Vitamin A deficiency (VAD) is an important problem in many developing countries. If population data are unavailable or outdated for common indicators of VAD, a surrogate indicator may suggest whether a problem is likely and full-scale assessment is necessary. The Global Vitamin A Initiative report suggests that the under-five mortality rate (U5MR) might serve that purpose; a U5MR >70 indicates that VAD of public health importance is likely. Should the U5MR be used as a surrogate indicator? If so, is 70 the appropriate cutoff? U5MR data were collected for countries where the United Nations Children's Fund supports programs, along with unpublished vitamin A survey data from the World Health Organization. U5MR distribution was compared in countries with and without VAD. Sensitivity and specificity using U5MR cutoffs of 70 and 50 were calculated and compared with the presence of VAD as proven by a survey. Countries with a U5MR >40 have a VAD problem, about 70% of countries with a U5MR between 20 and 50 have VAD, and about 25% of countries with a rate <20 still have VAD. With a cutoff of 50, the U5MR relates well to the existence of a VAD problem in regions where malnutrition problems are most prominent. Decreasing the cutoff from 70 to 50 increased the sensitivity of the criterion and included nine additional deficient countries. A cutoff of 40 would increase sensitivity but would include countries with localized rather than nationwide problems. It is proposed that a country with a U5MR >50 likely has a VAD problem that requires immediate or continuing action. These actions include proper, formal assessment of the vitamin A status of the population, vitamin A capsule distribution and other program actions. Countries with a U5MR between 20 and 50 should assume they have a VAD public health problem until surveys prove otherwise.