Davis 2005.

Methods	Randomisation list was prepared with random number tables. Double blind 76 participants randomised, 74 analysed.
Participants	Inclusion: aged 19 years or younger with moderate or severe dysmenorrhoea, regular menstrual cycles for at least 1 year, 21 to 35 days of menstrual cycle length, condom users. 300 adolescents were screened for eligibility. Exclusion: pregnancy, history of pelvic pathology, abnormal genital bleeding, recent miscarriage or abortion, use of other medications likely to interfere with metabolisms of oral contraceptives. Age: OC group: 16.7 +/‐ 2, Placebo group: 16.9 +/‐ 2 Source: medical centre, college campuses Location: USA
Interventions	1. Ethinyl estradiol 0.02mg, 0.1mg levonorgestrel 2. Placebo Duration: 3 cycles
Outcomes	Pain severity (5 point scale) Rating of worst pain intensity Use of analgesic medication Absence from work or study Adverse events Discontinuation rate
Notes	3 publications from one study
Risk of bias
Bias	Authors' judgement	Support for judgement
Adequate sequence generation?	Low risk	Random numbers table
Allocation concealment?	Low risk	A ‐ Adequate
Blinding? All outcomes	Low risk	Double
Incomplete outcome data addressed? All outcomes	Low risk	76 randomised and 74 analysed
Free of selective reporting?	Low risk
Power calculations	Low risk	Sample size was calculated for power to detect differences in main outcome of dysmenorrhoea between the OC group and the placebo group
Dropouts reported	Low risk	7/ 150 (4 from OCP and 3 from placebo group)
Baseline comparability	Low risk