Table 3.

Decreasing and increasing the grade of evidence

Decrease grade if*	Serious or very serious limitation to study quality Important inconsistencies in results Some or major uncertainty about directness of the evidence Imprecise or sparse data (relatively few participants and/or events) High probability of reporting bias
Increase grade if	Magnitude of the treatment effect is very large and consistent Evidence of a large dose-response relation All plausible confounders/biases would have decreased the magnitude of an apparent treatment effect

*Each quality criterion can reduce the quality by one or, if very serious, two levels. See BTS GRADE guideline production manual for further details (https://www.brit-thoracic.org.uk/quality-improvement/guidelines/).

BTS, British Thoracic Society.