Programmable electronic medical systems - related risks and learning from accidents

out being hospitalized. More patients stay at home – or at least outside the reach of health-care providers – for self-treatment, in which they use medical devices, such as dialysis or insulin injector implants. Remote diagnosis and even treatment is already being applied for specific forms of illness, for example cardiac failure problems, in the sense of telemedicine or even telesurgery. Risks related to the application of a medical system result from inappropriate use; for example , a selected device cannot do the anticipated job because it was not designed for it, or because of functional failure, through user error, technical failure or a design flaw. Technical failures and design flaws may also introduce unanticipated occupational risks. In a networked configuration of multiple medical systems, erroneous behaviour of a subsystem may lead to functional failure of another subsystem: wrong data from a source used as valid input for a target system then performs a function incorrectly. To date, active medical devices hardly are without programmable components. In future, even passive devices like heart valve prostheses will most likely be equipped with (embedded) smart systems, for instance to monitor the deterioration of such a component. We might 'safely' say that the dissemination of program-mable electronic systems in medical devices is just accelerating. Furthermore, the spread and use of (active) medical devices outside health-care institutions is also growing fast. Thus, the likelihood of undetected functional failures when applying a medical device is potentially increasing rapidly. Case studies [1][2][3] show that different sorts of software design flaws may result in functional failures even though the device is working as designed. In this paper, the criticality of medical devices and their application is explained, and different approaches for evaluation of the risk and the benefit are given; the relation to safety integrity levels is presented. After reviewing some examples, the reporting of device-related incidents and the learning from data about such incidents will be discussed. Principles of organizational learning help to pinpoint requirements for a viable medical device reporting system that also helps to specify requirements of dependable programmable electronic medical systems (PEMS). Whether a medical device is critical ultimately depends on the actual purpose for which it is The application of medical devices is often combined with some risk. This risk needs to be evaluated and compared with the benefits: based on this, evaluation of the device needs to fulfil a certain safety …