Grade of Recommendation

The strength of a recommendation — in GRADE, strong vs weak (conditional); in older systems, the A–D letter grades (SIGN) or numbered levels (Oxford CEBM). Strength reflects how confidently a clinician should act on it. It is driven by the certainty of the evidence plus the balance of benefits and harms, patients’ values and preferences, and resource use — so it is distinct from the level/certainty of the evidence itself.

The pyramid (the level/certainty of evidence) is just one input. To set a recommendation’s strength, a panel also weighs how the benefits stack up against the harms, what patients value, and the cost. The output — strong or weak/conditional — sits on a different axis from the pyramid tier.

A strong recommendation means “almost all well-informed patients would want this — just do it.” A weak/conditional one means “it depends — share the decision.” Read the grade as a guide to action under uncertainty, not as a badge for the evidence: the two can diverge in either direction.

In the ED you act on recommendations, not raw papers. Knowing strength is separate from evidence quality stops you over-trusting a confident-sounding “Grade A” and stops you dismissing a strong recommendation just because the evidence behind it is thin — sometimes a large, obvious benefit justifies acting on little data.

• GRADE strength: strong vs weak / conditional
• Determined by: certainty + benefit–harm balance + values + resources
• Older grades (SIGN A–D, Oxford levels) → superseded by GRADE
• Strength of recommendation ≠ level / certainty of evidence

A strong recommendation on low-certainty evidence — GRADE methodologists give whole-body CT/MRI cancer screening as the textbook case. The benefit (fewer cancer deaths) is very uncertain, but the harms are near-certain: false positives, anxiety, and invasive follow-up tests with their own complications. So a panel can legitimately make a strong recommendation against screening despite low-certainty evidence on benefit (Guyatt et al., J Clin Epidemiol 2013; GRADE guideline 15). The strong grade here is driven by the benefit–harm balance, not by the (low) certainty of the evidence — the exact dissociation examiners test.

• Confusing strength of recommendation with level / certainty of evidence.
• Ignoring that values, preferences and resource use also drive the grade.
• Relying on outdated letter grades (SIGN A–D / Oxford levels) instead of GRADE strong–weak.