Childbirth training workshop

(Iran)

(Valiani 2014, randomised trial)

73 per 100

40 per 100

(24 to 65)

Mothers alone versus control:

RR 0.55

(0.33 to 0.89)

(1 study, 60 women)

㊉㊉㊀㊀

LOW^a,b

73 per 100

43 per 100

(27 to 69)

Couple versus control:

RR 0.59

(0.37 to 0.94)

(1 study, 60 women)

27 per 100

61 per 100

(31 to 118)

Mothers alone versus control:

RR 2.25

(1.16 to 4.36)

(1 study, 60 women)

27 per 100

58 per 100

(29 to 112)

Couple versus control:

RR 2.13

(1.09 to 4.16)

(1 study, 60 women)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Nurse‐led applied relaxation training programme

(Iran)

(Bastani 2006, randomised trial)

404 per 1000

89 per 1000

(44 to 174)

RR 0.22

(0.11 to 0.43)

(1 study, 104 women)

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

481 per 1000

212 per 1000

(115 to 385)

RR 0.44

(0.24 to 0.80)

(1 study, 104 women)

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

‐

‐

‐

‐

Psychosocial couple‐based prevention programme

(USA)

(Feinberg 2015, randomised trial)

394 per 1000^c

209 per 1000

(126 to 355)

RR 0.53

(0.32 to 0.90)^c

(1 study, 147 women)

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Psychoeducation

(Finland)

(Rouhe 2013, randomised trial)

325 per 1000

228 per 1000

(159 to 328)

RR 0.70

(0.49 to 1.01)

(1 study, 371 women)

㊉㊉㊀㊀

LOW^a,b

475 per 1000

632 per 1000

(527 to 765)

RR 1.33

(1.11 to 1.61)

(1 study, 371 women)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Antenatal education programme for physiologic childbirth

(Iran)

(Masoumi 2016, randomised trial)

437 per 1000

450 per 1000

(315 to 651)

RR 1.03

(0.72 to 1.49)

(1 study, 150 women)

㊉㊉㊉㊀

MODERATE^a

0 per 1000

80 per 1000

(CI not estimable)

Relative effect not estimable

(1 study, 150 women)

570 per 1000

479 per 1000

(353 to 650)

RR 0.84

(0.62 to 1.14)

(1 study, 150 women)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Pelvic floor muscle training exercises

(China)

(Wang 2014, randomised trial)

49 per 100

43 per 100

(18 to 100)

RR 0.87

(0.37 to 2.04)

(1 study, 90 women)

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Antenatal education on natural childbirth preparation with training in breathing and relaxation techniques

(Sweden)

(Bergstrom 2009, randomised trial)

630 per 1000

599 per 1000

(365 to 983)

RR 0.95

(0.58 to 1.56)

(1 study, 977 women)

㊉㊉㊉㊀

MODERATE^d

152 per 1000

138 per 1000

(102 to 187)

RR 0.91

(0.67 to 1.23)

(1 study, 977 women)

663 per 1000

663 per 1000

(603 to 723)

RR 1.00

(0.91 to 1.09)

(1 study, 977 women)

122 per 1000

139 per 1000

(100 to 192)

RR 1.14

(0.82 to 1.57)

(1 study, 977 women)

‐

‐

‐

‐

‐

‐

‐

‐

Computer‐based decision aids (information programme, decision analysis)

(UK)

(Montgomery 2007, randomised trial)

496 per 1000

486 per 1000

(407 to 585)

RR 0.98

(0.82 to 1.18)

(1 study, 478 women)

㊉㊉㊉㊀

MODERATE^d

496 per 1000

412 per 1000

(337 to 506)

RR 0.83

(0.68 to 1.02)

(1 study, 478 women)

202 per 1000

220 per 1000

(156 to 313)

RR 1.09

(0.77 to 1.55)

(1 study, 478 women)

202 per 1000

212 per 1000

(150 to 303

RR 1.05

(0.74 to 1.50)

(1 study, 478 women)

303 per 1000

376 per 1000

(291 to 485)

RR 1.24

(0.96 to 1.60)

(1 study, 478 women)

303 per 1000

291 per 1000

(221 to 385)

RR 0.96

(0.73 to 1.27)

(1 study, 478 women)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Decision aid booklet

(Australia)

(Shorten 2005, randomised trial)

Baseline: 23.2%

Follow‐up: 49.4%

Change from baseline: 26.2%

Baseline: 29.6%

Follow‐up: 52.2%

Change from baseline: 22.6%

Relative effect not reported

Difference in absolute change from baseline: ‐3.6% (NS)

(1 study, 227 women)

㊉㊉㊉㊀

MODERATE^a

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Intensive group therapy (cognitive behavioural therapy and childbirth psychotherapy)

(Finland)

(Saisto 2001, randomised trial)

484 per 1000

436 per 1000

(315 to 600)

RR 0.90

(0.65 to 1.24)

(1 study, 176 women)

㊉㊉㊀㊀

LOW^a,b

286 per 1000

235 per 1000

(143 to 389)

RR 0.82

(0.50 to 1.36)

(1 study, 176 women)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Psychoeducation sessions by telephone

(Australia)

(Fenwick 2015, randomised trial)

419 per 1000

339 per 1000

(235 to 494)

RR 0.81

(0.56 to 1.18)

(1 study, 184 women)

㊉㊀㊀㊀

VERY LOW^a,b,e

247 per 1000

173 per 1000

(96 to 304)

RR 0.70

(0.39 to 1.23)

(1 study, 182 women)

419 per 1000

482 per 1000

(352 to 666)

RR 1.15

(0.84 to 1.59)

(1 study, 184 women)

161 per 1000

176 per 1000

(92 to 333)

RR 1.09

(0.57 to 2.07)

(1 study, 184 women)

‐

‐

‐

‐

‐

‐

‐

‐

Prenatal education for husbands of pregnant women

(Iran)

(Sharifirad 2013, randomised trial)

50.0%

(number of events not reported)

29.5%

(number of events not reported)

Relative effect not reported

P < 0.05

(1 study, 88 women)

㊉㊀㊀㊀

VERY LOW^b,c,f

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Role play versus standard education using lectures

(Iran)

(Navaee 2015, randomised trial)

56 per 100

37 per 100

(22 to 63)

RR 0.66

(0.39 to 1.12)

(1 study, 67 women)

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Interactive decision aid versus educational brochures

(USA)

(Eden 2014, randomised trial)

37%

Number of events unclear

41%

Number of events unclear

P = 0.72

Number of participants unclear

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Individualised prenatal education and support programme versus written information in pamphlet

(Canada, USA)

(Fraser 1997, randomised trial)

237 per 1000

213 per 1000

(175 to 263)

RR 0.90

(0.74 to 1.11)

(1 study, 1275 women)

㊉㊉㊉㊀

MODERATE^a

690 per 1000

607 per 1000

(400 to 918)

RR 0.88

(0.58 to 1.33)

(1 study, 1275 women)

490 per 1000

529 per 1000

(475 to 593)

RR 1.08

(0.97 to 1.21)

(1 study, 1275 women)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Rates of maternal morbidity and neonatal outcomes were similar in the study groups (maternal–uterine rupture or dehiscence, hysterectomy, blood transfusion; neonatal–perinatal deaths, Apgar score less than 7 at 5 minutes, admission to NICU)

㊉㊉㊉㊀

MODERATE^a

ǂThe corresponding risk (absolute effect with intervention) (and its 95% confidence interval) is based on the assumed risk in the comparison group ((i.e. risk with control) and the relative effect of the intervention (and its 95% CI).

About the certainty of the evidence (GRADE)*

High: this research provides a very good indication of the likely effect; the likelihood that the effect will be substantially different^† is low.
Moderate: this research provides a good indication of the likely effect; the likelihood that the effect will be substantially different^† is moderate.
Low: this research provides some indication of the likely effect; however, the likelihood that it will be substantially different^† is high.
Very low: this research does not provide a reliable indication of the likely effect; the likelihood that the effect will be substantially different^† is very high.

^*This is sometimes referred to as ‘quality of evidence’ or ‘confidence in the estimate’

^†Substantially different = a large enough difference that it might affect a decision.

CI: confidence interval; NICU: neonatal intensive care unit; NR: not reported; NS; not significant; RR: risk ratio; VBAC: vaginal birth after caesarean.

^aDowngraded one level for serious risk of bias (due to inadequate randomisation processes).

^bDowngraded one level for serious imprecision (due to small sample size and few events).

^cReanalysed, based on: control event rate (40%, n = 71); intervention event rate (21%, n = 76); odds ratio (OR) 0.36, 95% CI 0.15 to 0.86).

^dDowngraded one level due to serious imprecision (95% CI includes appreciable benefit and harm).

^eDowngraded one level for serious indirectness (follow‐up analyses, not described in the trial report, indicated that the impact on caesarean sections was due to reduced birth complications arising from foetal position (e.g. breech birth) and labour progression).

^fDowngraded two levels for very serious risk of bias (due to inadequate randomisation processes and reporting issues).

Implementation of clinical practice guidelines combined with mandatory second opinion

(Argentina, Brazil, Cuba, Guatemala and Mexico)

(Althabe 2004, cluster‐randomised trial)

Mean baseline rate: 24.6 (39,175 women)

Mean follow‐up rate: 24.9 (39,638 women)

Mean rate change: 0.3

Mean baseline rate: 26.3 (34,735 women)

Mean follow‐up rate: 24.7 (35,675 women)

Mean rate change: ‐1.6

Mean difference in rate change:

‐1.9 (‐3.8 to ‐0.1)

㊉㊉㊉㊉

HIGH

Mean baseline rate: 9.1 (39,175 women)

Mean follow‐up rate: 9.0 (39,638 women)

Mean rate change: ‐0.1

Mean baseline rate: 8.9 (34,735 women)

Mean follow‐up rate: 9.1 (35,675 women)

Mean rate change: 0.1

Mean difference in rate change:

0.2 (‐1.4 to 1.8)

Mean baseline rate: 15.4 (39,175 women)

Mean follow‐up rate: 15.9 (39,638 women)

Mean rate change: 0.4

Mean baseline rate: 17.4 (34,735 women)

Mean follow‐up rate: 15.6 (35,675 women)

Mean rate change: ‐1.8

Mean difference in rate change:

‐2.2 (‐4.3 to ‐0.1)

‐

‐

‐

‐

‐

‐

‐

‐

Mean baseline rate per 10,000 livebirths (39 175 women): 5.9

Mean follow‐up rate per 10,000 livebirths (39 638 women): 7.5

Mean baseline rate per 10,000 livebirths (34 735 women): 3.2

Mean follow‐up rate per 10,000 livebirths (35 675 women): 4.3

Mean difference in rate change: 0.66 (‐4.0 to 5.3) (re‐analysed)

㊉㊉㊉㊉

HIGH

‐

‐

‐

‐

Mean baseline rate (39,175 women): 1.1

Mean follow‐up rate (39,638 women): 1.0

Mean rate change: ‐0.1

Mean baseline rate (34,735 women): 1.1

Mean follow‐up rate per 10,000 livebirths (35 675 women): 0.9

Mean rate change: ‐0.2

Mean difference in rate change (95% CI):

‐0.1 (‐0.4 to 0.3)

㊉㊉㊉㊉

HIGH

Mean baseline rate (39,175 women): 3.1

Mean follow‐up rate (39,638 women): 3.1

Mean rate change: 0.0

Mean baseline rate (34,735 women): 4.3

Mean follow‐up rate per 10,000 livebirths (35 675 women): 3.4

Mean rate change: ‐1.0

Mean difference in rate change (95% CI):

–0·9 (–1·9 to –0·0)

㊉㊉㊉㊉

HIGH

Implementation of clinical practice guidelines combined with audit and feedback

(Canada)

(Chaillet 2015, cluster‐randomised trial)

Baseline: 6671/28,698 (23.2%)

Post‐intervention: 6767/28,781 (23.5%)

Baseline: 5484/24,388 (22.5%)

Post‐intervention: 5128/23,484 (21.8%)

RD ‐1.8% (‐3.8 to ‐0.2)

㊉㊉㊉㊉

HIGH

Baseline: 1256/14,717 (8.5%)

Post‐intervention: 1172/13,019 (9.0%)

Baseline: 971/11,478 (8.5%)

Post‐intervention: 763/10,067

(7.6%)

RD ‐1.7% (‐3.0 to ‐0.3)

Baseline:

2404/28,698 (8.4%)

Post‐intervention: 2598/28,781 (9.0%)

Baseline:

1995/24,388 (8.2%)

Post‐intervention: 1931/23,484 (8.2%)

RD – 0.6% (‐0.07 to 1.28)

‐

‐

‐

‐

‐

‐

‐

‐

Baseline:

138/28,698 (0.48%)

Post‐intervention:

141/28,781 (0.49%)

Baseline:

161/24,388 (0.66%)

Post‐intervention:

167/23,484 (0.71%)

RD 0.03%

(‐0.11 to 0.23)

㊉㊉㊉㊉

HIGH

Baseline:

3869/28,698 (13.5%)

Post‐intervention:

4244/28,781 (14.7%)

Baseline:

3293/24,388 (13.5%)

Post‐intervention:

3576/23,484 (15.2%)

RD 0.3%

(‐1.2 to 1.8)

Baseline:

1018/29,107 (3.5%)

Post‐intervention:

1156/29,211 (4.0%)

Baseline:

1172/24,823 (4.7%)

Post‐intervention:

1070/23,902 (4.5%)

RD ‐0.7%

(‐1.3 to ‐0.1)

Baseline:

3947/29,107 (13.6%)

Post‐intervention:

5002/29,211 (17.1%)

Baseline:

3936/25,823 (15.9%)

Post‐intervention:

4261/23,902 (17.8%)

RD ‐1.7%

(‐2.6 to ‐0.9)

Baseline:

14/29 107 (0.0%)

Post‐intervention:

28/29,211 (0.0%)

Baseline:

35/24 823 (0.1%)

Post‐intervention:

20/23,902 (0.1%)

RD ‐0.06%

(‐0.08 to ‐0.03)

Physician education by local opinion leader (obstetrician‐gynaecologist)

Audit and feedback

(Canada)

(Lomas 1991, cluster‐randomised trial)

Control:

66.8% (61.7 to 72.0)

Opinion leader education:

53.7% (46.5 to 61.0)

‐

㊉㊉㊉㊉

HIGH

Control:

66.8% (61.7 to 72.0)

Audit and feedback:

69.7% (62.4 to 77.0)

‐

Control:

18.7% (15.4 to 22.1)

Opinion leader education:

21.4% (16.8 to 26.1)

‐

Control:

18.7% (15.4 to 22.1)

Audit and feedback:

18.6% (13.9 to 23.2)

‐

Control:

14.5% (10.3 to 18.7)

Opinion leader education:

25.3% (19.3 to 31.2)

‐

Control:

14.5% (10.3 to 18.7)

Audit and feedback:

11.8% (5.8 to 17.7)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Control: 1.2 (0.0 to 2.4)

Opinion leader education: 0.9 (0.0 to 2.6)

‐

㊉㊉㊉㊉

HIGH

Control: 1.2 (0.0 to 2.4)

Audit and feedback: 5.9 (4.2 to 7.6)

‐

Education of public health nurses on childbirth classes

(Finland)

(Hemminki 2008, cluster‐randomised trial)

160 per 1000

198 per 1000

(159 to 242)

OR 1.29

(0.99 to 1.67)

㊉㊉㊀㊀

LOW^a,b

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Peer review plus mandatory second opinion for caesarean section indication

(Taiwan)

(Liang 2004, interrupted time series study)

Change in level of total caesarean deliveries at 12 months^c: ‐2.4% (‐11.4 to 6.7);

change in slope^c: 1.34% (‐2.5 to 5.2).

㊉㊀㊀㊀

VERY LOW^d

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Audit and feedback using Robson classification

(Chile)

(Scarella 2011, interrupted time series study)

Change in level of caesarean deliveries during intervention^c: ‐11% (‐23.2 to 1.2), NS; change in slope^c ‐1.1% (‐6.4 to 4.2), NS

Change in level of caesarean deliveries in the immediate post‐intervention period compared with the intervention period^c: 8.6% (2.1 to 15.2), P = 0.022;

change in slope^c: ‐0.3% (‐1.6 to 0.9), NS

㊉㊀㊀㊀

VERY LOW^c

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Audit and feedback plus financial incentive

(Iran)

(Mohammadi 2012, controlled before‐after studies (reanalysed using interrupted time series methods))

Change in level of caesarean deliveries during the intervention^c: ‐14.6% (‐24.4 to ‐4.8), P = 0.02;

change in slope^c: ‐0.07% (‐1.5 to 1.3), NS

㊉㊀㊀㊀

VERY LOW^d

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Audit and feedback plus 24‐hour in‐house coverage by dedicated physician

(USA)

(Poma 1998, interrupted time series study)

Change in level of total caesarean deliveries (primary and repeat caesarean sections) at 24 months^c: ‐6.6% (‐10.1 to ‐3.2); change in slope^c: ‐0.11% (‐0.25 to 0.02) (data reanalysed)

㊉㊀㊀㊀

VERY LOW^d

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

ǂThe corresponding risk (absolute effect with intervention) (and its 95% confidence interval) is based on the assumed risk in the comparison group ((i.e. risk with control) and the relative effect of the intervention (and its 95% CI).

About the certainty of the evidence (GRADE)*

High: this research provides a very good indication of the likely effect; the likelihood that the effect will be substantially different^† is low.
Moderate: this research provides a good indication of the likely effect; the likelihood that the effect will be substantially different^† is moderate.
Low: this research provides some indication of the likely effect; however, the likelihood that it will be substantially different^† is high.
Very low: this research does not provide a reliable indication of the likely effect; the likelihood that the effect will be substantially different^† is very high.

^*This is sometimes referred to as ‘quality of evidence’ or ‘confidence in the estimate’

^†Substantially different = a large enough difference that it might affect a decision.

CI: confidence interval; NR: not reported; NS: not significant; RD: risk difference; RR: risk ratio.

^aDowngraded one level for serious risk of bias (pilot study with no sample size calculation; unit of analysis error).

^bDowngraded one level for serious imprecision (confidence interval includes null effect)

^cTwo standardised effect sizes are obtained from ITS analysis: change in level (also called ‘step change’) and change in trend (also called ‘change in slope’) before and after the intervention. Change in level = difference between the observed level at the first intervention time point and that predicted by the pre‐intervention time trend; Change in trend = difference between post‐ and pre‐intervention slopes. A negative change in level and slope indicates a reduction in caesarean section rate.

^dDowngraded one level for possible confounding (unclear whether the intervention occurred independently of other changes over time).

Insurance reforms equalising physician fees for vaginal and caesarean section deliveries

(USA)

(Keeler 1996, interrupted time series study)

Caesarean section rates for non‐breech deliveries decreased by 1.2% (22.5% before reform versus 21.3% after reform)

㊉㊀㊀㊀

VERY LOW^a

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Insurance reforms equalising physician fees for vaginal and caesarean section deliveries

(Taiwan)

(Lo 2008, interrupted time series study)

The change in the level of total caesarean section rate following the rise in VBAC fees was ‐1.68 (95% CI ‐2.3 to ‐1.07); the change in slope was ‐0.004 (95% CI ‐0.05 to 0.04)^b

The change in the level of total caesarean section rate (for all indications and order of birth) following the rise in vaginal birth fees was 1.19 (95% CI ‐0.01 to 2.40) and the change in slope was ‐0.43 (95% CI ‐0.78 to ‐0.09)^b

㊉㊀㊀㊀

VERY LOW^a

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Collaborative midwifery‐labourist care (versus private model of care)

(USA)

(Rosenstein 2015, interrupted time series study)

Primary caesarean rate among privately insured women decreased from 31.7% to 25.0% (OR 0.56, 95% CI 0.39 to 0.81). Interrupted time series analysis estimated a 7% drop in the primary caesarean rate in the year after the intervention, and a decrease of 1.7% per year thereafter

㊉㊉㊀㊀

LOW^c

VBAC rate increased from 13.3% before to 22.4% after the intervention (OR 2.03, 95% CI 1.08 to 3.80)

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Labourist model of obstetric care (versus traditional model of obstetric care)

(USA)

(Srinivas 2016, controlled before‐after study)

Non‐labourist before:

28.5% (46,486 births)

Non‐labourist after:

31.8% (42,348 births)

Labourist before:

32.6% (47,206 births)

Labourist after:

33.6% (35,210 births)

OR 1.02

(0.97 to 1.1)

㊉㊉㊀㊀

LOW^c

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

Non‐labourist before, % (N): 6.2 (10,018)

Non‐labourist before, % (N): 4.8 (6339)

Labourist before, % (N): 3.8 (5549)

Labourist after, % (N): 3.5 (3814)

OR 1.07 (0.88 to 1.30)

㊉㊉㊀㊀

LOW^c

‐

‐

‐

‐

Non‐labourist before, % (N): 0.4 (557)

Non‐labourist after, % (N): 0.4 (476)

Labourist before, % (N): 0.2 (216)

Labourist after, % (N): 0.2 (223)

OR 1.09 (0.69 to 1.72)

㊉㊉㊀㊀

LOW^c

Non‐labourist before, % (N): 0.3 (398)

Non‐labourist after, % (N):

0.2 (247)

Labourist before, % (N): 0.2 (310)

Labourist after, % (N): 0.2 (171)

OR 0.75 (0.48 to 1.18)

㊉㊉㊀㊀

LOW^c

ǂThe corresponding risk (absolute effect with intervention) (and its 95% confidence interval) is based on the assumed risk in the comparison group ((i.e. risk with control) and the relative effect of the intervention (and its 95% CI).

About the certainty of the evidence (GRADE)*

High: this research provides a very good indication of the likely effect; the likelihood that the effect will be substantially different^† is low.
Moderate: this research provides a good indication of the likely effect; the likelihood that the effect will be substantially different^† is moderate.
Low: this research provides some indication of the likely effect; however, the likelihood that it will be substantially different^† is high.
Very low: this research does not provide a reliable indication of the likely effect; the likelihood that the effect will be substantially different^† is very high.

^*This is sometimes referred to as ‘quality of evidence’ or ‘confidence in the estimate’

^†Substantially different = a large enough difference that it might affect a decision.

CI: confidence interval; NR: not reported; OR: odds ratio; RR: risk ratio; VBAC: vaginal birth after caesarean.

^aDowngraded one level for serious risk of bias (due to possible confounding of outcome; unclear whether the intervention occurred independently of other changes over time).

^bTwo standardised effect sizes are obtained from interrupted time series analysis: a change in level (also called ‘step change’) and a change in trend (also called ‘change in slope’) before and after the intervention.

Change in level = difference between the observed level at the first intervention time point and that predicted by the pre‐intervention time trend; change in trend = difference between post‐ and pre‐intervention slopes. A negative change in level and slope indicates a reduction in caesarean section rate.

^cObservational study which start at low certainty evidence according to GRADE (we did not downgrade or upgrade the certainty of evidence).

Multifaceted programme comprising education programme for hospital staff and women, audit of surgeon practices, public health campaign, monitoring rates of caesarean sections and neonatal outcomes

(China)

(Runmei 2012, controlled before‐after study)

Change in level of caesarean deliveries during intervention: ‐13.4% (95% CI ‐19.6 to ‐7.1)^b

Change in slope of caesarean deliveries: ‐0.72% (95% CI ‐3 to 1.5)^b

㊉㊀㊀㊀

VERY LOW^c

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

"We found a significant increase in the incidence of all obstetric complications, with the exception of placental abruption, after 2004"

㊉㊀㊀㊀

VERY LOW^c

‐

‐

‐

‐

"The incidence of birth asphyxia did not increase after 2004 (P = 0.303)"

㊉㊀㊀㊀

VERY LOW^c

‐

‐

‐

‐

Multifaceted programme comprising transmission of information on caesarean section, training of health workers on best obstetric practices and inclusion of caesarean section rates as a criterion for hospital funding

(Portugal)

(Ayres‐De‐Campos 2015, interrupted time series study)

In the period between 2009 and 2014, representing the possible influence of the programme:

rates of caesarean section in the study region decreased by 20.0% (from 36.0% to 28.8%, time trend P < 0.001)^b;

rates of instrumental vaginal delivery increased by 33.1% (from 13.7% to 18.2%, time trend P < 0.001)^b;

rates of VBAC increased by 99.8% (from 16.4% to 32.8%, time trend P < 0.001)^b

㊉㊀㊀㊀

VERY LOW^c

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

‐

The incidence of hypoxia‐related complications decreased by 14.1% (from 0.71% to 0.61%, time trend P < 0.001)^b

㊉㊀㊀㊀

VERY LOW^c

ǂThe corresponding risk (absolute effect with intervention) (and its 95% confidence interval) is based on the assumed risk in the comparison group ((i.e. risk with control) and the relative effect of the intervention (and its 95% CI).

About the certainty of the evidence (GRADE)*

High: this research provides a very good indication of the likely effect; the likelihood that the effect will be substantially different^† is low.
Moderate: this research provides a good indication of the likely effect; the likelihood that the effect will be substantially different^† is moderate.
Low: this research provides some indication of the likely effect; however, the likelihood that it will be substantially different^† is high.
Very low: this research does not provide a reliable indication of the likely effect; the likelihood that the effect will be substantially different^† is very high.

^*This is sometimes referred to as ‘quality of evidence’ or ‘confidence in the estimate’

^†Substantially different = a large enough difference that it might affect a decision.

CI: confidence interval; NR: not reported; VBAC: vaginal birth after caesarean.

^aMultifaceted interventions with components targeted at women, healthcare professionals or healthcare organisations.

^bTwo standardised effect sizes are obtained from interrupted time series analysis: a change in level (also called ‘step change’) and a change in trend (also called ‘change in slope’) before and after the intervention.

Change in level = difference between the observed level at the first intervention time point and that predicted by the pre‐intervention time trend; change in trend = difference between post‐ and pre‐intervention slopes. A negative change in level and slope indicates a reduction in caesarean section rate.

^cDowngraded one level for serious risk of bias (due to possible confounding of outcome, unclear whether the intervention occurred independently of other changes over time).