Skip to main content

Table 1 The features of evidence grading captured in a GRADE plot (adapted from Evid Based Med 2011;16:65-9)

From: Graphical displays for effective reporting of evidence quality tables in research syntheses

Grade

Design

Risk of bias

Inconsistency

Indirectness

Effect size

Evidence Quality

 

Studies are either described as randomised-control trials (RCTs) or observational.

Explains the limitations of the study based on assessment of blinding and allocation process, follow-up and withdrawals, scarcity of data, other methodological concerns e.g. incomplete reporting, subjective outcomes.

Inconsistencies due to unexplained (statistical) heterogeneity. The same weakness is only downgraded once.

Presence of indirectness in the PICO elements that affect the generalisability of participants and outcomes from each study to population of interest.

Relates to imprecision of the estimated effect based on the reported odds ratios or relative risks or mean differences for comparison. This is based on the confidence intervals, sample size and number of events.

 

High

Randomised controlled trial

No problems

All/most studies show similar results with or inconsistency across studies is explained by a dose response

Population and outcomes broadly generalisable

Effect size more than 5 or less than 0.2 for all studies/meta-analyses included in comparison and significant

 

Moderate

  

Lack of agreement between studies (e.g. statistical heterogeneity between RCTs, conflicting results)

 

Effect size more than 2 or less than 0.5 for all studies/meta-analyses included in comparison and significant

 

Low/Very low

Controlled observational study

Problem with 2 or more elements

Serious lack of agreement between studies

Some problem with 2 or more elements

Not all effect sizes more than 2 or less than 0.5 and significant; or if effects observed not significant

 

Examplea: Thiamine vs Placebo for Pelvic Pain

Randomised trial

No limitations

Consistent

Indirect

Precise

Moderate

 

Initially assigned a high strength level

→ No Change

→ No change

→ Relegation

→ No change

┘

  1. abased on evidence profile shown in Fig. 1 and BMJ 2012;344:e3011 doi:10.1136/bmj.e3011