What RHTP Measurement Actually Asks of Your Data

Rural Health Transformation Program (RHTP) initiatives need clinical, claims, health information exchange (HIE) and program data that can be read together: tied to the same person, the same place and the same time period. That is what it takes to produce quantifiable metrics you can defend, including one at the county or community level. When those sources can’t be joined, the metrics are harder to defend.

The Centers for Medicare & Medicaid Services (CMS) has been specific about the measurement expectation. According to the CMS RHTP applicant webinar, states identify at least four quantifiable metrics per initiative, with at least one at the county or community level. (For the program overview, see RHTP data and measurement with CareFlowIQ.)

What is CMS asking states to measure?

The Rural Health Transformation Program commits $50 billion over five years, with an award to every state. First-year awards average $200 million, ranging from $147 million to $281 million (CMS, December 2025). The program sets five strategic goals, including technology innovation (CMS, September 2025).

Each initiative a state funds has to answer for itself: at least four quantifiable metrics per initiative, and at least one that shows up at the county or community level, not just in a statewide total.

States define their own metrics. In practice they tend to fall into four groups: access, equity, quality and cost. The metric list is a policy decision. Whether you can calculate it is a data question.

One more constraint shapes the data plan. EMR replacement is capped at 5 percent of funding where a HITECH-certified EMR was in place on September 1, 2025. For many states, the path to better measurement is connecting the systems already in place, not replacing them.

Why do those measures need connected data?

A metric is a ratio, and the two halves of a ratio rarely live in the same system.

Take a simple measure: the share of adults with diabetes in a county who had a follow-up visit after an emergency department stay. The emergency visit may arrive through an HIE feed. The diagnosis sits in an electronic health record (EHR). The follow-up shows up in claims. The program enrollment that tells you whether the person was part of an initiative sits in a program spreadsheet.

Remove any one source and the metric can be hard to calculate or hard to trust. Add the county requirement and every record also needs a reliable place attached to the person.

Medicaid data and quality leaders know this problem well. A measure is only as defensible as the trail from the dashboard number back to its source records. Program managers feel it from the other side: quarterly reports are due, and subrecipients send data in whatever format their systems produce.

Where does fragmented data break measurement?

Four places come up again and again.

• Systems that weren’t built to be read together. EHRs, claims, HIE feeds and program reports use different structures, definitions and update schedules. Someone has to reconcile them, usually by hand. Rural systems start further behind: in 2023, 36% of rural hospitals routinely engaged in all four domains of interoperable exchange (send, find, receive and integrate), compared with 47% of urban hospitals (ONC Data Brief No. 71, 2024).
• Identity matching. The same person can appear under different identifiers, spellings or addresses across sources. If records don’t link, one person becomes two, or two become one, and every rate built on them shifts.
• Many names for one condition. Diabetes can appear as a diagnosis code in claims, a problem-list entry in an EHR, a lab value, a medication or a free-text note. Unless you define the condition once and map each source to that definition, you are counting different things and calling them the same thing. In one study of 173,503 adult patients, different diabetes definitions identified anywhere from 7% to 13% of the same population, and the patient lists they returned overlapped by only 50% to 86% (Richesson et al., JAMIA, 2013).
• Lag. HIE feeds can move quickly. Claims arrive after the fact. Program reports land on a reporting calendar. A metric that mixes them needs a clear rule for what “current” means.

A fifth problem sits on top of these: provenance. Reviewers, legislators and CMS may ask where a number came from. If you can’t show the source, the definition and the refresh schedule, the number is hard to defend. We cover this in more depth in Why Data Foundation Comes Before Dashboards.

What data questions should you answer before you start?

Work through these before you connect anything. Each one is easier to settle before data starts moving.

1. Which four or more metrics will each initiative report, and which one is at county or community level?
2. What is the exact definition of each metric: numerator, denominator, time window and exclusions?
3. Which source system holds each piece of the definition, and who owns it?
4. How will you match a person across EHR, claims, HIE and program records, and how will you handle records that don’t match?
5. How will each person be assigned to a county or community?
6. How is each condition defined, and how is that definition mapped to codes, labs, medications and notes in every source?
7. How often does each source refresh, and what is your rule for combining fast and slow feeds?
8. Can you trace any reported number back to its source records, and does your documentation show how?

If you can answer all eight, you have the groundwork for a measurement plan. If you can answer only some, you have found where to start.

How does a phased RHTP rollout work?

CareFlowIQ works in three connected phases, Connect, Activate and Expand, so a statewide program builds on one data foundation and widens at its own pace without re-platforming. Scope is defined with your state partner, and each phase uses the answers to the questions above. See how this fits state programs on our RHTP data and measurement page.

1. Connect: establish the data foundation. Connect FHIR, HIE and EHR sources and ingest claims and clinical data through secure, read-only integration, with documented provenance for each source. Integration effort depends on the data your partners can provide.
2. Activate: create the Whole Patient View™. Build a Whole Patient View that moves from population to cohort to patient, apply risk stratification and segmentation, and calculate the agreed metrics, including the county or community view.
3. Expand: widen cohorts and use cases. Add initiatives, cohorts and regions, track outcomes against configured measures at the program level, and adjust as measurement needs change.

This approach supports defensible reporting. It doesn’t promise any particular result, and it works with available data, so the gaps it surfaces are part of what you learn.

Ready to plan your program’s data foundation? Start a partner conversation.

Common questions

How many metrics does each RHTP initiative need?

CMS expects states to identify at least four quantifiable metrics per initiative, with at least one at the county or community level.

Do we have to replace our EMR to measure well?

Not necessarily. CMS caps EMR replacement at 5 percent of funding where a HITECH-certified EMR was in place on September 1, 2025. For many states, connecting existing systems may be the more practical route.

Why does identity matching matter so much?

Metrics count people. If the same person appears as two records, or two people merge into one, the rates are wrong and the county view moves with them.

What is the first step before connecting data?

Write down the definition of each metric, then map each part to a source system and an owner. The eight questions in this article give you the order.

Sources

• Centers for Medicare & Medicaid Services. CMS Announces $50 Billion in Awards to Strengthen Rural Health in All 50 States. December 29, 2025.
• Centers for Medicare & Medicaid Services. CMS Launches Landmark $50 Billion Rural Health Transformation Program. September 15, 2025.
• Centers for Medicare & Medicaid Services. Rural Health Transformation Program applicant webinar.
• Gabriel MH, Richwine C, Strawley C, Barker W, Everson J. Interoperable Exchange of Patient Health Information Among U.S. Hospitals: 2023. ONC Data Brief No. 71. May 2024.
• Richesson RL, Rusincovitch SA, Wixted D, et al. A comparison of phenotype definitions for diabetes mellitus. Journal of the American Medical Informatics Association. 2013;20(e2):e319–e326.