Case Study No. 8. Tablets are more acceptable and give fewer problems than syrups among young HIV-infected children in resource-limited settings in the ARROW trial