Privacy in the healthcare sector is an issue of growing importance. The adoption of digital patient records, increased regulation, provider consolidation and the increasing need for information exchange between patients, providers and payers, all point towards the need for better security and protection of patient data. This study proposes a mechanism for preserving data privacy by ensuring the protection of patient data in a distributed clinical decision support system using the Random Forest decision tree algorithm. This mechanism facilitates knowledge building using statistics based on patient data from multiple sites. The results of experiments performed by inducing the SCHIZO data set of patients from Biomedical Informatics Research Network (BIRN) with three types of decision tree classification algorithms (J48, Random Tree and Random Forest) embedded in the Waikato Environment for Knowledge Analysis (WEKA) software showed the J48, Random Tree and Random Forest decision tree algorithms to have accuracy of 87.7551%, 89.7959% and 91.8367% respectively in the training phase. The Random Forest decision tree algorithm which gave the highest accuracy was then selected and used to induce the SCHIZOPH data set of patients from Biomedical Informatics Research Network (BIRN) and gave an accuracy of 94.2029% in the testing phase. Two of the 23 attributes in the SCHIZOPH data set were then blocked and the Random Forest decision tree algorithm was used to induce the data set again and it gave an accuracy of 94.2029%. The experiments coupled with restricting a fraction of attributes from sharing statistics as well as applying constraints on privacy demonstrate the usefulness of the decision tree algorithm in preserving privacy of data in a distributed clinical decision support system.