Public reference · Evidence map
Illustrative public reference—not client work or a results claim.
Healthspan measures and endpoints: an evidence map for decisions.
“Healthspan evidence” is not one category. This map separates six layers so a concept, program milestone, measurement result, trial result, regulatory status, or adoption signal is not mistaken for another.
How to use the map
Locate the evidence before judging its strength.
- Name the decision.Screening, trial design, product claim, regulatory submission, coverage, procurement, or care delivery require different evidence.
- Place each source in one layer.Do not let award, publication, or program status stand in for the next layer.
- Record the gap.Identify the evidence layer missing between what is known and what the decision requires.
- Set the next source check.Define the protocol, dataset, analysis, decision, or implementation evidence that could close the gap.
| Layer | Core question | Evidence that belongs here | What it cannot establish alone | Next decision |
|---|---|---|---|---|
| 1. Concept and construct | What aspect of health, function, resilience, capability, or lived experience matters—and to whom? | Conceptual model; patient and clinician input; population; setting; meaningful aspect of health; proposed measure. | That the concept can be measured accurately or changed by an intervention. | Is the target concept sufficiently defined to measure? |
| 2. Measurement performance | Does the method produce reliable, valid, interpretable data under intended conditions? | Verification; analytical validation; clinical validation; usability; data provenance; missingness and error analyses. | That a valid measure is an accepted endpoint or that changing it improves health. | Is the measure fit for the defined context of use? |
| 3. Endpoint validity | Can the measure support the specific inference and decision assigned to the endpoint? | Endpoint definition; estimand; timing; meaningful change; longitudinal properties; multiplicity; statistical analysis plan. | That an intervention will achieve a favorable effect. | Can the endpoint be used in the proposed study and interpretation? |
| 4. Intervention effect | Does an intervention improve the prespecified endpoint relative to an appropriate comparator, with acceptable harms? | Registered protocol; randomized or otherwise fit-for-purpose design; effect estimate; uncertainty; follow-up; harms; attrition; subgroup analyses. | Regulatory authorization, coverage, generalizability, or real-world feasibility. | Does the result support a bounded claim and next-stage evaluation? |
| 5. Regulatory, policy, and payment status | What use, claim, coverage, payment, or pathway has an accountable authority actually established? | Authorizations; labeling; guidance; coverage decisions; model rules; contracts; dated official records. | That an intervention is appropriate for every person, organization, or setting. | What lawful, funded, and governable use is available now? |
| 6. Implementation and real-world learning | Can the intervention be delivered, accessed, trusted, sustained, and monitored in practice? | Workflow; training; adoption; adherence; access; equity; cost; safety monitoring; real-world outcomes; failure and stop criteria. | That observed associations are causal or that one implementation generalizes. | Continue, adapt, pause, stop, or scale—with what monitoring? |
Current examples
Three source objects, three different evidentiary roles.
FDA digitally derived measures paper
Supports questions about concept selection, verification, validation, usability, error, and evidence proportionate to intended use. It does not validate a specific measure or endpoint.
Inspect the source record →ARPA-H PROSPR
Establishes a funded healthspan research portfolio and intended work on measures, trial methods, and interventions. Program activity is not favorable endpoint or treatment evidence.
Inspect the PROSPR object →XPRIZE Healthspan finalists
Establishes 20-finalist status and separately identifies 10 of those teams as $1 million Milestone 2 Awardees, plus announced trial plans. Neither status establishes safety, efficacy, endpoint validity, regulatory status, or generalizability.
Inspect the source record →Minimum evidence-map row
For each claim or decision, record: source and date · evidence layer · population and setting · measure or intervention · comparator · endpoint · follow-up · effect and uncertainty · harms · missingness · bias · applicability · funding/conflicts · current decision use · next source check.