Background: India and other developing countries need their own reference intervals for various medical parameters because these populations differ from Western population for the genetic profile, anatomical structure, dietary habits, and lifestyle. Such reference intervals have not been worked out so far for most parameters because it is difficult to have a large database on healthy people in these countries. Aims and objectives: Large hospitals generally have a huge database of laboratory values but a substantial proportion of them belong to sick subjects whose values cannot be included for establishing normal reference intervals. Thus, the database remains unutilized. We propose a simple method to utilize these data for establishing reference intervals. Materials and methods: A simple double filtration method is used to exclude all outliers and abnormal values that could finally provide uncontaminated data on healthy values. This method is based on quartiles and interquartile range. The method is illustrated on a dataset from the laboratory of a large tertiary care hospital. Results: The filtered values have been seen to follow a smooth distribution pattern and can be used to establish our reference intervals using the usual 2.5th and 97.5th percentiles. The method is illustrated for A/G ratio in the data from our hospital, and the reference interval obtained. Conclusion: Double filtration method can be used on hospital laboratory data to establish reference intervals of medical parameters.