Table 3.

Conclusions not supported by the evidence

Reviewers' conclusion	Assessors' comment
“Studies are of insufficient duration to identify a reduction in mortality” w18	“The studies . . . suggest an increased mortality (13 v 7)”
“X is the cornerstone of treatment” w51	“The comparisons with Y and Z are not that convincing; see graphs”
Treatment X might promote healingw20	“The study that suggested this . . . may be flawed in design, conduct, and analysis”
Treatment X “is associated with a substantially reduced risk” w11	“Too strong . . . as the confidence interval crossed 1”
“This review found a significant decrease in . . .” (one variable)w31	“Many small trials. . .with many outcome variables . . . implying a high risk of reporting bias”
Treatment X “is associated with a reduction in death or oxygen requirement” w53	“Significant reduction was seen only for the combined outcome, not for death or reduction in oxygen requirement separately”
“There is evidence to support the early use of X in disease Y” w47	“This conclusion . . .is unwarranted and misleading” as “the review might concern disease Z” (which is known to respond to treatment X)
“X is associated with short term improvements” w29	“Given the limitations of the included studies we would recommend that the conclusions be modified to be more cautious”
“There are too few data to draw any reliable conclusions” w50	“Potential adverse effect is so important that it should be mentioned in the conclusions”