Johnson 1995.

Methods	RCT cross‐over ‐ using here only data from first part.
Participants	Inclusion criteria: women scheduled for attempted ECV with the fetal spine in the midline (back‐up or back‐down) on US examination. N = 26. All women approached agreed to participate. Exclusion criteria: oligohydramnios (AFI < 5 cm), fetal or uterine anomalies, ruptured amniotic membranes, active labour, engagement of presenting part, fetal heart rate decelerations.
Interventions	Intervention:vibroacoustic stimulation (B). Fetal acoustic stimulation for 1 to 3 seconds with a Western Electric Division AT&T (Phoenix) model 5C electrolarynx over the fetal head, or over the nurse's upper arm (dummy). Physician blinded by leaving the room during the intervention. N = 12. Comparison:placebo. N = 11.
Outcomes	Persistent midline spine position on US (stimulation 1/13, control 13/13); failed ECV attempt. Data on method of delivery not included because followed cross‐over treatment.
Notes	2 hospitals in Arizona, USA, 1 January 1993 to 31 December 1994. After randomisation, 1 from the treatment group and 2 from the control group were excluded because the breech was found to be deeply engaged in the pelvis during the initial ECV attempt. None had changed position to the spine lateral position, and no further attempts at ECV were made. In keeping with the pre‐stated protocol for this review, these women have been included in the outcomes as originally allocated. Those women in whom ECV failed were crossed over to the other intervention arm. This review considers only data from the first intervention, according to the original allocation Results of the 'cross‐over' part of the study are not included.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Table of random numbers.
Allocation concealment (selection bias)	Low risk	Selection of sequential envelopes generated by a table of random numbers and handed out by research nurse.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Clinicians were blinded, but the nurse applying the stimulation and the women could not be blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Clinicians reported blinded, but the nurse applying the stimulation and the women could not be blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	3 women (12%) (1 treatment and 2 control) excluded, as breech was deeply engaged. This loss should be insufficient to affect the comparison.
Selective reporting (reporting bias)	Unclear risk	We did not assess the trial protocol.
Other bias	Unclear risk	No baseline data are available for assessment.