Characteristics of prospective studies included in the systematic review.

aall adjusted OR/RRs are as reported by authors, bOR/RRs calculated by authors, raw data not available, cbaseline prevalence used, draw data on the % using OC not available, on the basis of the odds ratios reported, the proportion of women using contraceptives were calculated to well exceed 10%, eper number of visits, fnumber or percentage of assessments rather than number of participants reported, gwoman-years, calculated by authors, hraw data was not available, but parent study from which this cohort was derived reported 29% using HC [89], ibaseline prevalence of women using combined methods of contraception was 53.5% of which most went into the longitudinal analysis, but raw data was not reported, junadjusted ORs clustered by clinic so included as reported. Key: HC = hormonal contraception, COC = combined oestrogen- and progesterone-containing methods of HC, POC = progesterone only containing methods of HC, u-HC = unspecified HC, CS = cross-sectional study, LC = longitudinal cohort, RCT = randomised controlled trial (either LC used or CS data used), OR = odds ratio, RR = risk ratio, NC = no contraception used, TL = tubal ligation.