AHM2007 Medical E-science session
Brief notes on this parallel session from E-science all hands meeting on Tuesday morning.
First talk in this session discussed the CARMEN project which aims to provide repositories and tools for neuroscience electrophyisology data. There was a short discussion on the challenges of persuading scientists to put the data in. The speaker’s (Paul Watson) view was that this would probably need to be driven by funders and journals. He made the point that this was a specific area where it was difficult to persuade people of the value of open data as it is highly competitive.
The second talk by Sarah Thew focussed on looking at the requirements for visualisation techniques within the ADVISES project. She discussed the challenges of getting a clear view from the scientists of what they want and the approach to sorting this out through developing scenarios and story boards. Looking at epidemiology as an example there is a move towards data-driven (-omics in my language) styles of reserach and the need for more effective use of maps. Apparently maps are not widely or effectively used in epidemiology. Part of the problem is that existing tools are not very easy to use for these scientists.
In terms of the process for defining requirements she made the interesting point that starting with a large piece of paper ‘levels the playing field’ meaning that users, implementors and developers are all on the same level. Go from this through a process of building an example based on a specific scenario and then move on to building an actual prototype tool. Interactive visualisations are very powerful but are not straighforward to build. They are going through multiple stages of prototyping and testing to develop useful tools.
The issue of differing attitudes to data access from different members of the user group was raised again. I asked the question as to whether the focus on a cyclical prototyping led to a risk of developing something that was only useful to a small group. The groups development approach aims to be modular which answers part of the question but equally they aim to develop ontologies that are useful for that specific group. This does have risks but someone has to start the development somewhere.
The final talk was on the clinical challenges of E-science by John Ainsworth. Here privacy and data security is a serious issue. Talk about PsyGrid and how the clinical community shaped the project. PysGrid aims to developing application and middleware to support studies and clinical trials in mental health and to use the technical outputs to conduct a study of people suffering from schizophrenia over time.
He identified a series of risks but most interesting here was probably looking closely at the response from different parts of the user community from project managers through clinical academics to research nurses. In particular it is very easy to alienate the RNs because they will have to learn a new system and may not benefit very much. This is in some ways similar to the issues faced by researchers using an e-notebook. Hopefully they will see more direct benefits but there is an analogy here. The PI’s probably see the biggest benefits from deploying these type of systems. Final point was essentially that doing the basics well solves a lot of potential problems.