Mathematica Medicare Beneficiary Chronic Condition Data

Mathematica’s Data Innovation Lab

Mathematica’s Data Innovation Lab has access to 100% linkable claims, encounter, pharmacy, and enrollment data for 170+ million lives enrolled in Medicare and Medicaid programs—including managed care. Mathematica provides direct-access to data at the provider, facility, therapeutic, condition, geographic, or population-level. Mathematica also provides indirect access to support client’s needs around real-world evidence, population health, market intelligence, and other services for the healthcare and life sciences industries.

Understand the full Medicare population

Over 63 million Americans are enrolled in traditional Medicare or Medicare Advantage plans, many of which were dually eligible for Medicaid. Due to challenges with data completeness of Medicare Advantage data, the focus of research and insights around Medicare data has leaned primarily towards Medicare FFS beneficiaries. However, now with Mathematica’s corpus of Medicare Advantage enrollment and more complete encounter data, insights can be understood with respect to quality, utilization, and access to care for all Medicare beneficiaries.

Data Products for Medicare Beneficiaries with Diabetes and Other Chronic Conditions

Mathematica is offering insights into the Medicare population through access and integration of the national Medicare and Medicaid databases. This includes inpatient, professional, ambulatory, long-term care, pharmacy, and enrollment data from CMS’s databases. The data are summarized at the lowest level of granularity possible for each population—state, core-based statistical area, or county. These files are derived from beneficiary- and claim-level data that is linkable across the Medicare and Medicaid programs. These data cover 2017 through 2019.

Filling a Critical Gap for Public Payer

As the experts on Medicare and Medicaid data analytics, Mathematica enables health plans, health systems, and researchers to gain insights into populations that are historically underserved and disproportionately impacted by the COVID-19 pandemic.

Data Highlights:

• Users can gain a high-level snapshot of Medicare enrollees and HCCs at the county level, with counts by year and HCC to assess the feasibility of market-based analyses

• The basic data files include counts of beneficiaries with any diagnosis data, mean unscaled HCC risk score, counts of beneficiaries with each HCC across years by state and county, and prevalence per 1,000 of each HCC per year among beneficiaries with at least one diagnosis code present for that year. An additional file is provided filtered to those with HCC 18 (diabetes with chronic complications), this file is limited to state level data due to CMS censoring rules.

• For 2019, a payer-based file is provided to compare HCC risk score and HCC prevalence per state between the Medicare FFS population and MA Encounter population.

• HCCs are assigned based on v24 of the CMS HCC risk adjustment software.

• NOTE: Count and prevalence columns with NULL values are censored per CMS regulations for low sample size (<11 cases). Counts and prevalence with 0 are true 0s.

• An expanded list of chronic condition populations, or a data product focused on a population other than Medicare beneficiaries with diabetes, is available upon request.

Data Dictionary (hcc_by_state, hcc_by_state_diabetes, hcc_by_county*)

Variable: Definition

state = Two character state abbreviation

county_name* = Name of county (*only present on hcc_by_county table)

year = Program year of HCC run, based on prior calendar year’s data (i.e. 2017 bene and HCC counts are based on CY2016 claims)

num_benes = Count of all beneficiaries with at least one diagnosis code present in that year’s data.

mean_hcc_score = Mean unscaled HCC risk score for that state and year among beneficiaries with at least one diagnosis code.

hcc = HCC value and label.

benes_with_hcc = Count of beneficiaries with given HCC.

hcc_prev_per_1k = Prevalence of beneficiaries with given HCC per 1,000 beneficiaries with at least one diagnosis code.

(state_by_payer_19 includes the same columns as above with additional columns differentiating counts, prevalence, and average HCC score by MA Encounter vs. Medicare FFS populations)

Custom Products and Services Available

To learn about other data products and services offered by Mathematica, email DataInnovationLab@mathematica-mpr.com . We offer data products and analytics focused on the Medicare and Medicaid populations, with emphasis on cost, utilization, and quality of care benchmarks, provider network analytics, and health economics outcomes research.