Balk 2002.

Methods	Design: parallel‐group Number of women screened: "hundreds" (most not amenorrhoeic for one year) Number randomly assigned: 27 Number dropped out: seven (five in active arm: two for family reasons, three for adverse effects; two in placebo arm: one for lack of efficacy, one disliked taste) Number lost to follow‐up: one (active arm) Number analysed: 19 Intention‐to‐treat analysis: no Power calculation: powered to detect endometrial changes, but baseline proliferation rate underestimated and study thus underpowered for primary outcome Duration: six months Timing: January 1998 to June 2000 Location: university hospital clinic Funding: academic research grants
Participants	Inclusion criteria: postmenopausal women 40 years of age with no vaginal bleeding for one year or over 30 years of age with oophorectomy or premature ovarian failure, omnivorous, intact uterus, normal endometrium on Pipelle biopsy, normal mammogram within previous year Exclusion criteria: tamoxifen usage, endometrial cancer, allergy to soy, hormone therapy on past year, using phytoestrogen supplements (diet logged for two weeks before study) Age, years: active arm 56.8 ± 5.9; placebo arm 57.9 ± 8.2 Recruitment method: primary and tertiary clinics, newspaper and radio advertisements, research institute website
Interventions	Phytoestrogen: isoflavone Formulation: soy flour and corn cereal (Nutlettes): 3/8‐cup serving contains 92 mg of isoflavones. Mixed with placebo cereal (3:1) to increase similarity of taste Placebo: wheat cereal (Grapenuts) Dose, duration and timing of administration: 100 mg daily (1/2 cup cereal) for three months Given list of soy‐ and phytoestrogen‐containing foods to avoid
Outcomes	Menopausal symptoms: weekly log monitoring nine specific symptoms on 4‐point scale Compliance: daily dietary logs, check of unused cereal Adverse effects: weekly log monitored specific symptoms such as nausea, breast tenderness and gastrointestinal effects
Notes	Authors reported that recruitment was difficult. Participants were not required to have menopausal symptoms to be eligible for the study, although these were measured at baseline. Mean symptom score at baseline indicated that, on average, participants had mild to moderate hot flushes and/or night sweats
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated in blocks of six
Allocation concealment (selection bias)	Low risk	Sequentially numbered opaque sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Stated as double‐blind
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Stated as double‐blind
Incomplete outcome data (attrition bias) All outcomes	High risk	Very high dropout (8/27)
Selective reporting (reporting bias)	Low risk	All prespecified outcomes reported