danyarm/ma-a1c-heor-analytics

MA T2DM A1C Gap — HEOR Analytics Source Data

Graphify.md · §3A Compact Knowledge Graph Methodology · Patent Pending · graphifymd.com

Dataset Summary

This dataset contains the source data, clinical trial evidence, cost benchmarks, and population funnel underlying a Health Economics and Outcomes Research (HEOR) analysis of inadequately controlled Type 2 diabetes in Medicare Advantage (MA) plans (May 2026).

The analysis was conducted by Daniel Yarmoluk (Graphify.md), Mohan Nadarajah (payer analytics), Dan McCreary (knowledge graph architecture), and Jim (commercial strategy) in a live working session.

The core finding: ~2.2 million Medicare Advantage members have A1C >8% (inadequate control by ADA 2024 standards). Annual excess medical cost vs. controlled: $9,600/member/year. Total exposure on a 100,000-member plan: $67.2M/year. Therapy intensification generates a 3.1× ROI over 3 years.

---

Files

File	Description
data/t2dm_ma_prevalence.csv	T2DM epidemiology in Medicare/MA — prevalence, incidence, A1C distribution
data/cv_outcome_trials.csv	9 published CV outcome trials (LEADER, EMPA-REG, SUSTAIN-6, CREDENCE, etc.)
data/cost_benchmarks.csv	Medical event costs, drug costs, PA metrics — all with sources
data/population_funnel.csv	10-layer actionable population funnel for 100K-member MA plan
process/methodology_overview.md	Step-by-step HEOR analytical process — replicable without proprietary data

---

Key Statistics

Metric	Value	Source
T2DM in Medicare Advantage	27–29%	CMS CCD Warehouse 2023
MA members with A1C >8%	~2.2M	NHANES 2017–2020 + CMS
A1C >8% — inadequate control rate	23–27% of T2DM	ADA SOC 2024
A1C >9% — poor control rate	12–15% of T2DM	NHANES 2017–2020
Annual excess cost (A1C >8% vs. controlled)	$9,600/member/year	Dall et al. Diabetes Care 2022
Therapeutic inertia delay	1.9 years mean	Khunti et al. Diabetes Care 2013
MACE reduction — GLP-1 (LEADER)	−14%	Marso et al. NEJM 2016
CV death reduction — SGLT2 (EMPA-REG)	−38%	Zinman et al. NEJM 2015
3-year ROI on intensification program	3.1×	Modeled (see methodology)

---

Intended Use

Permitted Uses

• Internal payer analytics and population health program design
• HEOR research and academic analysis
• Replication and audit of published claims
• Health economics education

Not Included / Protected

The Compact Knowledge Graph (CKG) encoding structure — the typed entity graph, relationship topology, and traversal logic used to integrate these data sources — is the proprietary methodology of Graphify.md, protected under pending patent claims (§3A, 2 provisional applications filed April–May 2026). This dataset publishes the underlying source data and analytical process but not the graph structure itself.

---

Data Sources

Source	Type	Access
CMS Chronic Conditions Data Warehouse	Federal administrative data	cms.gov/Research-Statistics
NHANES 2017–2020	Federal population survey	cdc.gov/nchs/nhanes
ADA Standards of Care 2024	Clinical guideline	diabetesjournals.org
HCCI Medicare Advantage Claims Analysis	Published research	healthcostinstitute.org
Dall TM et al. Diabetes Care 2022	Peer-reviewed journal	doi.org/10.2337
Khunti K et al. Diabetes Care 2013	Peer-reviewed journal	doi:10.2337/dc13-0375
LEADER (Marso SP et al. NEJM 2016)	RCT	doi:10.1056/NEJMoa1603827
SUSTAIN-6 (Marso SP et al. NEJM 2016)	RCT	doi:10.1056/NEJMoa1607141
EMPA-REG (Zinman B et al. NEJM 2015)	RCT	doi:10.1056/NEJMoa1504720
CREDENCE (Perkovic V et al. NEJM 2019)	RCT	doi:10.1056/NEJMoa1811744
CMS Star Ratings Technical Notes 2025	Federal regulatory	cms.gov
AHIP PA Survey 2023	Industry survey	ahip.org

---

Limitations

1. Cost estimates are literature-derived, not plan-specific claims. Calibrate with your own actuarial data.
2. RRR from CVOTs are applied to MA population; real-world effect sizes may differ.
3. Rebate assumptions are estimates — net drug costs vary by plan size and negotiation.
4. Therapeutic inertia delay (1.9 years) is from a 2013 study; current practice patterns may differ.
5. Model outputs are planning estimates, not actuarial certifications.

---

Replication

See process/methodology_overview.md for the full step-by-step analytical process including:

• Claims-based population extraction logic
• A1C stratification criteria
• ROI model structure and sensitivity ranges
• HEDIS C09 linkage

---

Citation

@dataset{yarmoluk2026ma_a1c_heor,
  author = {Yarmoluk, Daniel},
  title = {MA T2DM A1C Gap — HEOR Analytics Source Data},
  year = {2026},
  publisher = {HuggingFace},
  url = {https://huggingface.co/datasets/danyarm/ma-a1c-heor-analytics},
  note = {Graphify.md §3A CKG Methodology — Patent Pending}
}

---

Contact

Daniel Yarmoluk · Founder, Graphify.md
daniel.yarmoluk@gmail.com · graphifymd.com

§3A Compact Knowledge Graph Methodology · Patent Pending (2 provisional applications) · Non-provisional deadline April 2027