An active population health management program is key to improving performance, and it needs big data engagement analytics if it’s going to succeed.
Typically, plans collect data from multiple different sources, including member and provider demographics, enrollment/eligibility information, medical claims, pharmacy claims and HEDIS results.
Compound these sources with consumer/marketing, census, disease/case management, lab results, risk adjustment or even social media data and you begin to see just how diverse and complex the data sources are.
Big data analytics offers the best way to integrate and make sense of all this information.
How do big data analytics work?
Big data is defined as large or complex data sets that cannot be managed with traditional data processing and applications.
Challenges include analysis, capture, data curation, search, sharing, storage, transfer, visualization, and information privacy. Big data also includes the predictive analytics that help extract meaning from the data—rarely to a particular data set. The goal is to learn from the data so that we can make better decisions.
How do big data analytics help identify the highest risk patients?
Using big data for engagement analytics, plans can:
• Create member- and measure-specific predictions that help identify members who are at highest risk for undesirable outcomes;
• Segment members according to their barriers to engagement;
• Discover the most efficient intervention channel to change behavior; and
• Determine which members to target, what they should be targeted for, as well as the appropriate message and communication channel.
By building real-world member risk profiles, big data becomes the bedrock to building an effective population health outreach plan. Once plans understand each member’s barriers to engagement, plans can predict which members have the highest risk of undesirable behavior, then customize an approach for each at-risk member based on greatest area of need.
As a result, plans can focus interventions on the right members using language (and scripts) that are best suited to address those barriers, and ultimately change behavior.
How about members with messy profiles?
A significant number of members have “messy” big data risk profiles. They could be at high risk (i.e., exhibit undesirable behavior) for all (or most) measures, or they could be at high risk for some measures, moderate risk for other measures, and lower risk for the balance of measures.
If your overall goal is improving and sustaining engagement, the focus should be on the measure or groups of measures that are of greatest clinical importance.
If a member is high risk for multiple measures, including the risk of readmission, then the focus of the intervention should be engaging the member with the appropriate interventions so that both the admission and readmission are avoided.
With an engagement approach, plans may be able to not only alter the course of members' future admission and readmission behavior, but also affect their future compliance, future adherence and future satisfaction with the plan.
Saeed Aminzadeh is chief executive officer of Decision Point Healthcare Solutions, a healthcare engagement analytics solutions provider.