Summary of findings for the main comparison. Active versus expectant management of the third stage of labour (all women).

Active versus expectant management of the third stage of labour (all women)
Population: all women who expected a vaginal birth at 24 weeks' gestation or later and their babies Setting: UK and Ireland, hospital setting. The countries were classified as 'higher‐income' and 'lower‐income', with the border between lower‐middle‐income and upper‐middle‐income being the cut‐off. All studies included in this main analysis were undertaken in higher‐income countries (defined according to World Bank definitions 2018). Intervention: active management of the third stage of labour Comparison: expectant management of the third stage of labour
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Quality of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Expectant management of the third stage of labour	Active management of the third stage of labour
Severe primary PPH (clinically estimated or measured blood loss ≥ 1000 mL at time of birth)	24 per 1000	8 per 1000 (3 to 21)	RR 0.34 (0.14 to 0.87)	4636 (3 studies)	⊕⊝⊝⊝ Very lowa
Very severe primary PPH (clinically estimated or measured blood loss ≥ 2500 mL at time of birth)	See comment	See comment	Not estimable	0 (0 studies)	See comment	No data
Maternal mortality	See comment	See comment	Not estimable	0 (0 studies)	See comment	No data
Maternal Hb < 9 g/dL 24‐72 hours postpartum	71 per 1000	36 per 1000 (21 to 59)	RR 0.50 (0.3 to 0.83)	1572 (2 studies)	⊕⊕⊝⊝ Lowb
Admission to SCBU/NICU	52 per 1000	42 per 1000 (31 to 58)	RR 0.81 (0.60 to 1.11)	3207 (2 studies)	⊕⊕⊝⊝ Lowc
Neonatal jaundice requiring phototherapy or exchange transfusion	49 per 1000	47 per 1000 (27 to 83)	RR 0.96 (0.55 to 1.68)	3142 (2 studies)	⊕⊝⊝⊝ Very lowd
Neonatal polycythaemia treated with dilutional exchange transfusion	See comment	See comment	Not estimable	0 (0 studies)	See comment	No data
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; Hb: haemoglobin; NICU: neonatal intensive care unit; PPH: primary postpartum haemorrhage; RR: risk ratio; SCBU: special care baby unit
GRADE Working Group grades of evidence High quality: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate quality: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low quality: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low quality: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aRisk of bias: of the three studies providing data for this outcome, all are at low risk of bias for sequence generation (selection bias) and allocation concealment (selection bias). All are at high risk of bias for lack of blinding for clinicians and women and all are unclear for blinding of outcome assessment. Two studies are at low risk of bias for incomplete outcome data (attrition bias), and one is unclear. Two are at high risk of selective reporting (reporting bias) and one is unclear. One study is at high risk of other bias and two are unclear (see 'Risk of bias' tables and Figure 2). Downgraded 1.
 Inconsistency: there is some overlap of confidence intervals of the three studies, however, Tau² = 0.38, the P value for the Chi² test of heterogeneity is 0.08 and I² = 60%. These suggest the presence of heterogeneity, which cannot be explained by any of the subgroups or sensitivity analyses performed. Downgraded 1.
 Indirectness: directly answers the question.
 Imprecision: total (cumulative) sample size 4636 is less than the optimal information size of 18,590 (assuming α = 0.05, 1‐β = 0.80, relative risk reduction (RRR) of 25% from control event rate). Events = 88, Downgraded 1.
 Publication bias: assessment of funnel plot asymmetry not performed due to fewer than 10 studies included for this outcome.

^bRisk of bias: of the two studies providing data for this outcome, both are at low risk of bias for sequence generation (selection bias) and one is at low risk of bias for allocation concealment (selection bias) and the other is unclear. Both are at high risk of bias for lack of blinding of clinicians and women, both are unclear for blinding of outcome assessment (although unlikely to affect Hb measurements). One is at high risk of bias for incomplete outcome data (attrition bias) while the other is unclear. One is high risk of bias for selective reporting (reporting bias), the other is unclear. Both studies are at high risk of other bias (see 'Risk of bias' tables and Figure 2). Downgraded 1.
 Inconsistency: the confidence intervals of the two studies overlap. Tau² = 0.02, the P value for the Chi² test of heterogeneity is 0.31 and I² = 3%. Although Tau² is non‐zero, tests suggest an absence of unexplained heterogeneity.
 Indirectness: directly answers the question.
 Imprecision: total (cumulative) sample size 1572 is less than the optimal information size of 5804 (assuming α = 0.05, 1‐β = 0.80, RRR of 25% from control event rate). Events = 94. Downgraded 1.
 Publication bias: assessment of funnel plot asymmetry not performed due to fewer than 10 studies included for this outcome.

^cRisk of bias: of the two studies providing data for this outcome, both are at low risk of bias for sequence generation (selection bias) and allocation concealment (selection bias). Both are at high risk of bias for lack of blinding of clinicians and women, both are unclear for blinding of outcome assessment. One is at low risk of bias for incomplete outcome data (attrition bias), the other is unclear. Both are at high risk of bias for selective reporting (reporting bias) and both are at high risk for other biases (see 'Risk of bias' tables and Figure 2). Downgraded 1.
 Inconsistency: the confidence intervals of the two studies overlap. Heterogeneity: Tau² = 0.00; Chi² P value = 0.40, I² = 0%. This suggests an absence of unexplained heterogeneity.
 Indirectness: directly answers the question.
 Imprecision: total (cumulative) sample size 3207 is less than the optimal information size of 8066 (assuming α = 0.05, 1‐β = 0.80, RRR of 25% from control event rate). Events = 152. Downgrade 1
 Publication bias: assessment of funnel plot asymmetry not performed due to fewer than 10 studies included for this outcome.

^dRisk of bias: of the two studies providing data for this outcome, both are at low risk of bias for sequence generation (selection bias) and allocation concealment (selection bias). Both are at high risk of bias for lack of blinding of clinicians and women, both are unclear for blinding of outcome assessment. One is low risk of bias for incomplete outcome data (attrition bias), the other is unclear. Both are at high risk of bias for selective reporting (reporting bias) and both are at high risk of other biases (see 'Risk of bias' tables and Figure 2). Downgraded 1.
 Inconsistency: there is some overlap of confidence intervals of the two studies. However, Tau² = 0.11, P value for heterogeneity = 0.09 and I² = 66%, which suggest the presence of heterogeneity that cannot be explained by any of the subgroups or sensitivity analyses performed. Downgraded 1.
 Indirectness: directly answers the question.
 Imprecision: total (cumulative) sample size 3142 is less than the optimal information size of 8584 (assuming α = 0.05, 1‐β = 0.80, RRR of 25% from control event rate). Events = 149. Downgraded 1.
 Publication bias: assessment of funnel plot asymmetry not performed due to fewer than 10 studies included for this outcome.