Strong accreditation evidence is authentic, traceable, current and demonstrative of implementation. This note provides a CHSPS evidence hierarchy, sufficiency test, sampling logic and reviewer checklist designed to make evidence review more consistent without turning document count into a score.
1. The CHSPS evidence-quality model
Reviewers should ask two separate questions: Is the evidence authentic? and Is the evidence sufficient to support the requirement? A policy can be authentic but insufficient if it does not show implementation. A screenshot can be current but weak if it cannot be traced to the responsible process or source record.
2. Evidence hierarchy
| Evidence class | What it can establish | Typical examples | Common weakness |
|---|---|---|---|
| E1 · Controlled documentation | Design and stated responsibility | Policy, SOP, committee charter, plan | May show intent only |
| E2 · Implementation records | That the process occurred | Minutes, logs, approvals, completed forms | Single record may not show consistency |
| E3 · Demonstration / system output | How controls work operationally | System demonstration, access record, report | Can be staged or unrepresentative |
| E4 · Interview / observation | How people understand and execute the process | Staff interview, observed workflow | Memory/opinion alone is insufficient |
| E5 · Performance data | Whether the process is producing monitored results | Trend report, KPI, audit result, CAPA closure | Definitions or source lineage may be unclear |
3. Sufficiency test
For a material requirement, reviewers should generally seek a chain of evidence: controlled requirement → accountable owner → implementation record → sample or observation → monitoring/corrective action where applicable. Not every clause needs all five evidence classes, but higher-risk claims should require stronger corroboration.
An organization submits its incident-response policy. That establishes design. A stronger file adds a six-month incident log, two sampled investigation records, reviewer training records, a monthly security report and evidence that a recurring issue generated corrective action.
4. Sampling approach
CHSPS does not prescribe a universal sample size in this note. Sample size should reflect risk, volume, variability, previous findings and the consequence of failure. A reviewer may expand a sample when records are inconsistent, when exceptions cluster in one location/vendor, or when the initial sample identifies a control failure.
5. Evidence sufficiency labels
| Label | Meaning | Reviewer action |
|---|---|---|
| Supported | Evidence reasonably demonstrates both design and implementation. | Document rationale and proceed. |
| Partially supported | Some evidence exists, but one or more material elements remain unproven. | Request targeted clarification or additional evidence. |
| Unsupported | Evidence does not substantiate the requirement. | Evaluate against the applicable standard and finding rules. |
| Not applicable | The requirement is demonstrably outside the approved scope. | Document the applicability rationale. |
6. Reviewer checklist
- Identify the exact clause and the claim that must be demonstrated.
- Confirm evidence ownership, date/version and source.
- Look for proof of implementation, not only policy language.
- Sample enough records to test consistency and exceptions.
- Reconcile contradictions across documents, interviews and data.
- Record why the evidence is sufficient or why more is needed.
CHSPS references
Accreditation requirements are stated exclusively in the published, version-controlled CHSPS standards. This note explains method, risk and good practice; it creates no requirement and changes none.
- CHSPS Standards LibraryThe controlled Edition 2026.1 standards for all 16 active CHSPS program codes.
- CHSPS Standards InterpretationsFormal interpretations issued under document control when a requirement needs clarification.
- CHSPS Technical ReportsPublished CHSPS methods and analyses supporting evidence quality, assessment integrity and performance measurement.