As you may recall from earlier posts, after assigning the individual evidence ratings for each of the domains for downgrading and upgrading, you then combine into one summary rating for each outcome. Here's the summary of prior posts for how to get to that overall summary rating.
- GRADE provides evidence ratings for each outcome.
- GRADE uses 5 domains to rate down your confidence in the quality of evidence derived from Randomized Controlled Trial studies (RCTs).
- GRADE starts overall evidence rating for evidence derived from RCTs at the highest level (HIGH) and then subtracts, or downgrades, for insufficiencies in that evidence. However, there are only 4 levels total, so overall evidence cannot be rated lower than VERY LOW.
- The four rating levels in the GRADE approach are High, Moderate, Low, and Very low.
- Each domain is considered equal.
- Those 5 domains for downgrading are:
- Risk of bias = limitations in the design and conduct of the studies that impact validity
- Inconsistency = lack of reproducibility of the effect across multiple studies
- Indirectness = in any of the PICO elements; if tested in population that differs from the one of interest, if difference in the intervention itself, or use of surrogate outcomes
- Imprecision = when confidence intervals around the effect estimate include both benefit and harm and impact the clinical decision threshold
- Publication bias = difficult to assess, but GRADE provides some indicators for suspicion that positive studies have been selectively published for the topic
- So, for example, if the evidence is downgraded by one level for risk of bias and one level for inconsistency it would go from an overall rating of High to Low.
- After being downgraded by 3 levels when starting at High, the evidence rating cannot be further reduced.
- GRADE starts evidence derived from observational studies at an overall rating of Low and allows for upgrading in 3 instances.
- Only well-designed and conducted observational studies are eligible for upgrading.
- The 3 instances (domains) for upgrading are:
- Large effect
- All known confounding should be working against the direction of the observed effect
- Evidence can be upgraded one or two levels for each domain.
You know where to go for more detail - the official GRADE handbook
Copyright TheEvidenceDoc December 5, 2017.