A new approach: reusable real-time database for medical instrumentation applications.

Abstract

Medical instrumentation (MI) software development work differs from other software development works mainly in the testing and validation phase. For MI software, this phase has always been most difficult, time-consuming and yet the most doubted phase. The primary reason being the input to these programs (ie) the physiological signals like ECG, pulse, etc. To be more specific, the problem is because 1. These physiological signal patterns vary so widely from person to person and hence these programs have to be validated statistically on all categories of persons including those on the extreme ends. 2. There are also artifacts riding on these signals (inherent, intentional or due to inevitable physical movements) 3. These signals are real-time ones to the order of a second. Above all, 4. They are not reproducible patterns, if you need them for study and analysis at a later date. Though it is tough to find alternatives for the present technique of statistical validation, there can be ways to make use of this time-consuming validation process to its fullest potential. One such way is to make a kind of 'CAPTURE-RETAIN AND REUSE' databases that can reduce the future efforts in the validation phase substantially. The main areas where these databases can make vital contributions are: 1. Medical Instrumentation software testing and validation 2. Medical Instruments' results and performance comparison 3. Expert system building, testing and enhancement of its capabilities.