Levels of Evidence

To help you to make more informed decisions, from Autumn 2014 on the Hot topics course we will grade the level of the evidence (from research studies) and the strength of recommendations (from guidelines) wherever possible. 

Grading of evidence

This is an internationally recognized ranking system used in evidence based medicine to rank the quality of the evidence and therefore the strength of the conclusion you can take from it.

Evidence level	Evidence is from:
1a	A systematic review & meta-analysis of randomized controlled trials
Ib	At least one randomized controlled trial
11a	At least one well-designed study, but not randomized e.g. prospective cohort study
11b	At least one well designed experimental, but not controlled, trial
111	Well designed descriptive studies e.g. case-control, comparative studies and case series
1V	A panel of experts

Grading of recommendations

Guidelines then use a grading hierarchy to grade the strength of the recommendations they make. Confusingly, there are several different models in use but the most accepted one (which we have tried to follow) is the GREG (Guideline Recommendation and Evidence Grading) system:

Grade of recommendation	Based on:
A (Recommendation)	There is robust evidence to recommend a pattern of care (based on level 1 evidence)
B (Provisional recommendation)	On balance of evidence, a pattern of care is recommended with caution (usually based on hierarchy 11 evidence)
C (Consensus opinion)	On absence of directly applicable studies of sufficient quality, a pattern of care is recommended by consensus

For simple guidance on Evidence Based Medicine (e.g. how to tell your ARR from your RRR, different study types, how to calculate a NNT and what it means, how to interpret confidence intervals etc) please download our free KISS Guide to Evidence Based Medicine.

 For more on evidence levels, there is an excellent PatientPlus summary and see also the Centre for Evidence Based Medicine