Even when the program office finally acknowledged in 2011 that home visits were unworkable, the requirement of the geographic samplingstructure was not modified so that when investigators started recruiting in clinics, they were forced to restrict recruitment to women living in the randomly selected neighborhoods.