The project is organised in work packages as follows:
Work package 1 – HypoPredict Parameter Identification
Assessing the ranked list of patient parameters likely to be indicators of subsequent arterial hypotension.
Work package 2 – HypoPredict Engine Design
Designing a Bayesian Neural Network architecture based on the weighted parameter set identified in WP1, refining the algorithms and developing the HypoPredict engine to be used during clinical trials in each of the six clinical centres.
Work package 3 – HypoPredict clinical interfaces
Developing interfaces to the BANN to support data acquisition from patient records, instrumentation and test results.
Work package 4 – HypoPredict training cycle
Validating the HypoPredict parameter conversion methods and providing preliminary confidence levels for predictions.
Work package 5 – HypoNet development and ethical considerations
Defining mechanisms for obtaining patient consent and the GRID based framework to enable data capture from and software distribution to the clinical trials centres.
Work package 6 – Clinical trial
Attaining patient consent, recruiting research nurses, validating and updating the HypoPredict engine and demonstrating its accuracy and effectiveness.
Work package 7 – Consortium Management
Ensuring all aspects of the EC requirements for are met, protecting the project results and disseminating the benefits of the developed technology.
Work package 8 – Dissemination and Exploitation
Ensuring the successful dissemination of the project results into cross sector applications and participating in concertation activities with other ICT projects.
