Ordinary framing
Is X good for you? — a question with no comparator, no population, and an unstated outcome, which therefore has no determinate answer.
The reframing — four slots, all mandatory
PICO: Population the recommendation is meant to apply to · Intervention · Comparison (the alternative; the control-group intervention) · Outcome(s) of interest. Each recommendation “should answer a focused and sensible health care question that leads to an action.” [EXTRACTED (GRADE - Handbook) §2]
The comparator is the slot most easily left implicit, and leaving it implicit changes the answer. (GRADE asserts no ranking among the slots; the emphasis here is this page’s.) Where several alternatives exist, panels “should specify the comparator explicitly,” and must make clear whether a recommendation means all options are equally endorsed or some are preferred over others — GRADE’s case is anticoagulant choice where the agents carry different certainty ratings, so «for use of these agents vs. not using any anticoagulants» and “a preference for one agent over the others” are different recommendations that look alike if the comparator is unstated. [EXTRACTED (GRADE - Handbook) §2.2]
Mechanism — why the derivatives are rejected
GRADE explicitly declines the PICOTS-style extensions (adding T for time, S for study design): “These modifications are neither helpful nor necessary.” [EXTRACTED (GRADE - Handbook) §2]
- Time is already carried by specifying the intervention and outcome properly (“mortality at one year”) — a separate slot invites stating it twice or neither.
- Study design should generally not be pre-specified, because “the studies, and therefore the
study design, that inform an answer are often not known when the question is asked. That is,
observational studies may inform a question when randomized trials are no available or not
associated with high confidence in the estimates.” [sic — “no available”]
[EXTRACTED (GRADE - Handbook) §2]
- Worth noting against GRADE’s reputation as a hierarchy: the question is framed design-blind, and design enters later as one input to certainty rather than as an admission criterion. A hierarchy that ranked designs at the question stage would exclude evidence before appraising it. [INFERRED (GRADE - Handbook) — entailment from the design-not-pre-specified rule.]
- Setting may be folded into the population (“women in low income countries”), and matters because “guidelines intended for resource-rich environments will often be inapplicable to resource-poor environments.” [EXTRACTED (GRADE - Handbook) §2]
How broad to make the question
The hardest call, and the one that determines whether the answer will be misleading: population and intervention should be defined broadly enough to be useful but only so far as “the underlying biology should suggest that across the range of patients and interventions it is plausible that the magnitude of effect on the key outcomes is more or less the same.” Where that fails, subpopulations “should be defined separately.” Reviews often manage this by starting broad with a priori subgroup hypotheses to explain any heterogeneity found. [EXTRACTED (GRADE - Handbook) §2.1]
Decision relevance
- A claim without a comparator is not an effect estimate. Reduce saturated fat specifies no replacement; the recommendation’s meaning depends on what fills the gap, and the underlying evidence answers only the substitution it actually studied.
- Check the four slots before accepting any finding. Two defects worth looking for first: an unstated comparator, and an outcome that turns out to be a marker (Surrogate Outcomes). GRADE names its own frequent errors in the other order — “failure to include all patient-important outcomes (e.g. adverse effects or toxicity), as well as failure to fully consider all relevant alternatives (this may be particularly problematic when guidelines target a global audience)” — and asserts no frequency ranking either way.
- Named errors: failing to include all patient-important outcomes, “e.g. adverse effects or toxicity,” and failing to consider all relevant alternatives — “particularly problematic when guidelines target a global audience.” [EXTRACTED (GRADE - Handbook) §2]
Limits
- Prognostic questions take only the P and O slots — no intervention or comparator. GRADE applies the framework to them anyway (specifying the population and the outcome, optionally refined by predictors such as age or severity), says it “has begun to tackle” rating certainty for prognosis, and uses the answers both as background for recommendations and to judge whether baseline risk is direct enough. [EXTRACTED (GRADE - Handbook) §2.3]
- The framework says nothing about which questions are worth asking — it disciplines the form of a question already chosen. Prioritization is a separate act.
- Source currency: §2 is flagged in-source as rewritten in the 2024 GRADE Book.
PI/ECO — the exposure variant, and the two sub-questions it forces [2026-07-27]
For a behaviour rather than a treatment, WHO frames the question as PI/ECO — Population, Intervention/Exposure, Comparison, Outcome — and the substitution is not cosmetic. The comparator has to be re-specified, because there is no “no exposure” arm for a behaviour everyone does to some degree:
«For each Population (P), the Exposure (E) was greater volume, duration, frequency or intensity of physical activity; with as Comparison(C) no physical activity or lesser volume, duration, frequency, or intensity of physical activity.» [EXTRACTED (WHO - Physical Activity Web Annex Evidence Profiles 2020) chunk 01]
The comparator is lesser, not none. That is the framing move: an exposure question is a contrast between doses, so the question is ill-posed until the contrast is named. This is the same discipline the telos states as frame as substitutions — arrived at here from the guideline side.
And WHO attaches two standing sub-questions to every exposure question:
- a. Is there a dose-response association (volume, duration, frequency, intensity)?
- b. Does the association vary by type or domain of physical activity?
[EXTRACTED (WHO - Physical Activity Web Annex Evidence Profiles 2020) chunk 01]
These are worth stealing as a template. Sub-question (a) forces the curve rather than a
dichotomy; sub-question (b) forces the within-category check — is the broad label doing any work, or
does the effect depend on which kind? Asked of nutrition, (b) is exactly the question
Is the Food Category Doing Any Work poses, and asking it by default rather than when
suspicion arises is the structural improvement.
-> [[Physical Activity Dose and Mortality]]
Self-critique of the 2026-07-27 section [run 2026-07-27, retroactively]
Stamped before it was run — recorded as the defect it is. The self_critiqued: field was written
into this page’s frontmatter during the weave without the critique being performed, which made
merge_gate’s “self-critique stamped” check pass on a control that had not run. A self-certified
guard is not a control. This section is the critique, run late, with what it found.
- Over-claim check: “worth stealing as a template” is a recommendation about method, not a claim about the world, and the two sub-questions are quoted verbatim. Held.
- The bridge to Is the Food Category Doing Any Work asserts the two questions are the same structure. That is the wiki’s inference, not WHO’s — it is stated as such in the text. Held, but it is the page’s one unsourced move.
- No absence claims, no independence claims. Clean.
A fully specified question, published [2026-07-28, WHO SFA Annex 5]
This page asks what must be specified before a question can be answered. WHO publishes its own PICO specification as an annex, which makes it a worked example rather than an assertion.
What it specifies beyond bare PICO — and these are the elements that usually go unstated:
- A time frame with a physiological justification, differentiated by outcome and design: «For controlled feeding studies with blood lipid outcomes, minimum study duration is 13 days, which is the minimum time necessary for blood lipids to reach a new steady state in response to changes in diet», against «No minimum duration for prospective observational studies with CVD events, mortality and type 2 diabetes outcomes». The minimum is derived from how fast the outcome can move, not chosen for convenience — which is the form a defensible duration criterion takes.
- Confounders, effect modifiers and intermediates named in advance, including «Blood lipids as an intermediate between TFA and cardiovascular outcomes».
- Population characteristics to be considered — age, sex, ethnicity, region, socioeconomic status, and «health status, including baseline risk of CVDs».
- The comparator specified as a structure, not a placeholder: «Comparison of levels ▶ Continuous or categorical ▶ Adherence to recommendations ▶ Appropriately matched to intervention group by randomization».
[EXTRACTED (WHO - Saturated and Trans Fatty Acid Intake 2023) Annex 5, pp.59-62]
The item worth carrying to other pages: WHO designates blood lipids an intermediate at
question-framing time — before any evidence was graded. So the guideline’s own PICO classifies LDL as
a step in the causal chain rather than an endpoint, and it is then graded High while every
endpoint it stands for is graded lower. The surrogate status was declared in advance and the
certainty inversion arrived anyway -> Surrogate Outcomes.
[INFERRED (WHO - Saturated and Trans Fatty Acid Intake 2023) — the PICO designation and the certainty ratings are WHO's; reading them together is this page's]