Newnham 1990.

Methods	Randomisation: randomised by draw of sealed envelopes. Effectiveness of randomisation was assessed by comparisons of several parameters within the 2 groups which included maternal ages, parities, gestational ages and primary indications.
Participants	Women of least 34 weeks' gestation. Exclusion criteria ‐ no contraindications to contraction stress tests. Country: Western Australia. 172 women randomised.
Interventions	Women randomised to fetal acoustic stimulation group were monitored for an initial 5 minutes. If the pattern is non‐reactive, a 3‐second vibratory acoustic stimulation was applied to the maternal abdomen in the region of the fetal head. The stimulus was repeated a second and a third time, also at 1‐minute intervals, if satisfactory FHR accelerations had not occurred. Women randomised to non‐fetal acoustic stimulation group were monitored for an initial 20 minutes. If the trace was non‐reactive, the fetus was stimulated manually and the test was continued for a further 20 minutes. If satisfactory accelerations were not found, the women were then sent for a meal. On her return, a nipple stimulation contraction stress was performed if the subsequent test remained non‐reactive after a further 20 minutes.
Outcomes	Primary outcome: FHR reactivity. This was defined by the presence of 2 accelerations of greater than 15 bpm and of 15 seconds' duration or more within a 20‐minute period. In tests in which accelerations had been provoked by fetal acoustic stimulation, the definition of reactivity required 1 of the 2 accelerations to have been unprovoked.
Notes	All tests were performed with Corometrics 115 monitors using Doppler FHR transducers. The tests were performed by specially trained fetal intensive care midwives with a nurse: patient ratio of 1:1. The vibratory acoustic stimulus had an audio frequency of 75 Hz, a sound intensity of 74 db at 1 m in air and a stimulation duration of 3 seconds. Randomisation was performed after informed consent.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by draw of sealed envelopes.
Allocation concealment (selection bias)	Low risk	Sealed envelopes.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No evidence of blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Follow‐up 100%.
Selective reporting (reporting bias)	Unclear risk	No outcomes documented in methods section.
Other bias	Low risk	No evidence of any other form of bias.