The Power of Cloud for Healthcare Data

If you answer, yes, to any of these questions you might be dealing with data silos.?

·???????? Does your data live in different places throughout your organization??

·???????? Do your technical and business areas use different tools to work with your data??

·???????? Do your business rules vary across the company??

·???????? Do you have to spend time on reconciling analytic results that differ??

?

Does any of this resonate?? One consideration that is possible to change the answers to “no” is creating a reliable single source of truth with common tools and platforms.? Over two decades ago, I embarked on my data career by constructing secondary data warehouses and data marts.? This data was mainly used by our actuaries and underwriters for pricing.? I realized early on data has no meaning on its own. ?That’s when I discovered analytics as a system of thinking about our data.? I have observed first-hand how analytics brings data to life and provides a form of translation that enables data to tell a story.?

?

Data is power, but only if you know how to use it.? I saw firsthand within my family and community how healthcare was unaffordable.? This became my passion: finding ways to improve the healthcare affordability challenge.? To truly impact this, I knew I needed to pursue an actuarial career so I could influence the pricing of our products along with giving our members richer benefits.? For my career aspirations, I saw the perfect trifecta being the areas of data, analytics, and actuarial.? This passion has driven me as an actuary and data and analytic thought leader for more than twenty years’, during which time I have strived to innovate and transform health care from within.? I am excited to share my knowledge of data and analytics across healthcare, with a focus on my most recent years at Priority Health.? Priority Health is an integrated health plan for the Corewell Health system operating in Michigan.? Corewell Health and Priority Health have a vision where health is simple, affordable, equitable and exceptional.?

?

Unfortunately, health has not been affordable.? Currently we have seen national healthcare expenditures of $4.1T in 2020 and is expected to grow to $6.2T by 2028. This is not a trend we want to see.? The healthcare industry has been on a journey for many years now to transition from a fee-for-service payment to value-based care payment.? As an integrated healthcare delivery system, we can innovate and test new ways of payment without impacting the overall risk within our organization.? We have an aligned vision and mission to improve the health and wellness of our community.? Everything we do is about placing our members/patients at our core.

?

As an integrated delivery system with both a health network and a health plan, the value of appropriate data sharing seamlessly is critical.? As we look across our industry and customer base, data agility, speed to decision making, and providing proactive analytics need to be the focus.? Integration allows us to know the whole person better and share data in more efficient ways to address holistic patient needs.? Prior to our migration to a cloud-based platform this meant creating multiple copies of data with minimal governance and controls.? One central source of key business data assets with governance and access policies applied to the databases and schemas allows appropriate team members to work from the same single source of truth and not rely on legacy extract, transform, and load (ETL) processes.?

?

We want to transform the old way of thinking about data sharing and allow for more real time information and key insights.? We often reference the quote by Albert Einstein - “We can’t solve problems by using the same kind of thinking we used when we created them.”? One of the most important uses of our data is identifying what is driving the overall cost of healthcare.? We are leveraging our data and analytics for fast business decisions in our employer group analytics and medical trend management.? We have many initiatives to improve the cost of healthcare and well-being of the community.? These initiatives are tracked and accounted for in our pricing.? We want to provide the data and analytic needs to help our providers, employer groups, and business partners have what they need to address the affordability challenge with us.? For example, we want to provide our provider organizations throughout Michigan with appropriate shared views of our members information to help them proactively address health concerns.? We do not want them to wait to address helping patients close care gaps until after their physician visit.? Rather, point this data upstream where the provider can address these gaps during a physician visit providing a favorable outcome for the patient.?

?

Looking ahead to the future of the industry, imagine this: providing members with direct access to their data, allowing them to gain personalized insights from a single, reliable source.? We already have members opting into a program called “All of Us”, a research program funded by the National Institutes of Health (NIH).? Members are taking charge and control of their health and wellness.? Consider how many people are using watches and phones to track their data.? These same members will want the ability to incorporate their provider and payer data along with genomic data and other types of data (e.g. steps, heart rate, nutrition, etc.) to get upstream in their own health and wellbeing.? ??

?

Priority Health has been leveraging our data with a cloud-based data storage and analytics company.? Enterprises choose a cloud provider like AWS, Azure, Google cloud, or others and set up their accounts within a chosen region.? Data is then ingested from various sources, transformed, and analyzed.? I have been part of many data warehouse builds, transitions, and migrations from different data platforms.? Some have gone well, and others have not.? However, in reading “Not Impossible” by Mick Ebeling, "If you're always worrying about what-if-I-fail, you're never going to give yourself much of a chance to succeed."? Being resilient is key when you are in the data and analytic world.? At Priority Health I have seen collaboration between the business and technology teams that is truly unique.? We have thrown out the old way of thinking, truly listening to each other, and working together to build what our company needs.? We have seen the success that we can continue to take beyond our data platform and throughout our transformational journey.? We have had our business leading the development.? The old way of thinking would have been something like this: you give requirements to your Information Technology area and then they will build it, creating a prototype, sharing, and then you both go back and forth with rebuilds.? I find this way of thinking very ineffective and wasteful.? We have removed silos and instead are using the talent across our whole company to build what the business requires.? We took time to not just pick up the old environment and move it, but to engineer our data in a way that allows us to leverage it in new ways so it can be faster and more efficient and make it AI-ready.

?

In our on-premises data environment, we had challenges with lengthy data refresh cycles.? This combined with extremely lengthy processes for new data transformations – with a heavy reliance on IT and too many copies of data.? With our data platform, we focused on shrinking the “time to usable data” by having both business and IT working in the same platform and with common tools.? Our IT partners source the data, which is ingested, or landed into a raw layer (which we call “PH Land”), and then our business data engineers model the data into the key domains that power our business demands.? The combination of our cloud and technology partners has taken refresh processes that used to run an entire night down to under one hour.? This is a complete refresh of the entire warehouse, not limited to an incremental load.? Now, this has transformed and built a foundation for everyone at Priority Health to leverage our data and turning it into our most powerful asset.?

?

The value of the cloud for healthcare data has brought us the gift of time, and we could all use more of that.? For example, one of our population health queries always failed before completion. To be successful, it had to be broken into monthly chunks and took eight-ten hours to run monthly and now runs in less than 10 minutes.? Processing twenty years of eligibility data can happen within seventy minutes.? In healthcare, on-premises technology has limitations, and typically you could effectively only work with 3-5 years of data.? This performance is astonishing, and we are seeing daily how this data platform is changing processing time.? Before this data platform, business transformations and logic were often kept outside of IT-managed production environments, so anything could be fixed rapidly.? Now we all work in one set of tools, with transformation all happening in our cloud platform. ?It provides the right level of agile access to data while also providing full auditability, security, governance, along with other IT appropriate controls.? Our environment also makes it easier to increase the level of self-service that can be created for business teams such as provider and population health.? Having a single source of truth with the ability to instantly scale changes the way we think about leveraging the tools in our toolkit, such as Tableau and PowerBI, and provides opportunities for data science applications at scale.?

?

We have been front runners in testing new features in our data platform. Those experiences have unlocked new use cases with our data.? As a not-for-profit health plan, our margin is small, and we invest it back into making health care more affordable for our members.? This means the more we can customize our analysis to our member population and their unique needs, the better we can drive positive health outcomes for our members and overall lower costs.? We focus a lot on population health analytics and creating extensive care management algorithms.? We examine specific social determinants of health barriers such as transportation, medication adherence, food insecurity, etc.? With our cloud data environment, our data scientists can source the data they need directly to build the AI/ML models.? We have been an early adopter of capabilities that lets our data scientists do their custom python work directly in our cloud environment.? They can build, maintain, and troubleshoot complicated models as easy as simple models.? There is no special access required for different tools and no python knowledge required for maintenance.? To top it off, all data stays in our cloud environment where the environment is more scalable, efficient, and governed.?

?

We are also leveraging an application within our cloud platform.? We see opportunities to create agility with delivering applications quickly to our customers that can be tailored to their exact needs.? Let me highlight a couple of examples where we are building data apps.? There is a risk adjustment model with different variables from claims (such as age, disability status, medical conditions, etc.) for a member that gets summed up to a risk score.? We have converted the model from SAS to SQL so we can do custom runs on-demand.? We are now implementing the model and intend to provide a user interface to do “what if” analysis and recalculate risk scores on the fly for certain diagnoses.? This step requires complex transformation logic for which python is most flexible.? We have created a machine learning model that we are using an app to monitor it’s use.? At a macro level, apps can help to visualize and provide better understanding of the data and predictive models.? Teams could leverage these apps to create member lists for different operational processes and write back to the database.

?

At Priority Health we are making data more actionable in the healthcare ecosystem.? Healthcare data is evolving.? I remember when I first started my career over twenty years ago, mainly insurers were only covering medical claims. ?Then gradually pharmacy claims were covered, followed by supplemental benefits (dental, vision, hearing, other) as well as lab results and transition of care plans.? And now we collect social determinates of health data as attributes impacting the delivery of care or health of the patient.? We can all guess what’s coming next for our evolution of data.? We are collecting new data like never before and this creates new opportunities and challenges.? We spend a lot of time looking through the data and how cost of care can come down, so it is affordable and accessible.? A lot of our efforts are using this single source of truth data to look for actions to achieve both those outcomes.?

?

We are not only transforming our data, but we are transforming how we work with our data and use it through our analytics.? By migrating to the cloud, we have accelerated the decision-making process.? Things are taking much less time, and we are seeing some great successes.? This gift of time will allow us to continue to advance our analytic maturity to be more proactive, predictive, and prescriptive, less reactive.? We are already beginning to see the future of how we will be interacting with data and analytics.? I have been through the data transformation and now I’m ready to see analytics be transformed from static dashboards and reporting to interactive large language model machines.?

?

Timely, accurate, and proactive information is key to the success of our company.? To support our growth aspiration and help providers transition to value, we knew we had to take the necessary steps to modernize our data warehouse and advance our analytics.? This makes for happy customers (employer groups, providers, vendors, and our members) and builds trust while removing the friction of data concerns. ?Our providers are more engaged and able to focus on our collective #1 priority, the health of our members.? Time will only tell what our future holds, as we take a 2-day process and it now runs in 10 minutes, just think what that will give our analytic teams time to explore and discover.

?

Endnotes:

·????????? Corewell Health is a not-for-profit health system providing health care & coverage with an exceptional team of over 60,000 dedicated employees—including more than 11,500 physicians & advanced practice providers and more than 15,000 nurses offering services in 22 hospitals, 300+ outpatient locations and several post-acute facilities—and Priority Health, a provider-sponsored health plan serving over 1.3 million members.

·????????? The National Institutes of Health (NIH) is?the nation's medical research agency?— making important discoveries that improve health and save lives.