State-level all-payer claims databases (APCDs) provide state governments and researchers lots of detail about payment patterns, but two APCD leaders recently discussed ways in which they can answer lots of other questions involving non-claims data as well. 

A recent Civitas Networks for Health webinar included a presentation by Kristin Paulson, J.D., M.P.H., president and CEO of the Center for Improving Value in Healthcare (CIVHC), the designated administrator for the Colorado All Payer Claims Database. Paulson, who has been with CIVHC for over 13 years, said she’s had a chance to “see the evolution of what people are able to do with claims data and what kind of insights can be derived using claims data that go far beyond just the cost data that's included in the claim itself.”

First, she detailed some of the high-level questions that can be answered about who's represented in the claims: Who is receiving care? What are their demographics? Where do they live? What kinds of services do they receive? Where do they receive those? How much do they pay? Do they follow appropriate care patterns?

From the provider side, they can track: Who was providing the healthcare services? Why were those services needed? How much do they charge? What was negotiated? What was covered? How far are patients traveling for care? "That's a huge piece that's currently being looked at in terms of geographic access, quality of care, and population health,” Paulson said. 

Another thing they can track is emergency rooms being used as primary source of care. They can also track: What does the case mix look like? Who's accepting Medicare and Medicaid patients? Who's not? And then finally, the cost of care: Has funding changed over time? What are the drivers for spending? They can look at how payments for service differ by setting location or plan. Is total cost of care in line with expectations? 

CIVHC also collects a lot of information that doesn’t specifically involve claims files. It has a drug rebate file that reports by payer and by plan spend, as well as a count of prescriptions that are brand specialty and generic. It also has a requirement to report by therapeutic class. 

Paulson noted that alternative payment models are a big focus. To gather this data, it uses the Non-Claims Payment Common Data Layout (NCP CDL), a file structure created by the APCD Council, a learning collaborative of government, private, nonprofit, and academic organizations focused on improving the development and deployment of state-based APCDs.

“Collecting these data is critical to be able to capture the full cost of care and to understand how we are progressing away from fee-for-service payments,” she said. 

She explained that using the NCP CDL allows for easier cross-state comparisons and is an effective mechanism to be able to do multi-state collaboration and comparison. CIVHC adopted this model in 2025. It previously had a model that was based on the HCP LAN Framework. “These two frameworks had a slightly different focus, so that's why it was important for us to keep both of those,”  she said. “There is a crosswalk that was developed to allow folks to convert between those two different frameworks. The HCP LAN Framework focuses on the level of risk that providers are assuming in these alternative payment models. The expanded framework focuses on identifying the purpose of the spending.”

“Cost is an important part of what is in an APCD and a unique differentiator, but it is far from the sole purpose of using those claims data,” Paulson noted. Some Colorado-specific use cases include reporting on primary care investment and tracking progress toward the goal of 50% of expenditures being in alternative payment models by 2030. There are a number of things that are population-based: modeling the COVID vaccine rollout for high-risk populations and modeling the impact of Medicaid policy changes across populations.

Paulson closed by mentioning some creative ways that CIVHC is using claims and non-claims data to provide transparency and an evidence base for organizations making the biggest difference on the ground in Colorado. 

Last December Healthcare Innovation wrote about one of these: A Colorado study revealed that healthcare providers participating in health information exchange (HIE) significantly reduce emergency visits and hospital admissions, especially benefiting Medicaid patients, by improving care coordination and data access. We spoke with Kelly Joines, the study’s co-author and chief strategy officer at nonprofit Contexture, which runs HIEs for Arizona and Colorado. The study was conducted along with CIVHC.

Another example is Project Angel Heart. “We used their program data and combined that with claims and were able to determine the return on investment of food-as-medicine interventions,” she said. “The evaluation actually led to that program growing in size and serving the entire state instead of just the Denver metro area. 

CIVHC is working with another group conducting a study they call the Colorado Food Cluster. This is a grant-funded program working in some impoverished multi-generational high-needs zip codes. A thousand families are being given $500 per person per month to support food purchasing for that household. “We are using community qualitative surveys, the claims data, grocery store data, employment data, school attendance data, and more to look at the impact of this model over the next several years and find out what impact this has on regular healthcare, on well child visits, on medication adherence, asthma emergency visits, Medicare wellness visits, and more,” Paulson explained. 

CIVHC also is working with the Colorado Village Collaborative, a group that's creating some tiny-home villages in the Denver metro area to look at what that means for the health and healthcare of the individuals participating in those programs, specifically looking at ED utilization, medication adherence, and overall intensity of services required in the healthcare system.

APCD in Virginia

Kyle Russell, M.H.A., is CEO of Virginia Health Information (VHI), the nonprofit organization that operates the Commonwealth’s Health Information Exchange (HIE), Emergency Department Care Coordination Program (EDCCP), All Payer Claims Database (APCD) and several other data collection systems.

He mentioned that in Virginia, APCD was the perfect resource for measuring the primary care workforce. “We have a great research relationships with our local university here in Richmond, Virginia Commonwealth University. They use the APCD for all kinds of different use cases throughout the year, and one that's consistently gotten a lot of traction is looking at the size and adequacy of our primary care workforce,” Russell said, “because if you were just to look at how many licensed practitioners we have, that doesn't tell the whole story. What volume of care are they providing? How many are actually active? How many have provided a certain amount of services, and what do those services look like, and how are they distributed across very granular views of the state? So they use the APCD to do all of that analysis, and really built quite a workshop around workforce analysis that's come out of the APCD. I think it is a great use case that that could be replicated in other states.” 

Russell noted that Virginia is not as far along as Colorado in terms of using non-claims data, but added they can't wait to get there, “because the amount of demand and interest in adding non-claims events to our APCD is enormous, and I can't speak highly enough about the work that the National Association of Health Data Organizations has done to build the non-claims payment files into the common data layout that makes it so much more accessible and easier to implement for all states that want to do this,” he said.

Prescription drug affordability is another use case drawing interest in Virginia. “We've actually had a bill be passed five years in a row, and be vetoed five years in a row in Virginia around prescription drug affordability, and it got even more attention than ever this past year,” Russell said. “But I will tell you the part of it that everyone agreed with — and they also agreed the bill wouldn't work without — was making sure that it would update the APCD code section, so that we would get the expanded rebate information. That was viewed as a vital component. I’m not sure if we'll get a prescription drug affordability bill in the same form this upcoming year, but I can almost guarantee you there will be a bill specifically to collect this information, because there is so much interest.”

Currently 25 states have APCDs. Russell stressed that they are amazing vehicles for states and local territories to understand what is happening with all different forms of healthcare in different fashions, and they continue to evolve. “I would say just as an advocate for the industry, we want to have more peers, we want to have more of this information flowing, because it adds so much value in the states that have already them.”