their patients alike. Targeting the Savvy Consumer Today’s healthcare industry is becoming more competitive than ever before. If the utilization of an organization’s services drops, so do its profits. Rather than simply seeking out the nearest hospital or clinic, consumers are now more likely to make positive choices between healthcare providers. Paralleling efforts common in other industries, healthcare organizations must make more effort to market themselves effectively to both existing and potential patients, building long-term engagement and loyalty. The keys to successful healthcare marketing are timeliness and relevance. If you can predict what kind of health services an individual prospect might need, you can engage and influence them much more effectively for wellness care. Venky Ravirala, Chief Analytics Officer at Influence Health, explains, “Healthcare organizations risk losing people’s attention if they bombard them with irrelevant messaging. We help our clients avoid this risk by using analytics to segment their existing and potential prospects and market to them in a much more personal and relevant way.” Faster and More Flexible Analytics As its client base has expanded, the total volume of data in Influence Health’s analytics systems has grown to include over 195 million patient records, with a detailed disease encounter history for several million patients. Ravirala comments: “With so much data to analyze, our existing method of scoring data was becoming too complex and time-consuming. We wanted to be able to extract insights at greater speed and accuracy.” By leveraging predictive analytics software from IBM, Influence Health is now able to develop models that calculate how likely each patient is to require particular services and express this likelihood as a percentage score. Micro-segmentation and numerous disease-specific models draw on demographic, socioeconomic, geographical, behavioral, disease history, and census data and examine different aspects of each patient’s predicted healthcare needs. “The IBM solution allows us to combine all these models using an ensemble technique, which helps to overcome the limitations of individual models and provide more accurate results,” comments Ravirala. “It gives us the flexibility to apply multiple techniques to solve a problem and arrive at the best solution. It also automates much of the analytics process, enabling us to respond to clients’ requests faster than before, and often give them a much deeper level of insight into their patient population.” For example, Influence Health decided to find out how disease prevalence and risk vary between different cohorts within the general population. By using very sophisticated cluster analysis techniques, the team was able to discover new comorbidity patterns that improve risk predictability for over 100 common diseases by up to 800%. This helps to reliably differentiate between high-risk and very high-risk patients—making it easier to target campaigns at the patients and prospects who need them most. With insights like these in hand, Influence Health is able to use its healthcare marketing expertise to advise its clients on how best to allocate marketing resources. “Our clients make significant budgeting decisions based on the guidance we give them,” states Ravirala. “We help them maximize the impact of one-off campaigns—such as health insurance marketplace campaigns when Obamacare began—as well as their long-term strategic plans and ongoing marketing communications.” Reaching the Right Audience By enabling its clients to target their marketing activities more effectively, Influence Health is helping to drive greater revenue and enhance population health. “Working with us, clients have been able to achieve return on investment of up to 12 to 1 through better targeted marketing,” elaborates Ravirala. “And it’s not just about revenues: by ensuring that vital healthcare information gets sent to the people who need it, we are helping our clients improve general health levels in the communities they serve.” Influence Health continues to refine its modeling techniques, gaining an ever-deeper understanding of the critical attributes that influence healthcare decisions. With a flexible analytics toolset at its fingertips, the company is well equipped to keep improving its service to clients. Ravirala concludes, “In the future, we want to take our understanding of patient and prospect data to the next level, identifying patterns in behavior and incorporating analysis with machine learning libraries. IBM SPSS has already given us the ability to apply and combine multiple models without writing a single line of code. We’re eager to further leverage this IBM solution as we expand our healthcare analytics to support clinical outcomes and population health management services.” “We are achieving analytics on an unprecedented scale. Today, we can analyze 195 million records with 35 different models in less than two days—a task which was simply not possible for us in the past,” says Venky Ravirala, Chief Analytics Officer at Influence Health.
Questions for Discussion
1. What did Influence Health do?
2. What were the challenges, the proposed solutions, and the obtained results?
3. How can data mining help companies in the healthcare industry (in ways other than the ones mentioned in this case)?