Surgery, antibiotics, facial cleanliness and environmental improvements (SAFE) are recommended for trachoma control. Programmes assess clean faces in children, but no standard definition of a clean face exists. We conducted a randomised controlled trial of face-washing to develop a valid and repeatable definition of a clean face. A total of 424 children were randomised to washed and unwashed groups after a first observation. Three additional observations were made throughout the day. Photographs were taken at each observation. No difference was observed in wet nasal discharge, dust, food or flies on the face between the face washed and unwashed groups at baseline or after washing. A difference was observed in the presence of ocular discharge (P < 0.001) and dry nasal discharge (P < 0.001) after washing. Agreement among observers was highest for flies (Kappa = 0.89, 95% CI = 0.87-0.91), followed by nasal (Kappa = 0.64, 0.62-0.66) and ocular (Kappa = 0.48, 0.46-0.50) discharge. The ability of any definition to identify whether a face had been washed decreased at each observation. This study suggests that the absence of ocular and dry nasal discharge can be used as an indicator of 'clean face', although it is not a good predictor of whether a face has been washed and is difficult to recommend
