Tom Roper's Weblog

Twenty years on

I haven’t used this blog in anger for a very long time but the time seems ripe to revive it. Twitter - I refuse to call it X- is no longer useful. The advertisements are annoying and laughable, and the search function simply doesn’t work properly any more. Unlike some, I can’t say that I have found the tone of Twitter discussions any worse than it was before Elon Musk bought it: the rot, by which I mean the ill-tempered tone of many posts, set in long before that.  I find it hard to become excited about the rehabilitation of various dubious individuals who had previously been banned. No, the death blow for me was the end of free access to Twitter’s API and the imposition of a subscription for TweetDeck. I don’t intend to delete my account, but I use it less and less. 

Some people have decamped to other places: Threads, Mastodon, BlueSky and probably more, but none of these has gained sufficient users to become useful; moreover, they seem to be places where the worthy talk to one another and no one else.    

That leaves this blog. I started it twenty years ago, but have neglected it of late. Let’s see what happens. 

Video, videre, vidi, visum

I have ventured into the world of the YouTube influencer. I am unlikely to make my fortune, as the number of people interested in the finer points of biomedical database searching is finite, but I was asked by Health Education England to make one outlining a method for translating database alerts from NICE’s Healthcare Database Advanced Search (HDAS) syntax to the Ovid native interface, in preparation for the retirement of HDAS next year.

I followed this up with one on translating alerts into the other native interfaces, EBSCO and ProQuest, using the Systematic Review Accelerator projects Polyglot Search Translator tool, and most recently one on spelling errors, and when they should be deliberately included in search strategies.

I have another one in preparation on the use of proximity operators. I had a brilliant idea for another recently, which I didn’t write down and have since forgotten. If there is any aspect of literature searching readers would like me to cover, do comment below. 

Shane Godbolt, 1943-2019

Shane and me at the launch of the North Thames Regional Documents Database

Shane Godbolt died on 24th November 2019. I am not qualified to write a full obituary, but these are some memories.

I first met her in the 1990s, but her fame preceded her: Information Sources in the Medical Sciences, whose 3rd and 4th editions she co-edited with Leslie Morton, was a great help to a newcomer to medical librarianship like me.  At that time she ran the NorthWest Thames regional library system, while I worked as Librarian of North Middlesex Hospital in North East Thames, under the wise direction of John Hewlett. 

Shane’s collaboration with Leslie Morton was important, not just for the books they worked on together, but for the work he started and she continued, building a network of libraries and librarians in the medical education centres of the British Postgraduate Medical Federation. It was fitting that, when Leslie came to be interviewed for the MLA’s oral history project, Shane was the interviewer, along with T. Mark Hodges. It is such a pity that some of the tapes were lost before they could be transcribed, but what survives may be seen here.  

I think the first time I met her was when she gave the 1992 Bishop lecture. Shane was one of the few people to achieve both an FLA by the dissertation route in 1975, and a honorary FCLIP. Her FLA dissertation, the Incomparable Mac, later published as a book, formed the basis of her Bishop lecture which she was invited to deliver (now the Bishop-LeFanu lecture) in 1992. The Mac in the title refers to John MacAlister, who as both the founding secretary of the Royal Society of Medicine and Secretary of the Library Association epitomised the relationship between the professions of medicine and librarianship. Apple, whose British marketing department must have heard the title of the lecture, sent some promotional material, believing that the Incomparable Mac must be one of their computers.

I came to work for Shane in 1995; i had already moved from North-East to North-West Thames, running the medical library at Edgware General Hospital, and the two North Thames regions merged. In those days the internet was a project, and Shane, while she was no technophile, saw the potential. of the internet I was appointed as Senior Project Officer for the internet and regional documents database projects at the newly merged North Thames Regional Library and Information Unit, of which Shane was the overall head. 

It was an extraordinary place to work. There were a hundred and one projects at different stages of progress. A visit to Shane’s office could mean a chance encounter with distinguished foreign visitors from the US Medical Library Association, or the National Library of Medicine, or Finland or India or anywhere in the world. The plates were kept spinning with the help of Susan Fairlamb, her PA, and her husband Paul. 

Shane was always open to new ideas; this was the time of the evidence-based medicine movement. David Sackett delivered  the 1995 Bishop lecture (I still have a cassette recording of it somewhere) and Shane deployed her librarians to the second UK workshop on teaching evidence based medicine. In some portakabins in the grounds of the Whittington hospital at Archway, we rubbed shoulders with some of the founders of EBM and first heard the dread words critical appraisal. 

Shane also had the knack of persuading others to take on tasks without her interlocutor ever realising that they had been persuaded. She invited me to come and talk to the Editorial Board of Health Libraries Review about the internet: I think I may have suggested that they might want to think about making the journal available online. The next thing I knew, I was a member of the Editorial Board. 

I moved on from North Thames unit to the Royal College of Veterinary Surgeons and was not there to see it join with the South to form one unit for the whole of London. I followed Shane’s career, which did not stop for anything as trifling as retirement, and she continued to be a regular and inspiring figure at professional meetings nationally and internationally. 

Tributes will probably concentrate on her international work, for which she was very well-known. Her work for British medical librarianship, though, should not be ignored: she gave the profession our own scholarly journal, she brought together the health libraries of the capital for the first time, and she could be relied upon to listen and offer original and wise advice to anyone who asked. She will be very much missed. 

Text Mining and Machine Learning in Systematic Reviews

A group of librarians from NICE, Public Health England, universities and NHS Library and Knowledge services were privileged to attend a workshop on Text Mining and Machine Learning in Systematic Reviews, led by James Thomas, Professor of Social Research and Policy at the EPPI-Centre. James designed EPPI-Reviewer software to manage all types of literature review, including systematic reviews, meta-analyses, 'narrative' reviews and meta-ethnographies, and leads Cochrane’s Project Transform.

James outlined the problem: we systematically lose research, and then spend a great deal of effort and money on trying to find it again. We need to use correct methods, and, moreover, need to be seen to be correct. There are quantitative issues as well: Cochrane reviewers screen more than 2 million citations a year. Can this considerable human effort be made more manageable by the judicious use of text mining and machine learning? While tools are being developed to help this task, their development is uneven, as is their adoption.

James distinguished between three types of machine learning, rules-based (unfashionable in computer science circles, he warned), unsupervised, and supervised, and gave us opportunities to try out tools based on these approaches using our own devices.

Rules-based approaches are accurate, but fragile – they either work, or fail completely. Unsupervised approaches work by leaving a machine to identify patterns in the data, for example by clustering documents, for example LDAVis based, you don’t need me to tell you, on Latent Dirichlet Allocation.

Supervised approaches require a human or humans to give the machine training data; after a while, from a 280,000 row spreadsheet in an example James quoted, a statistical model can be constructed which can then be used with new material to determine whether or not a study is a randomised controlled trail or not. Training data comes from people, including data generated for other purposes, data created for the project itself and crowd-sourced data, as in the case of Cochrane Crowd, which mobilises Cochrane Citizen Scientists to decide whether or not the subject of a database record is an RCT.

In systematic reviews, these approaches may be used to identify studies by citation screening or classification, to map research activity, and to automate data extraction, including performing Risk of Bias assessment and extraction of statistical data. Readers may be familiar with tools that take a known set of citations, and use word frequency counts, or analysis of phrases and adjacent terms to create word or phrases lists or visualisations. Similarly, term extraction and automatic clustering can be used to do statistical and linguistic analysis on text, for human review, and, if deemed useful, modification of an initial search strategy. Voyant Tools is one example, as are Bibexcel, Termine and even the use of Endnote’s subject bibliography feature to generate lists of keywords.

Citation networks can be used for supplementary searching – will this change, James asked, if or when all bibliographic data becomes open? Useful tools here, apart from traditional ones such as Web of Science, include VosViewer. We also spent some time playing with EPPI-Reviewer , the EPPI-Centre’s own tool for systematic reviewers and with Carrot2 Search.

In the future, James suggested that there is a great deal of interest in a “surveillance” approach to finding evidence, which can automatically identify if a review or some guidance needs updating. Cochrane are developing the Cochrane Evidence Pipeline which aims to triage citations found by machine or crowd-sourced methods can either be triaged by the relevant Cochrane Review Group, or assessed using machine-learning.

While the workshop focussed on systematic reviews, for a jobbing librarian like me in a clinical setting, searches to support systematic review will make up only a small part of the workload. Nevertheless, searches still need to be conducted soundly and rigorously. Can artificial intelligence and machine learning help? Certainly some of the tools James showed are useful when formulating search strategies. A group within London and Kent Surrey and Sussex NHS Libraries is developing a search protocol for the region. We may well find ourselves referencing some of these tools. It is always stimulating to hear a world leader in a field talk, and I’m sure all the workshop participants would join me in thanking both Professor Thomas for giving up his time, and Health Education England for organising the workshop.

The tools James described, and more, may be found on the EPPI-Centre website

See also the National Centre for Text Mining’s page of software tools

For a systematic review on the subject see:

O'Mara-Eves A, Thomas J, McNaught J, Miwa M, Ananiadou S. Using text mining for study identification in systematic reviews: a systematic review of current approaches. Syst Rev. 2015 Jan 14;4:5. doi: 10.1186/2046-4053-4-5.

For a more recent overview, I would recommend Julie Glanville’s chapter on Text Mining for Information Specialists in Paul Levay and Jenny Craven’s new book on systematic searching:

Glanville J. Text mining for information specialists. In: Craven J, Levay P, editors. Systematic searching: practical ideas for improving results. London : Facet Publishing 2018. p.147-169.

In memoriam Humphrey, 2011-2018

Helping me with my Greek irregular verbs

I first met Humphrey at the Cats Protection League shelter. Our previous cat Eustace had died at a ripe old age, and we had decided we could not live in a catless house. My daughter and I went to visit, leaving my wife, who had broken her ankle a few weeks before, in the car. We were shown to an enclosure containing a litter of five kittens. One, a tabby, was already spoken for. There remained two black and white kittens, and two black. Our attention was immediately seized by one of the black and white ones, who was charging up and down a ramp, knocking his siblings flying. One of the black kittens followed him. It was clear that these were the kittens destined to come to live with us, and, a few weeks later, so they did, exchanging their original names of Magpie and Robin for those we gave them, Humphrey and Percy. 

Humphrey established himself as number one cat within a few days. He was fearless, curious, and friendly to humans to a fault. He loved boxes, high places from which to leap on his prey, my armchair, hunting, our neighbours, crossing the road to the flats opposite in order to scare us, ‘helping’ any workmen who came to the house. He held his own in territorial disputes with other cats, notably Moonface, a coarse-haired stray who would stare balefully at Humphrey and Percy through the cat flap, and Esmerelda, a fluffy and buxom female from a house nearby. He could have held his own with Nimrod, the mighty hunter. He more than once caught rats, some of them of monstrous size. 

In a basket; never happier

In February of this year we took both cats to the veterinary practice we attend. We anticipated that Percy, who is fond of his food, would be the focus of the discussion. But the veterinary surgeons examining them felt Humphrey and said she felt a lump. After an exploratory laparotomy, the diagnosis was made: lymphoma. We tried chemotherapy, though it would only have bought him a little more time. The practice were splendid, and everyone took an interest in his case. He was a young, fit cat, and lasted well, but he was losing weight. In the end, we had to take him for his last ride to the surgery.

We have his ashes; we will scatter them in the garden in one of his favourite places. His brother Percy is bereft, as are we. 

On the 70th birthday of the NHS

 

My father and grandfather were both general practitioners in Cambridge, and starch supporters of the NHS. My grandfather, George Roper, practicing first from Regent Street, then from Lensfield Road, died in October 1948. He saw only a few months of the National Health Service, therefore, and had retired from practice, but he was delighted it had come into being. In practice from the early years of the century, and as a Medical Officer for Health, he saw extreme poverty among his patients in the slum parts of Cambridge such as Newtown and Mill Road. He tried to make health more accessible by the simple expedient of ‘losing’ bills when he knew his patient or their family would struggle to pay them. It is reputed that on his death many of these bills were found among his papers, to the amusement of my father, but the fury of other members of the family. He also caused some scandal among his fellow members of the medical profession by joining the Labour Party in the early 1930s, appalled at the conditions in which his patients had to live during the slump.

 

My father began practice after the war, taking over from my grandfather. He qualified from St Georges, did his house job and then served with the Royal Army Medical Corps in the Sudan. He was a staunch defender of the NHS throughout his life, retiring in the 1980s. His career took in advances such the discovery of the link between smoking and lung cancer (a heavy smoker, he stopped as soon as the evidence was known), organ transplantation, child vaccination programmes, fertility treatments and IVF. He died in 1997; I can guess at what he would have made of the NHS in 2018, and suspect I would share his view, that its founding principles, that it should meet the needs of everyone, that it should be free at the point of use, and that it should be based on clinical need, apply as much today as they did seventy years ago. Politicians have come and gone: the principles endure. 

ICML and EAHIL 2017: the short version: That he who runs may read

"That he who runs may read," read the epigraph to a legal reference work, whose monthly updates it was my duty, as a humble Senior Library Assistant in Hendon Reference Library, to interfile. It's a misquotation of Habbakuk Chapter II v 2, which reads in the Authorised Version in full as  ^"And the Lord answered me, and said, Write the vision, and make it plain upon tables, that he may run that readeth it". Be that as it may, the legal publishers intention was clear: they wanted to save the busy lawyer time, as we do with Ranganathan's Fourth Law: save the time of the reader.

I'm also aware that the way this blog is laid out, unchanged since 2004, does not make it easy to read large passages of text. So here are the six things I thought worthy of reporting back to my colleagues after ICML and EAHIL 2017

• What opportunities exist for us to get involved in teaching critical appraisal to patients?
• Evaluating search services: we presented on this, so did Sandra McKeown of Queen's University, Kingston, Ontario: what actions should we now take?
• Text mining tools for searchers: I, at least, really need to understand these, and srart using them systematically
  
• Clinical librarian services: so far we've developed ours in an ad-hoc way.  I was interested in Victoria Treadway's approach, where they conducted a knowledge needs survey, and built the service on that
  
• CADTH have provided some excellent resources for searchers
  
• There's lots of hard thinking to be done about librarian's competencies

 

 

ICML and EAHIL 2017: Friday

The final day came after an exuberant ceilidh the night before, and I found myself in the penultimate papers session on education and learning. First was Richard German of Otago University, New Zealand, on Researching Smarter. an online modular course, designed in-house by the library. Designed to develop research skills, it contained four core topics used across all programmes. It was successful and expanded beyond Medicine to the whole Health Sciences Division Their feedback elicited contradictory views: on videos for example, but they trimmed them from 3-4 minutes to 1-2. Note that online learning didn't reduce workload: the development and testing were very labour intensive.

Sarah Clarke, formerly of the Royal Free, though now working in a Norwegian nursing library, spoke on the Royal Free competency framework. This was timely, with the release of the MLA's competencies earlier this year, and Britain’s  PKSB for Health. There is little health content in the British library school curriculum. In 2009 she developed the first RFH competency model, covering three domains, searching, teaching and evidence-based medicine. In each, the librarian moves from novice to master. After trialling the framework, Sarah found it needed more online learning, and to be mapped to the CILIP PKSB for Health and MLA competencies. The RFH framework is more detailed than PKSB/MLA. Can it work outside university? She used a Delphi technique survey to validate it and found that it would work with the MLA competencies, though not with the PKSB for Health. (I wonder why not). I asked a question about its availability at the end, and Betsy Anagnostelis said that the EAHIL's new TrEDMIL group (Training, Education and Development for Medical and Health Information Specialists and Librarians - a group interested in a European medical librarianship qualification) will be taking the framework forwards and publishing it.

Then John Castle, a medical student, spoke on the James Lind Alliance’s Critical and Appraisal Resource Library (CARL) which aims to allow public to assess treatment claims. There are few freely available resources of quality that teach critical appraisal so CARL will find such resources, evaluate them, and promote their use. John mentioned previous work: Krause LK, Schwartz L, Woloshin S, Oxman A (2011). Interactive tools to teach the public to be critical consumers of information about health care: What's out there? Poster presentation at Cochrane Colloquium and the ECRAN project among them. CARL takes their key concepts from Testing Treatments. Covering 34 concepts in 3 groups: claims, comparisons, choices, CARL will populated with resources from systematic reviews, a snowball web search and systematic searches of MEDLINE and ERIC, categorised by format and user group. Of 500 resources, only 11 pass rigorous evaluation, e.g. Know your chances by Woloshin. Since our library's copy of this has vanished, I was delighted to find it's available as a free e-book. John's fellow speaker, a Norwegian librarian whose name I didn't catch, took us through their search methods for evaluations for resources.

Finally, Gethin White of @DrSteevensLib, urged us not to get fat on full text content. We collect lots of statistics about library resources. Library surveys, he said, correctly in my view, are too often a tool to find the answers you want to hear. They invited 100 stakeholders to 30-45 min interviews and achieved 83% participation! The interview format was hard to keep bias-free. Responses were generally favourable, but highlighted marketing; 60% had never attended library training. ‘Soft’ audiences are easily lost. They then developed self-published lib guides, more focused marketing, and made themselves more accepting of criticism. Gethin stresses it's not all about content, but communication

The EAHIL Assembly found me in a minority of one on the question of opting in or opting out of mailings from suppliers.

In the final parallel session on education and learning Linda Murphy presented a paper already published in the Western Journal of Emergency Medicine:

;Hansoti, Bhakti, Mark I. Langdorf, and Linda S. Murphy. “Discriminating Between Legitimate and Predatory Open Access Journals: Report from the International Federation for Emergency Medicine Research Committee.” Western Journal of Emergency Medicine 17.5 (2016): 497–507.

Linda explained models of subscription and open access publishing.In their paper she and her fellow authors had misidentified a title as predatory; its editor complained. Her search strategy consisted of variants on predatory or hijacked with the words journals or publishing.  Librarians, she said, were not contributing to the literature on predatory journals. Among the signs of a predatory publisher are orthographical and grammatical errors, poor websites, and parasitic titles that mimic reputable ones. Editorial Board members are frequently added without their consent. They may claim legitimate impact factors, or use fake impact factors. Adding insult to injury, they may also ask for payment for the removal of articles. Their officers may claim to be in European or North American city, but in reality are in India or China . She and her fellow authors Identified 150 emergency medicine journals using variety of tools. 98 were hybrid OA, 52 Gold OA, 25 legitimate, 21 predatory, 6 probably legitimate. The US Federal Trade Commission is now suing OMICS and two of their associated publishers. She offered a checklist for identifying predatory websites. Reading her paper, it seems that Jeffrey Beall played a unspecified part; Beall’s methods have been questioned, and in some circles he has been accused of using predatory publishing as a stick with which to beat the open access movement. Certainly his argument that, 'the real goal of the open access movement is to kill off the for-profit publishers and make scholarly publishing a cooperative and socialistic enterprise’ seems a little far fetched. His list no longer exists. 

The last three speakers in this session presented a paper on CPD among health librarians in Manila, an online course at Turku university in Finland, and supporting clinical public health, a paper mostly about curriculum development for an American university. They teach critical appraisal skills based on the JAMA users guides to the medical literature. Student choose an area, such as AIDS-free generation, childhood asthma, obesity, opioid epidemic and have to create a public health action plan, and a wellness and resilience action plan

At the closing ceremony there were very well deserved thanks and flowers to the local organising committee. We were invited to bid to host #ICML2021, the next  International Congress on Medical Librarianship.  The successful bid will be announced by November 2017, and launched at #EAHIL2018 in Cardiff, who invited us to attend with a well-made video including, inter alia, Nye Bevan, sheep and a male voice choir. 

ICML and EAHIL 2017: Thursday afternoon

What I have to say about this session is limited, I fear, as I was presenting myself, and I find it hard to take in other papers when my mind’s on not splitting my trousers on stage. I mean no disrespect to my fellow presenters, and hope they understand. I’ve reconstructed a lot of this from the Twitter archive on Symplur. 

In the first presentation, Danielle Rabb from CADTH, spoke on Single drug vs multiple intervention reviews: trends in systematic searching.  When searching for multiple interventions searches can become quite complicated. They tend to use very broads PICOs. Useful resources from her paper included their grey literature checklist, Grey Matters, the Cochrane Handbook, Cochrane's Methodological Expectations for Cochrane Intervention Reviewsand the European Νetwork for Health Technology Assessment's Guidelines for Information Retrieval. CADTH themselves, as well as Grey Matters, offer other tools for searchers, the Peer Review of Electronic Search Strategies (PRESS) instrument and their filters, all available on their Literature Searching Tools in Support of Systematic Reviews page. 

My paper, co-authored with Igor Brbre and Rachel Playforth, was What do users perceive to be the strengths and weaknesses of librarian- mediated and unmediated evidence/knowledge searches?  We hope to publish, so, very briefly, we surveyed those who had evidence searches from Brighton and Sussex NHS Library and Knowledge Service in 2016, to see what they thought were the strengths of mediated searches. We found that librarians’ searches were perceived to be significantly better for all the attributes we asked about, expect for one, the understanding of the search question. What can we do about this? More attention to the reference interview, perhaps. I set the scene, while Rachel bore the brunt of the presentation. We kept Igor in reserve to answer the awkward questions.  

Wichor Bramer presented Comparing the effectiveness of conceptual search methods: is a fast approach sufficient for the production of (sound) systematic reviews: a prospective, double- blinded, controlled study. Wichor's methods, which allow him to construct a systematic review search strategy in 70 minutes, are controversial, but full marks to him for putting them to the test with his co-authors Melissa Rethlefsen and Margaret Sampson. So far he’s tested his method with three searchers, all of whom missed some relevant papers, but he’s looking for more volunteers to join the study; I’d suggest contacting him on Twitter if you’re interested.  

Finally, Mark Clowes, now working at ScHARR , spoke on What’s the prognosis for health librarianship? The role of the information specialist in a prognostic review. Mark has moved from being a traditional liaison librarian role to an embedded librarian,  a co-producer of research. He told us of his involvement in a Health Technology Assessment in rheumatoid arthritis, which tried to assess which patients were more likely to develop severe symptoms. He used the text mining tool VOS Viewer to look for additional candidate markers of rheumatoid arthritis, which gave some reassurance that major search concepts had been covered, . 

At the end we took part in a panel for questions and answers; I’m afraid panel discussions reduce me to a gibbering state. It’s like a job interview, but in front of a panel of several hundred people. I think I coped with the questions reasonably well, but don’t ask me what they were. There’s no evidence of any gaffes on Tiwtter, at least. We were fortunate in our session chair, Marshall Dozier,  past president of EAHIL, who put us at our ease and encouraged us.  

ICML and EAHIL 2017: Thursday morning: technology, Mark Murphy, research and evidence-based practice

On Thursday morning, after the official conference run, I went to a technology parallel session. One speaker failed to show up, so we heard Michelle Kirkwood on Value of accessing Evidence Based Clinical Reference Content from within the Clinical Workflow. Michelle works for Greater Glasgow and Clyde, the largest health board in Scotland, possibly in the whole of Britain, in the eHealth directorate. Drawing our attention to the Clinical Decision Support (CDS) Roadmap for NHS Scotland and an excellent page of resources prepared by the US Agency for Healthcare Research and Quality  In April 2016 they launched Trakcare, a patient information system with embedded evidence: an evidence icon links out to an external page maintained by NHS Education Scotland, taking them on to evidence resources, with no further login required. It links to local resources, for examples pathways, and a clinical portal brings together records from multiple systems. There are three sets of resources: national subscription services, national (i.e. Scotland) in-house services, and local services. National subscriptions include Dynamed (declaration of interest, they paid for Michelle's attendance at the conference). national (Scotland) resources such as SIGN guidelines, and local resources, for example the Adult Therapeutic Handbook. The library and knowledge service offers search services. It too them twenty-four months to develop a guidelines app, a lesson to those of us who think an app can be put together quickly. To overcome the problem of a large staff turnover, they had to do a lot of awareness-raising, putting it in clinical governance bulletins, on the web and Twitter and running pop-up events as well, of course, as new doctor inductions. Among the lessons learnt were the need to have a consistent approach to use of own devices on network and to provide more content for non-medical staff. Next steps include making it –look the same across all health boards, and implementing in other systems (there are 754 in use!). They want to embed their clinical search and synthesis service bit will need to ration it: it offers a four-hour turn round for searches and is labour intensive. Michelle’s presentation may be seen here: https://librarynetwork.zendesk.com/hc/en-gb/articles/115003714549-What-is-Clinical-Decision-Support

The problem with IPs; How to Manage the IP Ranges Publishers Hold to Authenticate Your Library, presented by Ian Hames of PSI , a fascinating presentation, though ta little outside my area of professional practice — they don’t trust me to have anything to do with authentication these days.

The origins of PSI’s solution lie in subscription fraud problems, for example the abuse of personal subscriptions by institutions to defraud publishers, where a personal subscriber might share their access with others in the insitutition. PSI  helped publishers take down rogue agents and recovered significant sums for publishers. Publishers suspected that fraud was going online, and so PSI started c hecking IP ranges, finding that, while there was little fraud per se, publishers’ housekeeping of IP ranges was poor. Publishers’ records contained many unverified, or outdated, IP ranges.

So they built a database of institutions and their IP ranges. It works on a self-service basis so, libraries can edit the IP ranges associated with their institutions. The service is free to libraries, but publisher pay a subscription. To give us an idea of the scale, Ian mentioned Oxford University,who needed to change IP ranges. This required 400 e-mails, lots of problems, and no one  removed the old IP ranges. There are risks to not updating IP data, chiefly that duplicated or overlapping ranges can make nonsense of usage data. The systems tends to amplify dirty data. There’s also a problem of multiple institutions with the same ranges. While we have DOI for content, ORCID for researchers, there’s  nothing to unambiguously identify institutions. Their Global IP Access Database will launch soon at Global IP Access Database soon at http://theipregistry.org/

In some interesting questions, someone suggested that this could be used to detect fake publishers, while someone else asked if they had found any evidence of association with SciHub, to which the answer was yes, a lot, e.g. identify theft and massive downloads.

Then we were back for a plenary with Dr Mark Murphy,  a GP and lecturer at the Royal College of Surgeons in Ireland (he has a page on the Health Research Board Centre for Primary Care Research too) whose youth and enthusiasm meant that I struggled to keep up with his argument at times.

To cheers and applause, he declared himself a fan of librarians, and took us thro0ugh his library history, using libraries at UCD, Sligo University Hospital, ICGP and the RCSI. What separates doctors from homeopaths, he asked, the answer being evidence. But doctors still don’t implement evidence-based medicine. Quoting Richard Lehman,  whose splendid weekly review of the main medical journals is required reading for all, he said we still train doctors to be ignorant.  We [i.e. doctors-TR]over-estimate treatment effects and risks, for example NSAIDs and cardiac arrest. It’s hard to explain risk on the radio. We suffer from cognitive bias and variation is widespread.  He discussed the  competencies and challenges of modern medicine,  citing @Aoife Lawton’s  review of competencies needed for health librarians Lawton, A. and Burns, J. (2015), A review of competencies needed for health librarians – a comparison of Irish and international practice. Health Info Libr J, 32: 84–94. doi:10.1111/hir.12093

GPs need to practice shared decision making with patients with multi-morbidities.  He was not taught the EBM triad at med school, and papers too often ignore the outcomes that are meaningful to patients. We don’t share evidence well, the linguistic deceit of treatment threatens EBM and we rush into binary thinking. We practice too much medicine — care is often ineffective.

Librarians and GPs have in common that we are both on the outside but with important contributions to make. There’s over-diagnosis and over-detection (e.g. prostate-specific antigen tests, incidentalomas), we treat hypertension as a disease, not as a risk-factor. Dr Murphy quoted some of the authors who had formed his thinking: Petr Skrabanek  and his Follies and Fallacies in Medicine, James McCormack, Margaret McCartney and Iona Heath.

Corporate demands are prioritised over evidence. In ethics primum non nocere is not the most important thing, but patient autonomy. It’s normal to die, and to want to die at home. Ireland, he told us, is the next most litigious part of the world after Florida. Half his patients see him for free, half pay. It’s the main determinant of people’s health and care is fragmented by the cash nexus.

Research methodology is poorly understood, especially qualitative research which is under-valued. Critical appraisal isn’t happening and education is the poor relation. Our education doesn’t give sufficient emphasis to the real-world situation, e.g. multi-morbidity. Politics creates everything

He quoted  Richard Dawkins on the tyranny of the discontinuous mind. Evidence is much abused in politics. Dr Murphy uses Twitter to lobby politicians and journalists. He quoted the example of  Trump’s daughter Ivanka  and her #nationallibraryweek tweet.

He suggested a slogan, how to take advantage of your medical librarian. We improve patient care and there are huge opportunities for librarians in general practice. He (rightly) criticising that Gaiman quotation (Google can bring you back 100,000 answers. A librarian can bring you back the right one ) for underestimating our contribution

His talk provide a lot of questions, for example from Sue Thomas on how to achieve patient autonomy in the 'mess we’re currently in', - he says through high-quality decision aids in the consultation,. He sees 35 patients a day, and needs something that will work with that workload. Richard Lehman and others are putting NNTs (Numbers Needed to Treat) into formats patients can understand. Another question asked about multi professional education. Dr Murphy thought librarians could co-deliver teaching on methodology, critical appraisal. And Andrew Booth, who  asked a question in, I think, every session he attended, asked that if there are heart-sink patients,  are there also heart-sink evidence users?  Dr Murphy answered that yes, there were, and we needed to steer the debate towards the evidence.

Then there was another parallel session, and I attended one on research and evidence-based library and information practice (EBLIP). The first presentation was from Sarah Lewis and Tracey Pratchett on STEP, a set of freely available generic e-leaning modules, developed as part of Knowledge for Healthcare. It began with a phone call in January 2016. Sarah and Tracey secured funding from Health Education England. They had a distributed project team and a twelve-member steering group, and consulted to determine priorities, encourage use and buy-in. They had 139 responses from staff (13% response) and 173 from healthcare staff. Their top concern was how to develop a search when it retrieves too many results. Healthcare staff are not interested in advanced search techniques, e.g. thesaurus use. Both groups thought broadening a search should be a priority.  E-learning needs to be accessible anywhere and interactive. They derived recommendations for structure, design and content from the survey. They had no baseline measurement.  A Virtual Reference Group (I’m one of the 42 members) was set up,  using Yammer, to advise on learning objectives and content.

STEP has seven modules; each is short, taking around 20 minutes and interactive. Some of the lessons they learnt is that it’s hard to do things on a large scale and at a distance. Some found Yammer clunky, difficult to engage with. It’s hard to write e-learning materials. But the modules will launch this summer and they have shared their learning on the Knowledge for Healthcare blog http://kfh.libraryservices.nhs.uk/category/service-transformation/service-transformation-e-learning-project-step/

Next was Sandra McKeown of Queen’s University, Ontario, Canada, on evaluating the quality of a literature searching service, of great interest to me and my fellow delegates from Brighton and Sussex NHD Library and Knowledge Service, Rachel Playforth and Igor Brbe, as that afternoon we were to present a paper entitled What do users perceive to be the strengths and weaknesses of librarian-mediated and unmediated evidence/knowledge searches?

Sandra’s work has been published: McKeown S, Konrad SL, McTavish J, Boyce E. Evaluation of hospital staff's perceived quality of librarian-mediated literature searching services. J Med Libr Assoc. 2017 Apr;105(2):120-131. doi: 10.5195/jmla.2017.201

She and her fellow authors used a Critical Incident Technique to ask users to identify and talk about one experience. They discussed their survey with their library team and with experts, and the responses  improved its design. They tested it to ensure internal consistency, and calculated the sample size needed to provide 95% confidence.

68% had tried their own search. Satisfaction seems to be related to method of request. They found a significant relationship between satisfaction and librarian clarification of question (compare or finding that the one area where search requesters did not consider mediated searches superior was in the understanding of the search question).

They had few patient care requests. People prefer electronic submission, and are generally satisfied with service, but those using a paper form are less satisfied. They wondered if the paper form is not too PICO based, especially if they receive few patient care questions. For future research they want to look more closely at the method of communication. Their results could be subject to recall and response bias. Compare responders with non-responders?

The last paper was about Australian health libraries and how they could help their parent organisations achieve accreditation, the HeLiNS (Health Libraries for the National Standards) Project  Australian hospitals have to meet ten national standards, so they surveyed libraries to see what services they offered to support the standards, for example in governance by  managing guidelines documentation, and creating and maintaining repositories of institutional research. Serious safety and quality incidents have provoked an interest in improving resources

In questions, someone asked if Sandra’s research included searches for course work. No, she answered, they exclude them, as indeed we do at Brighton and Sussex NHS Library and Knowledge Service. Andrew Booth (him again)commented on the high methodological quality of the papers and asked how did presenters gain these skills? Two of the  speakers had  MScs in research or epidemiology, others team had access to an expert. (STEP is not a research project in this sense)All expressed the wish that there was better access to biostatisticians. Another question concerned the Chinese whispers problem of search requests, where the search request reaches the searcher from someone other than the original requester, for example the question  goes from consultant to registrar to junior loss in translation?