What is at the discipline's core?

What is the discipline of health informatics? Professor Graham Wright, director of the Centre for Health Informatics Research and Development (CHIRAD) and Dr Peter Murray, founding fellow, CHIRAD, gave an update of their work to define the knowledge base of health informatics at the BCSHIF meeting in October. Helen Boddy reports. 

When he first tried to put together a health informatics course some years ago, Graham Wright hit a problem: it was difficult to know what should be covered, given that there was no agreed definition of health informatics as a discipline. Now he is part of a team working to define the core knowledge of health informatics, which would be a useful resource for those setting up curricula, and for the industry when defining job specifications and required competencies. 

This work on a cognitive definition is now being funded by the BCS and the International Medical Informatics Association (IMIA) with the aim of creating an international definition of what constitutes 'health informatics'. 

'It will help us show that health informatics has a scientific base,' said Graham. 'And finally settle the arguments of what it should be called.' The IMIA may also use the definition in its current task of formulating its own strategic plan. 

The work began in early 2005 when a group of around 30 academics, practitioners and consultants from the UK, Australia, Canada and USA and from different health informatics domains got together in Otley to produce a framework matrix. They drew up a list of 239 elements or topics that made up the discipline (named ducks) that could be grouped together under 13 subject areas (named ponds) and together make up the framework matrix.

Graham, however, recognised that this matrix is 'just one view of the world' and that it could have gaps in it. Therefore, since the Otley meeting the team has been looking to see if the framework matrix is substantiated by other views of the world and trying to identify possible gaps. 

An initial approach was to see if other groups set the same task came up with similar ponds and ducks. However, when repeated with other groups, there was insufficient time to do the same amount of work and in one case a group even produced a narrower list than that of Otley.

The team has therefore turned to published papers to validate the 'Otley' matrix, assuming that examining authors' words should throw up a list of quintessential health informatics topics. Graham explained that he has been experimenting with different methods of searching through research papers to come up with a list of topics, which could then be cross-referenced against the Otley matrix. 

He has, for example, run different searches using various sorts of discourse analysis software and experimented using web crawlers. He has looked at how searches differ depending on whether searching keywords, MeSH headings or a combination of both. The task is complicated by the fact that the terms are contextualised but searches generally are not.

The team is currently piloting using Reference Manager to search keywords and MeSH headings in Pubmed and Scopus to provide lists of the most popular topics. They are considering whether this is the best method to capture topics. Another possibility, suggested Graham, is to ask volunteers to read through papers and define the key terms from each.

Graham and Peter have gathered together electronic copies of published papers from many sources: social citations, Pubmed and Scopus and conferences. The papers they have at their disposal are therefore not just those available on the internet, and not just peer-reviewed papers. They decided to collect papers up to ten years' old, as terms have changed over the years.

The decision now has to be made whether all the papers collected ought to be searched, or whether a smaller sample will suffice. There is also a question of how to avoid bias. For example, some people publish extensively, often on similar subjects, and they may therefore be over-represented among research papers. 

Once the validation has been completed, the matrix will go forward to a group of international experts so that an international consensus can be reached. A meeting of experts is being planned for Tennessee next year.

More Articles