Table 1.

Pros and cons of different approaches for incorporating adverse effect data in a systematic review*

Method	Look in the trials/studies included in the systematic review of benefit.	Look in all retrieved trials/studies of that intervention, even in those excluded from the analysis of benefit	Look for studies that specifically evaluate adverse effects of the intervention
Protocol	Should usually be the minimum recommendation	Studies rejected from analysis of benefit (e.g. because beneficial outcomes are measured in a different way, which cannot be combined with other studies), may be included to allow adverse effect data collection. Two sets of inclusion criteria will be needed – for benefit, and for adverse effects	Design separate strategy to identify studies that report adverse effects, including those that do not look at beneficial effects.
			Might amount to a separate review nested within a traditional Cochrane review
Pros	Less demanding on time and resources	More comprehensive than just looking at included trials	Most comprehensive
	Does not require new literature search strategy	Can potentially cover a more representative group of patients	May be able to evaluate rare, or long-term, or previously unrecognized adverse effects
Cons	Data may be very limited and biased towards common, short-term harms	Relatively time consuming as full-text articles of all potentially relevant studies need checking Data may be limited to well-recognized and commonly seen adverse effects.	Time and resource intensive
		Benefit and harm cannot be compared directly as the data come from different sources	Special techniques required in synthesizing data from a diverse range of sources
			Increased quantity of data but greater risk of biased and poor quality data
			Benefit and harm cannot be compared directly as the data come from different sources.

*Copied from reference [18].