Published on 15 November 2021, by Susan Smith
I wonder how many Knowledge & Library Services deal with data science?
There are lots of definitions, but here is an excerpt of one I felt kinship with:
“Data science combines the scientific method, math and statistics, specialized programming, advanced analytics, AI, and even storytelling to uncover and explain the business insights buried in data.” IBM (2020)
Much like our library collections, data needs to be curated and managed. Data sets need to be critically evaluated, analysed, and presented so we can gain knowledge and learning from them. Check out this day in the life video and then consider the similarities with managing evidence! The definitions also state that data science is multidisciplinary. Whilst it might not be obvious why a librarian might consider developing programming skills or advanced analytics, we do need to understand the basics and have a common language so we can work together to improve patient care.
Recently I talked to Quality Nursing Leads about how the library could support them in making data discoverable. They had created an MS Teams page for wards to share falls and pressure-ulcer data and were looking at how to make it searchable by specialty and ward. In an earlier meeting on quality improvement, we had been discussing creating a QI Hub where people could search data, research, lessons learned and patient experience videos. It was the first time any of us found out there was a searchable data portal in the Trust. This led to the question “could knowledge and library specialists help to make date more accessible?” My response: I am a librarian I catalogue things.
Later we were asked if we could do something similar for policies and procedures. This felt more comfortable ground, but it got me thinking. As librarians and knowledge mobilisers we work daily with information, it usually comes in the form of reports and research papers, but information and knowledge can also be presented in terms of dashboards, Statistical Process Control (SPC) charts and river diagrams. I’ve sat in many a meeting where people have referred to evidence in the forms of data and been frustrated it hasn’t included the literature. How many librarians would know how to run a search on gene sequences or other bioinformatics data?
In countries like the USA and China, the role of data librarianship is well established. CILIP now promotes different roles in data science. It wouldn’t be a long stretch for health libraries to catalogue data, teach data searching and appraisal, advise on bias in data collection and curation/stewardship of data libraries like the books and journals we have in our collections now. Some of you are already engaged in this kind of work e.g. in repositories, critical appraisal training, library chat bots or Mersey Care’s Iris.Ai to search open source papers.
Work has begun between Health Education England, Academy for Healthcare Sciences and Manchester University to develop a PG Cert in Clinical Data Science. This will be an exciting multidisciplinary course, teaching the basic data skills mentioned. We already have people who have these skills, or an interest in emerging technologies. Soon we will be launching a new Data Science community of practice. This will help us to identify what work we are already doing, develop case studies to share the learning, and identify future training needs to support extending services.
You might think that you don’t use data science tools, but have you tried the concept-mapping one in EBSCO Discovery? It has a nice little piece of machine-learning code which helps us visualise the information and demonstrates their relationships. Wouldn’t it be good if, when I searched for pressure ulcers, it could return not only an evidence summary and research articles but placed it alongside the data sets and videos of patient experience and lessons learned? We would see it in a wider context, and this would provide us with an opportunity for additional learning. In a shared system this becomes more powerful.
When first researching the topic, I was interested to find the Data Ethics Framework advocates for sharing the learning and was quick to check that the ideas in Releasing The Potential of Learning Health Systems was captured in the Knowledge Mobilisation workstream. I was happy to learn the lead investigator for the report, Tom Foley, had been invited to present as part of #Knowvember2021. I found the paper thought-provoking, inspirational, and scary in equal measure. Although lengthy, it is a very accessible read with some great explanations of tools and concepts which have already helped me build new connections. It provides an excellent model for how Knowledge and Library Services can work with colleagues in our organisations to help transfer knowledge from data into practice and drive quality improvement in the NHS. It is a reminder that NHS X’s Data Saves Lives needs to contextualise the data into a form which can be shared, and we can learn from. I am certainly using the report to leverage my own library service’s development and inform future strategy. Future roles within data sciences will be based on some very old skills developed by librarians, placed in new settings.
Knowledge and Library Services Project Manager (Digital and Data Science Module Development), Health Education England.