Good Healthspan Practice — Guide Text 健康管理良好实践规范 · 指南文本（总纲摘要）

This guide applies to service practice in longevity-medicine and health-management institutions — the full cycle of inquiry, measurement, interpretation, intervention and review, or any single module. It does not apply to clinical diagnosis-and-treatment guidelines (the province of medical societies), or to pharmaceutical production and clinical trials (the province of GMP / GCP / GLP / GVP). GHP is a good-practice guide and community consensus — not a standard, not a certification, not an organization; the criteria below are for institutions' self-assessment, and the secretariat only registers self-declarations, performing no conformity assessment.

Inquiry must establish a causal chain from phenotype to root cause (four information layers: exposure → root cause → function → phenotype), not mere symptom registration. Functional-medicine and traditional-Chinese-medicine history-taking are integrated side by side, delivered through structured tools so that competent practice becomes accessible to every practitioner.

Testing is integrated multi-omics with at least one molecular-level aging anchor; same-assay, alignable, re-testable, and data-compliant.

Interpretation places every abnormality in a phenotype–function–root-cause network. Root-cause positioning is the core hard requirement of this initiative: every interpretation is by nature an inference, and the causal chain must therefore extend to the root-cause layer (first candidate: hallmarks of aging) — stopping at functional imbalance does not qualify. Root-cause inference is mandatory (any method); hallmark measurement is a bonus, never enforced — the guide mandates no specific test.

Interventions target root-cause and functional layers first, symptomatic relief second: root first, goal-driven, attributable (pre-registered), executable.

After re-testing, the institution must answer "which interventions worked" — contribution decomposition + confidence grading + honest reporting; findings roll into the next cycle.

Individualized evidence generation for the world-model era: mechanistic priors + individual measurement + N-of-1 validation + contribution review. RCTs answer average effects in homogeneous populations and cannot fit heterogeneous, multi-morbid individuals — but arbitrary inference is equally unacceptable: every individualized causal claim must be registrable, falsifiable and auditable (in the spirit of FDA N-of-1 and real-world evidence). This guide governs health-management service documentation and constitutes no medical-treatment regulation.