Lessons

MEDMI has established a successful, cross-disciplinary partnership but several obstacles have been encountered, and lessons learned, throughout the project.

Partnership

  • A considerable amount of coordination between partners has been needed to build a shared understanding of outcomes and goals, requiring regular dialogue and interaction.
  • The role of senior academics has been essential in creating an effective collaboration, ensuring full co-operation and involvement from each institution, and committing staff and data for the project.
  • The basic methodologies that are taken for granted in individual disciplines were sometimes in conflict with one another, and it has been essential to build a common language between epidemiologists, meteorologists and software developers – to make sure the same meaning is assigned to important terminology.

Obtaining health data

  • It has been difficult and time consuming to obtain all health data the project originally hoped to use. Confidentiality, privacy and the licensing of data has meant that accessing databases, especially human health databases, has been challenging.
  • Linking health data to other information, such as spatially resolved environmental data, has been particularly problematic due to the rising risk of identification. The systems in place for data access and release are not designed for this sort of ‘mash-up’ project, causing long delays in accessing data.

Staffing

  • Finding and retaining staffing with the appropriate technical skills has been difficult. Individuals with experience of managing databases of ‘big data’, as well as software development skills for linking big data, are both hard to find and in high demand.
  • The project sought to employ one person to lead on the development of the tool. But the unusual combination of skills required for the project – database management, software development and front end interface – has been particularly difficult to find in one individual.
  • Like most technology, ‘big data’ has moved on since the initial project proposal was written in 2011.

Publishing research

  • Publishing the research and other findings of MEDMI has been difficult since it is a proof of concept activity. Reviewers of journals in the constituent disciplines of the project tend to look on the disciplinary content as lightweight and the cross-disciplinary content as confusing, while the cross-disciplinary journals are less attractive because of their generally lower impact ratings.