Estimating Illegal Abortion Rates Using Survey Respondents' Knowledge of Abortion Cases in Their Social Networks: An Innovative Method Tested in Ouagadougou, Burkina Faso

Clementine Rossier, University of California at Berkeley
Georges Guiella, University of Ouagadougou
Abdoulaye Ouedraogo, University of Ouagadougou
Blandine Thieba, University of Ouagadougou

This study tests a new method to estimate illegal abortion rates: the Third Party Information method. The TPI method asks respondents to report abortion incidences in their social networks and infers population level abortion rates from this data. This method was first developed in 2000 for a rural population in Burkina Faso: it is here adapted to the urban context of Ouagadougou, Burkina Faso. This study tests two hypotheses: 1) the TPI method yields higher estimates of abortion than two other existing survey estimation methods (the direct questioning and the residual methods); 2) the TPI method can provide reliable approximation of abortion rates. The three "survey" abortion rates are calculated with data from a survey of 1500 individuals representative of Ouagadougou. The "survey" rates are tested for reliability against abortion rates obtained using: 1) an ethnographic approach; 2) hospital abortion complications statistics multiplied by health professionals' estimates.