Objective. To describe the epidemiology of incomplete abortion (sponta
neous miscarriage and illegally induced) in South Africa. Design. Mult
icentre, prospective, descriptive study. Setting. Fifty-six public hos
pitals in nine provinces (a stratified, random sample of all hospitals
treating gynaecological emergencies). Patients. All women of gestatio
n under 22 weeks who presented with incomplete abortion during the 2-w
eek study period. Main outcome measures. Incidence of, morbidity assoc
iated with and mortality from incomplete abortion. Main results. An es
timated 44 686 (95% CI 35 633 - 53 709) women per year were admitted t
o South Africa's public hospitals with incomplete abortion. An estimat
ed 425 (95% CI 78 - 735) women die in public hospitals from complicati
ons of abortion. Fifteen per cent (95% CI 13 - 18) of patients have se
vere morbidity while a further 19% (95% CI 16 - 22) have moderate morb
idity, as assessed by categories designed for the study which largely
reflect infection. There were marked inter-provincial differences and
inter-age group differences in trimester of presentation and proportio
n of patients with appreciable morbidity. Conclusions. Incomplete abor
tions and, in particular, unsafe abortions are an important cause of m
ortality and morbidity in South Africa. The methods used in this study
underestimate the true incidence for reasons that are discussed, A hi
gh priority should be given to the prevention of unsafe abortion.