In order to evaluate strategies to reduce HIV transmission through breast milk and optimize both maternal and infant health among HIV-infected women and their infants, we designed and implemented a large, randomized clinical trial in Lilongwe, Malawi. The development of protocols for large, randomized clinical trials is a complicated and lengthy process often requiring alterations to the original research design. Many factors lead to delays and changes, including study site-specific priorities, new scientific information becoming available, the involvement of national and international human subject committees and monitoring boards, and alterations in medical practice and guidance at local, national, and international levels. When planning and implementing a clinical study in a resource-limited setting, additional factors must be taken into account, including local customs and program needs, language and socio-cultural barriers, high background rates of malnutrition and endemic diseases, extreme poverty, lack of personnel, and limited infrastructure. Investigators must be prepared to modify the protocol as necessary in order to ensure participant safety and successful implementation of study procedures. This paper describes the process of designing, implementing, and subsequently modifying the Breastfeeding, Antiretrovirals, and Nutrition, (BAN) Study, a large, on-going, randomized breastfeeding intervention trial of HIV-infected women and their infants conducted at a single-site in Lilongwe, Malawi. We highlight some of the successes, challenges, and lessons learned at different stages during the conduct of the trial.