Data mining techniques are extremely popular in the healthcare industry. The popularity arises from their flexibility, ability to provide reliable results and the absence of restriction on the data type. These methods are actively being used in improving line of treatment and diagnosis, customer relationship management, management of healthcare resources and fraud and anomaly detection. The main goal of this paper is to address the previous work done in the healthcare sector using data mining techniques and propose an appropriate framework to increase accuracy. We have also studied different statistical tests that can be used on such healthcare datasets.