Doing literature searches for people can feel a little bit like working a sausage machine. We shovel search terms in one end, wait for HDAS to do its magic and then shovel the resulting list of references back to the reader. Often we don’t engage with the material at all beyond weeding out articles in foreign languages or filtering out the systematic reviews.
Evidence synthesis is a different approach. It means taking a deep breath and plunging in to the material, sorting it, getting to grips with it and attempting to provide a structured response for the end user. Over the course of our sessions on evidence synthesis it became apparent that there are almost as many ways of doing this as there are librarians. Some preferred to use a framework, exporting the results into an Excel spreadsheet, rating them for quality and providing a structured response in the form of a table. Others preferred to use a more narrative approach, seeing what story – if anything – emerged from the research and attempting to tell this with the references at the end. Perhaps there’s no right or wrong way – although if we do it in a way we feel more comfortable with and enjoy using I would argue that would lead to a better end product.
Subjects that are amenable to being searched on HDAS make life a lot easier. You can round up all your references in one Word document and – if you’re lucky enough to have access to a bibliographic software package – export them into EndNote or something similar. I was very old-school and actually used pen and paper to make notes on the articles. I found it slowed me down enough to start thinking about the topic and while I was writing my mind was mulling everything over – something I don’t find occurs when I’m just cutting and pasting. At that point it struck me that compiling a narrative evidence synthesis wasn’t that dissimilar to doing an essay at college albeit with COPD and health literacy taking the place of crime in the nineteenth century. As it happened, in this particular instance, the research told a clear story which very much lent itself to a narrative approach but this might not always be the case. For more ‘on the one hand, on the other hand,’ results it might be an idea to group studies by whether they favoured the particular treatment being studied or not, or by the outcomes they measured; both of which allow us to indulge our inner Lawrence Llewellyn-Bowen and do a bit of colour coding, at least in a digital format.
All this takes time of course. If I’d been more pressed I’m sure I’d have made more use of cutting and pasting and foregone my pen and notebook and searches can, of course, expand and contract to fill the time available. However, I suspect few of us will ever end up in the Goldilocks zone of having just enough searches to do a thorough evidence synthesis for each one. In my experience it’s easy to tell which searches are for presentations or journal clubs – “have you got anything, anything at all, on x,y or z”- and searches made for more clinical reasons but even then that might leave us with too much on our plates. There is a real risk that we raise expectations by doing Rolls-Royce searches when we’ve got time only to find we’ve only got enough time to knock up a Morris Minor when we’ve got half-a-dozen to get through at the same time. However, I don’t think this is an adequate reason for not doing them. We’re hoping to share our evidence syntheses where people have given permission for their search results to be shared and it would be great to have a central repository somewhere on the old ‘do once and share’ approach. There’s now a mailing list set up for evidence synthesisers at email@example.com so why not chip in with your own thoughts, share your reviews and pass on a few tips?
Mid Cheshire Hospitals NHS Foundation Trust