A body sensor network that consists of wearable and/or implantable biosensors has been an important front-end for collecting personal health records. It is expected that the full integration of outside-hospital personal health information and hospital electronic health records will further promote preventative health services as well as global health. However, the integration and sharing of health information is bound to bring with it security and privacy issues. With extensive development of healthcare applications, security and privacy issues are becoming increasingly important. This paper addresses the potential security risks of healthcare data in Internet-based applications and proposes a method of signal scrambling as an add-on security mechanism in the application layer for a variety of healthcare information, where a piece of tiny data is used to scramble healthcare records. The former is kept locally and the latter, along with security protection, is sent for cloud storage. The tiny data can be derived from a random number generator or even a piece of healthcare data, which makes the method more flexible. The computational complexity and security performance in terms of theoretical and experimental analysis has been investigated to demonstrate the efficiency and effectiveness of the proposed method. The proposed method is applicable to all kinds of data that require extra security protection within complex networks.