PSD.17:4.5 - Perform and observe only the intervention that was chosen

Translate the decision into the actual teaching, facilitation, community, recognition or publication work needed. Keep proposed action and work planning separate from performance. A planned tutorial, a posted file and an actual guided exercise are different facts.

When performance is claimed, use the recovered A.13 and A.15.1 basis. An investigation of a proposed decision-support variant may enact an independently identified trial or teaching Method without pretending that the unadmitted candidate whole was already enacted as a U.Method.

Observe the receiving use that matters. Ask a receiving practitioner to reconstruct why a concern was retained, which premise limited a comparison, what remained outside the compared set, or why a blocker was the appropriate return. This is a possible intervention design, not a universal teaching recipe. Check whether it is suitable for the population and the professional situation.

Do not make participation in a real consequential decision an uncontrolled training experiment. Preserve the actual access, competence, confidentiality, challenge, source-use and fallback conditions. If sharing the original case would break them, use a qualified redacted or constructed case, or choose another route; acknowledge what that substitute cannot test.

Identify the actual changed object when a change is claimed. An amended example set, an altered display rule and a changed practitioner action are not the same transformation. Use A.3.4 and the direct effect or Work-to-change predicate when those claims matter. A performed intervention, an observed value and a causal-effect claim remain distinct; a log or temporal succession supplies no missing relation.