A comparison of women's, providers' and ultrasound assessments of pregnancy duration among termination of pregnancy clients in South Africa.

Objective To compare providers' and women's estimates of duration of pregnancy with ultrasound estimates for determining medical abortion eligibility. Design Cross-sectional study. Setting Public termination of pregnancy (TOP) services in three provinces. Sample A total of 673 women attending the above services for TOP. Methods Women participating in a medical abortion feasibility study in South Africa provided estimates of pregnancy duration and date of last menstrual period (LMP). Each woman also had clinical and ultrasound exams. We compared estimates using the four methods, calculating the proportion of women in the 'caution zone' (<= 8 weeks gestation by woman or provider estimate and > 8 weeks by ultrasound). Main outcome measures Mean gestational age by each method; difference between provider and LMP estimates and ultrasound estimates; and percentage of women in the 'caution zone'. Results Women's estimates of pregnancy duration were 19 days fewer than ultrasound estimates (95% CI = -27 to 63). Mean provider- and LMP-based estimates were two (95% CI = -30 to 35) and less than one day(s) (95% CI = -46 to 51) fewer than ultrasound estimates. Comparing provider and ultrasound estimates, 15% of women were in the 'caution zone'; this fell to 12% if estimates of 9 weeks or fewer were considered acceptable. Conclusions Provider estimates of gestational age were sufficiently accurate for determining eligibility for medical abortion. LMP-based estimates were also accurate on average, but included more extreme differences from ultrasound estimates. Medical abortion could be provided in TOP facilities without ultrasound or with ultrasound on referral.