The ideal long-term, randomized, placebo-controlled trial of hormone replacement therapy (HRT) from near menopause for up to 30 years to assess major morbidity and mortality is impractical because of high cost, participant retention, therapy compliance, and continuity of research staff and funding. Also the trial regimen may become outdated. It is nihilistic to demand such a long-term trial before endorsing HRT. However, medium-term trials using surrogate measures for long-term morbidity and mortality are possible and two are near completion. If these studies have been able to maintain reasonable participant retention, therapy compliance and minimal breach of protocol, they will set standards for trials of new HRT regimens. This paper discusses lessons learnt from past attempts at long-term trials and suggests the currently optimal protocol and cost of assessing new HRT regimens to optimize potential benefits and minimize adverse effects. A 5-7-year randomized, placebo-controlled trial of a flexible transdermal estrogen regimen ± either a selective estrogen receptor modulator, e.g. bazedoxifene, or micronized progesterone is discussed. Mild to moderately symptomatic women, 1-4 years post menopause, can be recruited via general practice and group meetings. Future trials should be funded by independent agencies and are high priority in women's health.