Silverlink is an interesting company whose approach is to apply science to drive healthcare behaviors in a scalable manner. They made some announcements today at the Health 2.0 conference. I recently got a chance to speak with the CEO (Stan Nowak) and his team. Silverlink has 50 healthcare clients representing 150 million lives (including 9 of the top 10 Managed Care Organizations) and has delivered more than 100 million HIPAA-compliant communication “touches”.
It has been estimated that $2 trillion is being spent on healthcare and that up to 70% of it is related to consumer behaviors. Silverlink estimates that a health plan with 15M members can save very large sums for very small changes in behavior. For instance:
- Improving treatment adherence 1% can save $800M
- Moving patients to generic drugs can save $80M per 1% change
- Small improvements in coordination of benefits could save $180M.
Add to this the growing sense among healthcare insurers that brand and retention are becoming more important and the opportunity is pretty clear – plans that can drive improve customer behavior in a way that builds their brand will do well.
In many ways managed care is moving from a B2B industry to a B2C industry. For instance, JD Power did a quality survey for healthcare in 2007 for the first time. They found that while information and communication was #3 on the list of things consumers cared about (coverage and doctors were first and second), healthcare payers ranked dead last behind even cable companies and credit cards! By and large these healthcare plans cannot easily differentiate on price or networks. This means that communication is the product and with 200M US consumers enrolled, insurers send $9B on (mostly defensive ) communication. As more and more of the spending power in healthcare moves to consumers, being bad at customer service will no longer be an option.
The traditional approach to “customer service” in healthcare has relied on nurses as the primary interactors. However, nurses are much more expensive than typical front line staff such as call center representatives and are generally in short supply. This means that nurse-based customer service has focused on those who are sick now and not on those “at risk”. Yet the value to healthcare plans of providing better service to those at risk has been shown repeatedly in lower costs and better care, for instance by keeping someone at risk of diabetes from actually becoming a sufferer.
Automation would seem to be the obvious way to address this but traditional call automation is ineffective in the complex world of healthcare. Silverlink’s insight is that automation can be effective if it delivers personalized communication at scale and with speed.
Silverlink has a series of new services (detailed in their announcement which you can find on their PR page) based on an approach it calls Adaptive Healthcare Communication Science or Adaptive HealthComm Science for short. The services are designed to drive member behavior using a 1:1 experience – what I have called “extreme personalization”. The services have three main components:
- Profiling Consumers
Using descriptive analytic techniques and relevant healthcare data to profile and segment consumers based on both characteristic attributes and past behavior. Member information, claims information, plan details and more all contribute.
- Building decision models
Using predictive analytic techniques and business rules to define decision models – what kind of intervention will be most effective at getting the desired behavior. Typically this begins with a fairly experiential model but it evolved thanks to â€¦
- Adaptive control
Using constant champion/challenger testing to assess and measure success and integrate what is learned back into improved models.
The new services also support dynamic pathing within calls, calculations during the call – both payment math and predictive scoring – and remember more preferences and interactions for ongoing improvement. Silverlink has seen some great results from this technology-based approach. While this approach – which is basically Enterprise Decision Management or EDM – is widely used in financial services it has yet to be applied in healthcare outside of fraud detection. This means that big movements in behavior are possible – not just shaving a few points here and there. One plan, for instance, used the technology on 1.4M out of their 25M members and was able to see a 4x improvement in identification of other coverage resulting in $20M opportunity for savings! When the Vioxx recall was announced at 8am, Silverlink was making personalized, interactive calls to patients in some plans by 11am to pass this information along.
From my point of view it was great to see the approaches and technologies of EDM – business rules, predictive analytics, adaptive control – being applied to a new area and I wish them luck.