Vitamin A deficiency (VAD) is a public health problem in India. According to the CNNS, in the states of Bihar, Chhattisgarh and Jharkand VAD is higher than 20% in adolescents. Spatialinformation about the variations in retinol concentration can help targeting efforts. This study analysed serum retinol data from 1,239 adolescents sampled from three states. The mean retinol for the primary sampling units (PSU) were calculated. Ordinary krigingpredictions of the concentration of retinol were computed using aggregated mean values for each PSU. A threshold of 0.7 μmol/L was used to indicate deficiency on the map of conditionalprobabilities. The geostatistical analysis (Figure 1a and b) showed marked spatial dependence in retinol concentration in the study area. The higher concentration of serum retinol was noticed near thecapital of Chhattisgarh and overlapping tribal area between Chhattisgarh and Jharkhand. Higher concentration was also seen in a larger proportion of the mapped area in Bihar, except forthe areas adjoining Jharkhand and western part of Nepal border. The probability of deficiency was higher in Jharkhand compared to the other states. State level deficiency prevalence estimates can mask vitamin A variations within a state. Spatial predications can provide important considerations to design targeted approaches.Uncertainties need to be accounted for while deciding upon the sampling design for future national micronutrient surveys.