Leveraging Analytics Solutions to Improve Value-Based Contracting, Population Health

Amid the pandemic and recent legislation, health care professionals have increasingly devoted time and attention to improving interoperability and closing gaps in care.

In this interview, Dan Ross, cofounder and CEO of CareJourney, describes how data integration and analytics solutions can help payers and providers better approach value-based contracting and enhance population health.

Can you comment on the challenges providers and payers face regarding data and analytics?

If there is anything I have learned in nearly 30 years of being in the data and analytics business, it is this: the challenges of data aggregation and integration, and actionability of analytics, cut across industries and technologies. It is certainly true in health care.

From an aggregation standpoint, it is difficult and expensive to assemble the people and technology you need to leverage your internal data in its many forms. In health care, we have an additional burden and responsibility to keep the data extra secure to protect patient privacy.

Wrangling internal data is difficult and challenging enough. One of the big issues we help our customers address is going beyond their own data to get a picture of what is happening in their city and market, with similar caregivers and plans, and their position in the value-based care ecosystem in terms of cost or quality.

When it comes to data integration, unfortunately, for all the incredible advances we have had in recent decades in computing power, hardware performance, and database architecture and tools, there has not been groundbreaking technology to ease the labor intensity of integrating different data sources.

To illustrate my point, some of these technological designs are like art. If you asked 10 artists to paint you a picture of a house, you are likely to get 10 nice paintings of houses that look nothing like each other. Similarly, you could ask 10 talented technologists to design a database to track claims data, and you are likely to get 10 great designs, but then someone must stitch those databases back together to integrate them. You must understand the design principles and reasoning behind those databases to figure out how to make them work together. That is the manual work needed almost every time for data integration.

This is often referred to as a "interoperability problem." There are standards that can help with some of this, but even then, there is a fair amount of leeway in what the final output is. It is like telling the artist, "You have to paint a house, but you can only use red and blue paint." That will help somewhat, but not a lot.

Finally, and perhaps most critically, the actionability of analytics: you must do something useful with all the data you have integrated, and distilling credible, actionable analytics from billions of claims or human resources records is not easy. I like to say the best analytics are those that are used, and that means adoption by people who can use them to make actual decisions.

In health care, especially, there is a premium on getting clinician buy-in, so solutions must be presented in a way that clinicians can make sense of. Solutions must be understandable and actionable in a way that does not necessarily require new workflows. This requires integrations into existing tools, since nobody wants another screen to log into.

Thank you, Dan. What is the impact of these challenges on value-based contracting?

We find people asking a few questions, especially in value-based contracting. The first one, and the most basic, is, should I be in this value-based contract?

These contracts fundamentally must have some element of performance to them that are based on benchmarks, as we know. The focus is moving beyond PCP-centric models to the specialty level. Organizations need to estimate whether they can be successful in value-based care models, and without data, they are flying blind on that question.

As you move toward a 2-sided risk model, the first principle is, should we be in this contract or not? Another common question is, what are my opportunities to succeed in this model or contract? What else is happening in the local market that might impact our success?

Are we particularly good at treating certain types of patients or handling certain situations? Does the market lend itself to my strengths? Who else is in the market? What models are they in? What are the spillover effects between different models? How does that impact our practices or patient populations? You must understand what you do well, and that comes from looking at your own data, but you also need a situational awareness of the market around you.

Lastly, once you enter a contract, you must ask, how are your performance and results comparing to what might be considered good performance? A lot of that comes down to benchmarking and what the expectations of the model are. Ultimately, organizations have to figure out how to apply limited clinical or care coordination resources to make some impact. The key is execution, so you need up-to-date and accurate data about what is happening to make this work.

Speaking of up-to-date and accurate data, how can improving data affect population health as a whole?

At some level, it seems obvious that taking true ownership of a population, in terms of its health and outcomes, requires data, software, and analytics. One of the basic tenets behind population health is that a provider or health plan needs to worry about 90% of the patients they are not interacting with to anticipate their needs and get ahead of any challenges through outreach or intervention. You cannot do that without good data and technology.

From a patient standpoint, there is a tremendous amount of analytic horsepower and professional expertise devoted to grouping people logically based on potential commonalities in their expected care journey. This allows us to design a better plan for each cohort based on what we know from past data about how certain populations tend to evolve over time.

In that sense, population health is getting more tailored to the individual patient and their needs. For example, if we can keep a prediabetic patient from full onset diabetes for even 5 years, that benefits the patient's quality of life, the costs of their care, and certainly the health care system.

Down the road, I think we will be talking about cohort health instead of population health, because it is going to become that tailored and specific to what people need.

What should stakeholders keep in mind when seeking new analytics or interoperability solutions?

I would say that one consideration is whether the technology can incorporate data beyond your own, as we have mentioned. The idea that you can only look at your own data to succeed in some of these contracts is wishful thinking. Look for solutions that also include a broader picture of the market at large.

Success or failure depends on getting this right up front. You do not want to wait until you are 9 months into a contract to figure out how you are doing.

Secondly, time to value is super important. Adopting a new analytics solution is a big investment by the busiest people in any company. Anything that requires a long data integration project before being deployed might entail several months or a year before you can derive value from it. The truth of the matter is some of that process is unavoidable, but ideally you can find a solution that has something you can implement on day one.

Lastly, it is important to consider the ability of some of these solutions to integrate into existing workflows. It is one thing if you must train 10 people on a new system and click paths to derive value, but if you want hundreds of people to adopt it, it would be much better if you could plug the solution into some existing tool or workflow.

Thank you. Is there anything else you wanted to add today before we wrap up?

I think one thing to point out is the increased focus on health equity in the discussion of value-based care and population health. This is a rather substantial data problem right out of the gate. I think people understand, intuitively, that there are inequities that need to be addressed, but to get your arms around that requires you to measure it somehow.

That means sifting through a lot of data to get a sense of it, and we are seeing more and more people asking for help. That has become another key need for some of the big data solutions that are out there.

Thank you for having me.

About Dan Ross

Dan Ross is the cofounder and CEO of CareJourney, which helps providers and payers along their value-based care journey and addresses strategic challenges including and beyond population health. CareJourney serves over 120 customers.

Mr Ross has spent most of his career in software and analytics with a focus on helping the largest companies in the world use data to make better decisions. He has spent the last 9 years of his career in health care.