The breakout group for projects with a biomedical/health interest had representation from all strands of the new programme and was included Stelios Alexandrakis & Anthony Thomas, University College London (Data Management Planning for Secure Services); Amanda Conway & Jennifer Crossley, University of the West of England (Managing Research Data: a pilot study in Health and Life Sciences); Michael Soljak, Imperial College London (Rapid Organisation of Health Research Data (ROHRD) Phase 1); Lindsay Wood, Newcastle University (Iridium); Bill Worthington, University of Hertfordshire (Service Oriented Toolkit for Research Data Management); and Jonathan Tedds, University of Leicester (BRISSkit), who took the chair.
After introductions, Jonathan posed the question: ‘what are the common issues facing the group?’. The conversations that followed went through several phases and moved back and forth through themes that have become familiar from my recent travels in the MRD bubble. This post attempts to gather the discussion and straighten the meander:
– the policy gap: Michael described work with datasets of the most demanding regulatory nature: non-consensually gathered clinical data. He was hoping to discover a consistent policy for sharing this data nationally. Lindsay’s concerns were for policy to manage the relationship with clinical trial participants. Amanda noted the need for mechanisms to accommodate the peculiar requirements of non-HE partners from the NHS and commercial sectors.
– the metadata jungle: Amanda questioned ‘is there a metadata standard for health related data?‘ The meeting agreed there was not. Anthony and Stelios described their work to model the DCC’s DMPonline with the DDI metadata scheme (and as a result of this rigour to have DMP plans answer the demands of ISO27001 standard for Information Security Management). Jonathan pointed out that DDI is complex and extensive and it will be good to focus it for biomedical use. In contrast, I noted that Dublin Core is not comprehensive enough and referenced work at DSpace at Cambridge to describe a variety of database artefacts, each with its own internal schema. Jonathan concluded this thread with a brief of description of how OpenXDS is being used to facilitate the exchange of patient health information; and, that metadata schemes need to be supported by Ontologies, for example Snomed, to express a standard terminology for clinical data.
– mitigating the risk: Jennifer asked ‘what can encourage researchers to practice good RDM?’. Stelios: a well regulated data collection regime helps. The ~60 researchers at UCL’s MRC Centre of Epidemiology for Child Health use thin clients with a closed virtual infrastructure to gather and use data. Jonathan expressed the hope that the impending JANET framework for trusted cloud provision will allow similarly secure infrastructures. Anthony cautioned that whatever the regime, individual conduct may remain the most challenging risk factor. Lindsay then questioned again: ‘what can motivate individual researchers?‘ Jonathan made the example of ‘the carrot’ of increased citation as a result of good sharing practice: early Hubble Telescope investigators are still being regularly cited due to re-use of their data ten years after it was first acquired.
To washup, the group resolved to followup this activity by commenting via this blog, and then perhaps by seeking a shared facility to building registers of metadata and ontology use, and policy examples. I suggest a subject specific google doc’ following the form of the programme commonalities example would be one way of achieving this.