The amount of data being collected and stored is huge and is expanding at a vivid pace at both the national and international level. Health care organizations correspondingly generate a large volume of information every day. The health care industry is rich in information but it needs to discover hidden relationships and patterns in this data. This paper intends to use data mining techniques to discover knowledge in a dataset that was provided by a research center in Tehran. By analyzing the drugs that were bought by each patient, this paper aims to predict what kind of physician each patient has referred to and what kind of disease they are suffering from. The dataset includes details such as sex, age and the names of the drugs prescribed for each patient. For labeling the instances, a group of pharmacy students and professors has determined each patient's disease. A number of experiments have been performed to compare the performance of different data mining techniques for predicting the diseases and the results illustrate that the proposed Stacking Model has higher accuracy compared to other techniques such as k-Nearest Neighbor (kNN), Naïve Bayes, Decision Tree etc.