It’s no secret that patients in community health plans experience outsized risk of chronic health conditions and diseases such as diabetes and depression that significantly impact health outcomes. For example, according to the CDC, 38% of Americans over 18 have pre-diabetes and many of these people are currently covered by community health plans. Unfortunately, community health plans typically lack a meaningful ability to predict members’ needs before they meet the criteria for a diagnosis of a serious medical disorder and are already on a poorer outcomes trajectory.

Author: The Siftwell Team

This impacts member health and the community health plan’s organizational health.

A key reason why predictive analytics tools have fallen short in the community health plan space is that these solutions have been built for more resourced, commercial, health plans. As it turns out, one size does not fit all. Population Health and Quality teams within community health plans, and the Analytics teams that support them, have been asking themselves:

  • Does integration cost time or give it back?
    • Traditional population health approaches often require them to devote limited resources to months of data collection, system implementation, and team training on yet another platform. Does this integration process ultimately make their ability to serve our members easier?
  • Are the insights clear and applicable to the community health plan segment?
    • Off-the-shelf analytics solutions built for commercial health plans are cumbersome to use and add more to teams’ overflowing plates. Results need contextualization and processing before they can be used for campaign design and execution, and often lack the granularity needed for local, complex population health needs. Community health plans lack data science and population health resources as is, and cannot afford to increase team sizes to be able to customize the outputs of off the shelf products.
  • They show which members are emerging in risk, but not what to do about it.
    • Predictive analytics falls short in helping plans understand how to “move the needle” for the members who need us most. They often know who needs support, but do not have clarity on what interventions to deploy against them and when, and how these initiatives are working.

How AI Can Help

Artificial intelligence (AI) grew out of traditional predictive analytics and builds upon its strengths and weaknesses. There is an acknowledgement that AI might help address the challenges above, but folks tell us that integrating AI seems like a massive lift that is suited only for the “bigger” payors. Many community health plans lack the internal know-how to evaluate and build AI tools, the financial resources to hire capable staff who would possess this know-how, and the time to seriously dive into this sector and see where AI could really benefit their operations.

As a result, unfortunately many community plans avoid AI altogether, assuming that the technology is too time-consuming or requires big technical teams and an even bigger budget.

However, without AI, they could fall behind when it’s time to recompete for state contracts.

The good news: AI should make life easier for everyone- not more complex. Integrating AI with Siftwell will give you time, not take it. Siftwell insights will add clarity, not extra work. And our very human team (who understands the challenges of the public sector) will enable your clinical operations teams to drive smarter and more tailored initiatives for more people.

Siftwell’s AI was built for community health plans’ clinical operations teams

Benefits can be realized by community health plans sooner than you might think. Here are two ways that Siftwell’s AI tools can help community health plans.  Both relate to identifying members within a population most likely to rise in risk due to new diagnoses, comorbidities, and/or non-compliance with preventative health measures, and why. From there, we focus on identifying the next best action for each of these members. We tell you the who, the why, and what to do about it.

Population health outcomes improvement 

AI-powered population health models organize data to optimize organizations’ bandwidth to promote health and wellness initiatives, in a tailored way, to those who need it most.

That is, they allow you to streamline your care and access improvement efforts, meaning the right resource for the right member at the right time.

Machine learning models are capable of assessing broader swathes of data to pinpoint specific member needs (as opposed to broad-risk cohorts). Therefore, the fidelity, accuracy, and precision of outputs outperform traditional predictive analytics approaches. All of that happens while reducing the workload for your team.

For example, the Siftwell Population Health Model aggregates unstructured data to create distinct Member Profiles. These Profiles—some as small as a few dozen people—contain members who share similar clinical risk factors, as well as social determinants of health (SDOH) and health related social needs (HRSN).

Once Member Profiles are created, Siftwell also assesses and offers direction on which available interventions will impact whom the most. For the first time, you can answer questions like:

  1. Given a specific Member Profile and their health-related social needs, which intervention(s) will drive the best outcomes for them and disrupt cost trajectories the most?
  2. Given a specific intervention at my disposal, which Member Profiles should be addressed to drive the best outcomes and the best cost trajectory impact?

Get razor sharp with Quality Improvement levers

VPs, Directors, and Managers in Quality departments understand that in the public sector, every initiative—and every penny—counts.

That’s why the Siftwell strategy involves identifying not only 1) who among your members needs extra support due to emerging clinical risk but also 2) who would be most likely to engage in a quality-related health activity (e.g., annual wellness visits, breast cancer screenings, etc.) that is associated with mitigating that clinical risk.

We use propensity-to-engage models and sentiment analysis to identify who is most likely to undertake a quality measure-associated health behavior or activity so Quality teams’ efforts are centered on the most impactable members. The information the teams have on each member is also rich in description of that member’s clinical and social risks so frontline staff can empathetically engage those members and maximize hit rate and conversion. Campaigns using Siftwell’s data help Quality teams prioritize their resources by growing their “numerator” for relevant quality measures such as annual wellness visits, and the community health plan benefits overall by improved member health.

The bottom line

The bottom line? Use Siftwell to find the members most likely to engage with you, and most likely to see a clinical benefit, and prioritize those members for outreach and interventions. This allows you to sustainably hit your quality goals while also driving population health outcomes.

There’s no need for collecting the “right” templated data. We’ll take what you already have, often pulled from mandated state reporting, and provide you with solutions that address your organization’s main concerns. Plus, Siftwell solutions don’t require any front-end work or organization-wide system training on your part. We bring AI and its benefits to you, without the daunting task of evaluating it yourself.

Our success is your action—the action you take to improve member lives in your communities.