Summary of findings 2. Ultrasound for diagnosis of birth weight discordance in twin pregnancies at 25% cut‐off.

Sensitivity	0.46 (95% CI 0.26 to 0.66)		Prevalencesa	19%			9%			27%
Specificity	0.93 (95% CI 0.89 to 0.96)
Outcome	No. of studies and participants	Study design	Factors that may decrease certainty					Effect per 1000 women tested
			Risk of bias	Indirectness	Inconsistency	Imprecision	Publication bias	Pretest probability of 19%	Pretest probability of 9%	Pretest probability of 27%	Test accuracy (certainty of the evidence)
True positives (women with diagnosis of birth weight discordance)	18 studies, 1679 participants	Cohort‐type studies with delayed verification (cohort type accuracy study)	Very seriousb	Not serious	Very seriousc	Not serious	None	87 (49 to 125)	41 (23 to 59)	124 (70 to 178)	⊕⊝⊝⊝ Very lowd
False negatives (women incorrectly classified as not having diagnosis of birth weight discordance)								103 (65 to 141)	49 (31 to 67)	146 (92 to 200)
True negatives (women without diagnosis of birth weight discordance)	18 studies 4792 participants	Cohort‐type studies with delayed verification (cohort type accuracy study)	Very seriousb	Not serious	Very seriousc	Not serious	None	753 (721 to 778)	846 (810 to 874)	679 (650 to 701)	⊕⊝⊝⊝ Very lowd
False positives (women incorrectly classified as having diagnosis of birth weight discordance)								57 (32 to 89)	64 (36 to 100)	51 (29 to 80)

CI: confidence interval.
^aThe prevalence used to represent the pretest probability are the median, first quartile and third quartile of the prevalences of included studies.
^bAt least 50% of the studies had unclear statements regarding index test, use of proper reference standard and flow and timing elements.
^cVery high unexplained heterogeneity in terms of sensitivity ranging from 0.1 to 1.00.
^dGRADE certainty of evidence downgraded one level for risk of bias and two levels for inconsistency.