Working with a national wellness incentives company, we learned that they were facing challenges communicating the value of their programs, and differentiating their approach to wellness. They had an immense amount of data collected across multiple programs, and due to the complexity and interrelated components of the programs, the outcomes were difficult to present. Reams of data were summarized in the tabular fashion, making it difficult for their clients to easily see trends or spot opportunities. This reporting not only limited their ability to tell an effective story to their clients but also hampered ongoing program effectiveness and metric tracking.
We examined their data to uncover trends and patterns and discovered that certain metrics, had more frequent movement while other healthcare metrics were slow to change. By decoupling these from each other, we were able to recommend a layered approach to demonstrating value by removing short-term focus on metrics that would typically take longer to move. In addition, we recommended limiting reporting on fewer key metrics that could be represented in a manner that was more visually compelling yet still effective at telling the important story that the wellness programs were effecting positive results.
We applied modern visualization approaches to creating an output that had more movement and was easy to interpret. Being able to show movement along a few critical factors that could be regularly measured and reported provided a more dynamic interface and an opportunity to meet with their clients more regularly. For example, the graph below shows members who were active in a walking program. The design makes it easy to see the green wave that demonstrates users moving into higher activity categories after two quarters as measured by the number of steps taken by participating users.
Understanding what metrics to demonstrate and showing them in a compelling manner achieved the goal of getting this client more connected with its constituents. It removed the onus of establishing a purely financial ROI metric that was difficult to quantify and replaced it with a set of measure that could change and program effectiveness. Our goal was to ensure that we focus on a few key metrics and show them the movement or change through our visual graphs and charts. We moved away from their tabular approach to a much cleaner and easier representation of their data. For example, the shorter-term outcomes of a nutrition program can be seen below:
A lot more employees are eating healthy
While the longer-term benefits such a lowered cardiovascular risk, can be shown in a more static “heat-map” that nicely shows the distribution of risk scores, which don’t change that frequently over time.
A corporate wellness company had serious challenge with their engagement levels and were not able to identify reasons for members disengaging and how they could proactively improve engagement in their programs.
zakipoint Health incorporated all their program participation data, including challenges, campaigns, message content across different customers, to build a disengagement model and run analytics to understand key drivers of disengagement. By bringing all the data in one place, we were able to show analysis of typical drop-offs across different segments. We built a prospective disengagement model using machine learning science to proactively score members that were at risk of disengaging and reasons behind it.
The following outcomes were achieved after running the analytics using our z5 Platform:
- Clearer definition of engagement metric, engagement indicators, framework for analyzing engagement (i.e. looking at the length of the user in the system as opposed to pure snapshot which includes new and old members)
- A robust defection risk scoring model. This core capability allowed our client to continuously score the member base and identify high risk members to profile, analyze and work on as opposed to focusing purely by program or demographic group or on the entire base
- A better understanding of the optimal challenge design
- Better ways to monitor health of the business through a subset of engagement indicators with recommendations on right cut offs, for example, the number of pedometer day readings
We have been working with a corporate wellness provider that has a challenge demonstrating its value and return to a particular client they serve. The client could not put all the data in one place. The biometrics data is there, but the client does not know how to use it to bring out the best results. They do not know about risk factors such as program ineffectiveness, declining population health, gaps in preventive care, unmanaged chronic conditions etc. affecting their population. We were called to help create a solution to help this provider go at risk with its client, and continue to show ongoing value by identifying the risk factors in the population that are driving costs.
On our initial data load, we took 2 years worth of claims data and processed it through our analytics platform (z5). Out of approximately 7,000 members in the population, there were 1,315 members that were suffering from chronic conditions, including arthritis, asthma, diabetes, depression, hypertension, CAD, CHF and COPD. Many members exhibited co-morbidities as well (multiple chronic conditions). For example, 169 members manifesting 3 chronic conditions cost the employer approximate $2 million.
In 2015, the employer paid approximately $19 million in healthcare costs, and in 2016 our client introduced a new series of coaching programs for their client.
Using our models, we have stratified the risk of the member population in order to help our client engage with the member population most in need of intervention. With a goal of attaining 10% realization of spend reduction, the predicted savings represent almost a 3:1 return on the coaching investment for this population. Using our platform, our client we can understand the actual disease conditions driving costs, pinpoint which members to target for intervention, and predict savings if risks are mitigated.