Case studies

Worcestershire Health Libraries have created a publications database to showcase all the published literature from NHS staff in Worcestershire.

The database highlights publications from Worcestershire Acute, Worcestershire Health and Care Trust and GP practices in the county. It links to the PubMed abstract of the articles. The database was created after a meeting with research and development highlighted the fact that no one knew, or was recording the research that was getting published by NHS staff in Worcestershire.

We knew that Birmingham and Solihull Mental Health Library had a publications database so we copied their system which uses our web development software (Easysite) functionality and is driven by a basic Excel spreadsheet.

Automated PubMed searches are run every month and highlight any new research from Worcestershire. We then add the new articles to the database. We publicise the database regularly through our communication channels and pick an article every month to make available in our libraries for staff to come and read.

The system is maintained by library staff and I would love to see it become a repository so that people can access the PDFs of the articles rather than the PubMed abstracts.

In 2007 The Christie Hospital NHS Foundation Trust developed an institutional repository for the discovery of research papers. Initially it was an in-house SQL database using Microsoft Access created by the Trust IT department, with a web version available via their internet pages. It linked out to some full text via Digital Object Identifier (DOI) and was searchable by keyword or author.

The content was populated mainly by library staff sourcing citations via Pubmed and ISI Web of Science imports which were then cleaned up, some items linked out to the full text, some had an abstract and others just had a citation. It was possible to produce reports out of the database to share with the research department. In addition to this Trust staff were encouraged to deposit information about new publications.

The content started in 2008 and they worked back to 1945. However by 2010 the project had outgrown the database and any developments were dependent on the time the Trust IT department had to commit to this project.

In 2010 they moved to a DSpace solution which is open source software with greater functionally than their current system and they bought it as an externally hosted system. They could create communities, continue to import records, but could now easily export to reference management software. This solution was also used by Manchester Metropolitan University so they were able to see it in use.

The Christie paid for the hosted version to remove the need for in house IT support and to manage maintenance and upgrades; however there is an annual cost for this.  As the DSpace solution is open source software it can be hosted internally, if the organisation has the support and infrastructure to make it web based, however this was not the preferred solution for The Christie for the reasons mentioned above.

The DSpace respository is the backend database which uses Dublin Core Metadata standards, these field headings are controlled for inputting consistency; the interface has changed hands a few times during The Christie’s time using DSpace and they currently purchase the interface through Atmire. They do store some full text but only after checking the Sherpa/Romeo site to ensure they are adhering to the permissions allowed by the publisher.

Where full text cannot be stored they link out and they include citations if not full text is available as with their previous solution. The library team continue to manage the updating and sourcing of the research papers.

Derby Teaching Hospitals have an Online Research at Derby Archive (ORDA). Working in partnership with Derbyshire Healthcare NHS Foundation Trust and Chesterfield Royal Hospital NHS Foundation Trust.

ORDA is the official institutional research repository for the NHS in Derbyshire. It captures, stores and preserves the research output and makes it available to the research community through Open Access Protocols.

The aim of the participating Trusts was to:

• make their combined research visible
• create exposure to the work done within the Trusts and to connect with internal and external colleagues
• upload PDFs rather than full-text links

In one month, using posters and social media to advertise, ORDA satisfied 7 inter library loan requests, saving both staff time and money.

The libraries use Sherpa Romeo to check for copyright restrictions and chose DSpace, a small software provider to host the repoistory.

The support provided was excellent but required strong communication and negotiating skills.

The partnership has two ORDA champions who spread the word amongst junior doctors. This has helped with usage. The next step is to get the community health service Trust on board and establish individual profiles for staff.

The Royal Devon and Exeter Research Repository came from a need from the research and development (RD) department to capture accurate publications data for their annual reports, which impacted on their research funding. This previously involved time-consuming and complicated searches on PubMed every year.

The Library approached RD with a list of NHS publication databases and repositories to demonstrate how this data was being captured in other institutions, and set up a trial to the Open Repository Lite DSpace software – a "budget" version of the repository software used in a number of higher education institutions.

RD were impressed with the possibilities of firstly capturing and reporting bibliographic data, and secondly using it to showcase trust research, as a repository would be freely available and searchable online. They did a cost-benefit exercise and given the staff time previously put into creating the reports, agreed to fund the software, for a 3 year fixed price period, with the provision that the repository itself would be managed and administered on a day-to-day basis by library staff – there wasn’t any additional funding for staff in the project.

The site build started in November 2014, working closely with the Information Analyst in RD and building it around data requirements and the divisions and departments structure provided by RD. The repository was initially populated with an initial batch load of data from 2013 and 2014, which had been gathered from previous reports. We have publications data going back to 1984, but given the time it was taking to quality check and format the data for uploading, the decision was made to add just 2 years’ worth of data initially, with the remainder to be added retrospectively.

To capture the current publications data PubMed and Google Scholar alerts (by author affiliation) were set up. Initially we used HDAS alerts too, but this produced a lot duplicates so now we just use PubMed and Google Scholar. New data is being manually added on a weekly basis by library staff. Data is auto-populated by the software if you have the PubMed ID or digital object identifier (DOI), which saves time, and there are usually only a couple of new articles to add every week. At the moment, this is being slotted into the daily work of the Band 5 Reader Services Librarian, as she has previous experience of repositories and Open Access, but will hopefully be cascaded to library assistants in the future.

Given how busy our clinicans are we opted for library staff submission, rather than self-submission. A simple online form set up where staff can submit details of their articles, but it has never been used! However, a lot of staff do engage with ResearchGate, so we may need to re-think the self-submission process.

Promotion

The repository officially launched in March 2015, promoted on the Trust intranet, RD intranet pages, Library website, Twitter (#RDEauthors), talks at Grand Rounds and local research events and following that, monthly inclusion in the RD newsletter. There was some interest from a few keen individuals, who sent us their list of publications, and have also set up researcher profiles.

Repository content

The primary focus is on published, peer-reviewed journal articles and case reports but we also include books and book chapters and conference posters – a portal for any Trust research output. The RD report requirements are the published articles, and they have a filter for what publication types they need (such as, no comments, letters, editorials, and so on), but it is easy to extract and filter the data for them using Excel.

Full-text

We do try to include full-text articles where we can. If an article is already freely available via open access then we upload a PDF copy to the repository (being sure to add the copyright/CC-BY licence statement to the cover sheet). If it isn’t available as open access, we look up the journal title on SHERPA Romeo to check which version of the article we can upload, and if there is an embargo on when we can do that. Then we contact the Royal Devon and Exeter author to see if they still have the appropriate version they can send us. Chasing this up can be a little time-consuming, and often we don’t get a response, so we need to work on this going forward.

Progress

Since the repository was launched in 2015 we have 1129 bibliographic citations, and 192 PDFs. It is being added to each week. RD are enormously happy with the amount of time it has saved when they create their annual report, and have now asked us to create quarterly reports for them too! The repository has had over 11,000 site views (from all across the world) and over 6,000 file downloads, proving that Royal Devon and Exeter Research is now discoverable!

Next steps

We need to continue to market the repository and get researchers on board. We’re continually strengthening our ties with RD and we’re trying to link articles to specific research projects, matching information in the trusts’ research information system, Edge.

We’re also in the last year of our fixed price deal with Atmire, after which the open repository lite software will no longer supported or developed. They are currently developing an alternative product, or else we’d have to go to the full open repository product, which would be much more expensive. We’re going to investigate joining up with the local University repository, or seeing what other products are available that we could transfer our repository too.

Case study provided by Cate Newell Reader Services Librarian at Royal Devon and Exeter Research Repository Manager.