B.1.3:5.1 - Episteme — Heterogeneous evidence into a guidance statement

This is a didactic evidence-composition case, not a clinical recommendation. The receiving question concerns a treatment’s pain outcome for an identified acute low-back-pain population within six weeks. The exact population and outcome, not a label or taxonomy operation, identify the subject.

• Inputs: E₁ is a randomized trial, E₂ an observational study, and E₃ a mechanistic model. Retain their designs, findings, uncertainty, assumptions, and scope. A bare “R=0.84/0.55/0.60” list has no declared common quantity here and cannot support a numerical aggregate. Likewise, “medium/wide/narrow” is not a substitute for actual ClaimScope.
• Synthesis: retain protocols, datasets, analysis scripts, and other participating carriers in the SCR. Establish any dosage or outcome mapping from its actual measurement basis; units named mg and IU do not themselves establish a conversion. Keep chronic cohorts and different outcomes separate unless a warranted transport relation supports the receiving claim.
• Necessary-premise case: if the proposed inference requires an outcome mapping that is unsupported, withhold that mapped inference. Retain the trial’s narrower source-scale conclusion rather than discard every source.
• Complementary case: a limited E₂ may help assess a confounding explanation left open by E₁; E₃ may constrain a mechanism without establishing the clinical effect size. Explain those contributions and their limits. The minimum of three unexplained scores neither captures them nor defeats the existing trial conclusion.
• Dependence case: if two reports reuse a cohort or share an outcome-measurement bias, do not count them as two independent confirmations. Different designs alone are insufficient to establish different biases.
• Contrary-result case: a credible E₂ result conflicts with E₁ in an overlapping subgroup. Keep the disagreement. Different baseline severity is a possible explanation only to the extent supported by the sources. Narrow or qualify the guidance claim; do not silently average the conflict away or call it an uninformative study.
• Completion: return the bounded guidance statement with its distinct supporting contributions, contrary result, scope, and unresolved interpretation. No common quantitative model has been supplied, so no aggregate R is returned. Whether another study is feasible and worth its total burden is a separate C.11/C.19.2 decision, not a condition for completing this synthesis.

Concrete completion of the case. Use the following invented findings only to illustrate the synthesis. P is the trial’s enrolled acute low-back-pain cohort; S is its identified high-baseline-severity subgroup. E₁ reports lower pain at six weeks with intervention X than with its comparator, both in P and in S. E₂ reports no such benefit in a separately observed cohort matching S; the sources have not resolved whether confounding or another difference explains the disagreement. E₃ supplies a mechanism compatible with benefit but no clinical effect estimate. No common quantitative support model is available.

The resulting guidance statement is: “The trial supports lower recorded pain at six weeks for X versus its comparator in P. The finding for S conflicts with the observational result and remains unresolved. The mechanism explains how benefit could occur but does not resolve that conflict or establish its magnitude. These sources support neither extrapolation beyond P and six weeks nor a combined reliability score.” The SCR retains all three contributions. A supported explanation of the disagreement or a corrected source result reopens the affected claim. A request for another population opens a separate question and synthesis. Identify the new claim, its subject and scope (C.2.1; A.2.6), and establish the evidence or warranted transport needed for that use. The request alone leaves the result for P unchanged. This is a completed bounded synthesis of the fictional findings, not advice to use X.

For Γ_epist^compile, map the retained claims into the journal’s scheme, carry the same limitations and any justified recalculation, and produce the compilation SCR and required hashes. C.2.1 identifies the target-scheme episteme “Guidance Statement v1.0”; later journal publication remains a separate occurrence.