A three-axis classification model applied to 5,129,584 Japanese adults from a national health insurance database reframes hypertension beyond a simple binary diagnosis. By simultaneously coding diagnosis status, treatment intensity, and blood pressure control into 27 possible states — condensed into seven clinically meaningful groups — the framework revealed that 36% of the cohort fell into hypertension-related categories. Critically, 11% had unrecognized hypertension and 7% were diagnosed but entirely untreated. The model showed strong internal consistency, with hierarchical cluster analysis agreement reaching a weighted kappa of 0.87, and diagnosis validity supported by 96.5% sensitivity and 91.8% specificity against antihypertensive medication records.
The finding that nearly one in five adults sits in a hypertension care gap — either undiagnosed or diagnosed but unmanaged — is clinically significant and aligns with patterns seen in global hypertension surveillance data, including the Global Burden of Disease estimates. Current single-axis definitions routinely obscure this heterogeneity, limiting both research precision and public health targeting. The three-axis approach offers a structured way to prioritize interventions: unrecognized patients need screening, while untreated diagnosed patients need engagement strategies. Limitations include the observational nature of claims data, potential coding inaccuracies, and the exclusively Japanese cohort, which may limit generalizability to populations with different healthcare access patterns. As a preprint not yet peer-reviewed, the model's clinical adoption should await independent validation. That said, this framework represents a genuinely useful — if methodologically incremental — advance in hypertension phenotyping at population scale.
