The increasing adoption of technology to support independent living at home through the extended use of community alarms and, ultimately, second-generation telecare systems poses some safety problems. The implications of any failure of the technology must be addressed in order to provide a safe and reliable care service. The risks of using home-based technology can be assessed by standard techniques under three categories: environmental factors; human factors; and technological factors. A safety classification system for telecare devices is proposed leading to a range of design guidelines which represent good practice. An example of the use of these design principles is a prototype second-generation telecare system, MIDAS, which is currently undergoing trials. In order to reduce the risks associated with device failure, the system incorporates distributed intelligence, built-in self-testing and redundancy. Potentially hazardous situations can therefore be controlled.