Use case 1: How does geographic variation shape healthcare payer strategy?

---

Use case 2: As a healthcare provider, where are the greatest opportunities within a service area to improve care delivery and patient outcomes?

---

Use case 3: As a self-insured employer, how do my healthcare spend drivers vary by region for my employees and their dependents?

---

Use case 4: As a life sciences organization, where can we focus efforts to maximize population health impact?

---

Use case 5: How can policymakers use demographic, health, and social risk data to inform policy decisions?

---

Data sources, caveats, and limitations

The interactive map included in this article displays findings from a variety of Milliman health products and data assets. The sections that follow are intended to help readers interpret the findings by summarizing the data assets behind the interactive map and highlighting important caveats and limitations. Users interested in learning more about the methodology or how these insights may be leveraged for their organization are encouraged to reach out to their Milliman consultant.

Milliman HCC Prevalence Analytics:

Milliman’s HCC Prevalence Analytics leverage 100% Medicare data, both the Research Identifiable Files (RIFs) and the Medicare Advantage encounter data, to quantify condition burden across fee-for-service and Medicare Advantage populations. These data support consistent measurement of HCC prevalence across geographies. We calculate the share of beneficiaries with selected HCC-defined conditions, enabling comparison of disease burden across counties, states, and population segments. These metrics support geographic benchmarking, highlight variation in documented morbidity, and inform population health, coding patterns, and risk adjustment insights.

We leveraged Milliman’s HCC Prevalence Analytics data to generate the following metrics in this interactive map:

• HCC: % with Diabetes Disease Group
• HCC: % with Heart Disease Group

Milliman Annual U.S. Healthcare Census:

Milliman Annual U.S. Healthcare Census is a dataset that combines the U.S. Census Bureau's American Community Survey (ACS), Current Population Survey (CPS), and other public data sources into a unified, standardized view of healthcare coverage and population characteristics across the 50 states and the District of Columbia. A calibration framework adjusts for known limitations in self-reported survey data and aligns coverage totals to external benchmarks from the Centers for Medicare and Medicaid Services (CMS), the U.S. Department of Veterans Affairs (VA), the Military Health System, state agencies, and other authoritative sources, producing consistent, comparable estimates across detailed government assistance, VA/military, and commercial coverage categories.

The dataset supports analysis across multiple geographic and demographic dimensions and accommodates individuals with multiple concurrent coverage types, supporting applications such as healthcare market assessments, state policy work, and longitudinal analyses of the coverage landscape.

The underlying ACS and CPS data are sample-based and rely on self-reported information that may be inaccurate or incomplete. Confidentiality protections and methodological assumptions may also affect results, and estimates may differ from other published sources.

We leveraged the Milliman Annual U.S. Healthcare Census data to generate the following metrics in this interactive map:

• % 18 and under
• % 19-34
• % 35-64
• % 65 and over
• % Disabled
• % High School or Less
• % Some College
• % Bachelor’s or Higher
• % 0-100% FPL
• % 100-138% FPL
• % 138-200% FPL
• % 200-400% FPL
• % 400%+ FPL
• % Non-Hispanic Black
• % Non-Hispanic White
• % Hispanic
• % Other Race and Ethnicity
• % Crowded Housing
• % VA/Military
• % Commercial
• % Government Assistance
• % Government Assistance: Dual Eligible
• % Government Assistance: Medicare (non-Dual)
• % Government Assistance: Medicaid (non-Dual)
• % Government Assistance: CHIP
• % Government Assistance: Other
• % Uninsured

Milliman Episode Grouper with Risk Adjustment (ERA):

Milliman ERA is a clinically meaningful episode grouper that reflects patients’ real-world care experiences in the healthcare system for specific medical conditions. Milliman ERA episodes—made up of procedural, acute condition, chronic condition, and oncology episodes—can equip organizations with the information needed to better analyze their specialty spending, identify opportunities for cost savings, or design targeted value-based payment models that promote efficiency and quality. Episode-specific risk adjustment models ensure apples-to-apples comparisons across providers, while risk-adjusted commercial and Medicare benchmarks ground every insight in market reality.

We leveraged Milliman ERA to generate the following metrics in this interactive map:

• Return to Hospital Rate within 7 Days of a PCI
• Return to Hospital Rate within 7 Days of a Knee Arthroplasty
• Risk-Adjusted Cost Distribution of PCI Episodes
• Risk-Adjusted Cost Distribution of Knee Arthroplasty Episodes

Milliman Population Health Grouper:

The Milliman Population Health Grouper is a total-cost-of-care grouping software that organizes claims data into events and assigns them to clinically meaningful categories that align with how patients interact with the healthcare system. Rather than focusing on how benefits are administered, the Population Health Grouper offers a care-oriented lens by aggregating claims data based on clinical areas, enabling more actionable insights into drivers of spend from a population health perspective. The software includes cost and utilization by site-of-service, as well as a divertible ED algorithm, designed to support ED utilization management efforts. The condition cohorts allow for further analysis at the subpopulation level for members with common, high-cost conditions (heart failure [HF], hypertension, coronary artery disease [CAD], diabetes, chronic kidney disease, end-stage renal disease, asthma, chronic obstructive pulmonary disease [COPD], and lower extremity osteoarthritis).

We leveraged the Population Health Grouper to generate the following metrics in this interactive map:

• Rate of PCIs among Patients with HF or CAD
• Rate of Knee Arthroplasties among Patients with Lower Extremity Osteoarthritis

Milliman Star Rating Market Intelligence:

Milliman Star Rating Market Intelligence (SRMI) is one of Milliman’s Medicare Advantage analytics solutions focused on Star Ratings performance and benchmarking. It uses CMS’s 100% Medicare Research Identifiable Files (RIFs), including the Medicare Advantage Encounter Data and Medicare Part D pharmacy claims data, to analyze quality measure performance across the Medicare Advantage and Part D landscape, enabling insight into performance trends, market movement, and geographic variation.

We leveraged SRMI Part D benchmarks to report statewide diabetes medication adherence rates for Medicare Advantage and Medicare Part D Prescription Drug Plan (PDP) beneficiaries.

Milliman Transformed Medicaid Statistical Information System (T-MSIS) Healthcare Utilization Metrics:

The Transformed Medicaid Statistical Information System (T-MSIS) is CMS’s nationwide Medicaid and Children’s Health Insurance (CHIP) database. It includes detailed information on enrollment, claims, encounters, providers, managed care plans, and other payments, including capitation and supplemental payments. As a standardized, comprehensive data source for all states, T-MSIS enables more consistent analysis of Medicaid and CHIP programs to support data-driven decision making by policymakers, healthcare organizations, researchers, and other stakeholders.

To illustrate the analytical value of the T-MSIS data, Milliman has summarized a defined subset of the Medicaid and CHIP experience using the 2022 and 2023 Release 1 versions of the T-MSIS Analytic Files Research Identifiable Files (TAF RIFs).

T-MSIS-based metrics included in the interactive map are based on fee-for-service claims and managed care encounters, with no completion adjustments applied. Service categories were assigned using Milliman’s proprietary Health Cost Guidelines™ (HCG) Grouper software and reflect analytic groupings used for this summary; they may not align exactly with state-specific reporting categories. TAF RIFs data quality varies by state, year, and metric. Additional data quality information is available through CMS’s DQ Atlas.

This report includes only the 50 states and the District of Columbia. Eligibility groups with limited coverage, special-purpose eligibility groups, and eligibility records with missing or unknown eligibility group codes are also excluded. Although the metrics are presented for state-to-state comparison, differences should be interpreted with caution because reported data quality, program structure, and covered services vary across state Medicaid programs. We have not adjusted these metrics to normalize for differences in program design, benefits, or reporting across states.

A state’s metric is redacted if it falls below 50% or above 200% of the nationwide average, based on the 50 states and the District of Columbia. A state’s metric is also redacted if the following DQ Atlas topics had a data quality assessment of “unusable.” This methodology is intended to improve interpretability, but it does not eliminate all data quality and comparability limitations. The nationwide averages in the report are calculated by excluding the redacted values.

We leveraged T-MSIS-based utilization results to generate the following metrics in this report:

• Facility Inpatient Days per 1,000 (Non-Maternity)
• Facility Inpatient Days per 1,000 (Maternity)
• Office/Home Visits per 1,000
• Prescription Drugs: Scripts per 1,000
• ED Facility Visits per 1,000, separated by the following major eligibility groups:
• Aged, Blind, or Disabled (ABD)
• Medicaid Expansion Adults
• Non-ABD, Non-Expansion Adults
• Children

---

General caveats and limitations

• Milliman does not intend to benefit or create a legal duty to any third-party user of its work product.
• The results included in this interactive map are based on historical data. These summaries are intended for exploratory comparison and to highlight the types of analysis that can be performed using Milliman’s products and data resources. Emerging and future results will differ from historical results.
• Milliman has developed certain models to estimate the values included in this interactive map. The intents of the models were to estimate the results for demographic, condition-specific, healthcare coverage, healthcare utilization and provider density metrics and to present results in an interactive map. Milliman has reviewed the models, including their inputs, calculations, and outputs, for consistency, reasonableness, and appropriateness to the intended purpose and in compliance with generally accepted actuarial practice and relevant actuarial standards of practice (ASOP).
• In developing this interactive tool, Milliman relied on data maintained by third parties, such as the ACS, CMS, and the Health Resources and Services Administration (HRSA). Milliman has not audited or verified this data or the combined data for accuracy or completeness. If the underlying information is inaccurate or incomplete, the content of our report may likewise be inaccurate or incomplete.
• Any user of this interactive map should possess a certain level of expertise in areas relevant to this map to appreciate the significance of the limitations and the impact of these limitations on the results. The user should also be advised by their own qualified professionals competent to properly interpret the material.
• The information displayed in the interactive map may not be appropriate for all uses. If users leverage information from the interactive map for purposes beyond the Milliman website, Milliman requests that the information be distributed with a reference to this article.

---

Acknowledgements

The authors would like to thank Stoddard Davenport and Maggie Alston for their thoughtful reviews and contributions to this insight, as well as Rebecca Driskill for her editorial contributions.

The interactive map was prepared with the assistance of many people. The following individuals provided valuable insights and support:

The American Academy of Actuaries requires its members to identify their credentials in their work product. All individuals above marked with an asterisk (*) and author Harsha Mirchandani are members of the American Academy of Actuaries and meet its relevant qualification requirements.