HDL 2004: Healthcare Digital Libraries Workshop 2004

 

to be held in conjunction with ECDL 2004

 

 

16st September 2004, University of Bath, Bath, UK

 

Workshop Chairs:

Patty Kostkova, City ehealth Research Centre, Institute of Health Sciences, City University, London, UK

Anne Adams, UCL Interaction Centre, UCL, London, UK

 

 

UKOLN University of Bath ECDL2004 banner

 

 

Abstracts and Talks

 

 

 Keynote Speech: Map of Medicine

Mike Stein, Medic-to-Medic Ltd, UK

 Have digital resources taken a wrong turn on the health ‘information journey’?                                                                                        

                                                                                                                                                                                                                 Talk Available
Anne Adams, UCL, London, UK
This paper discusses the health information needs and drivers of UK patients in terms of a health ‘information journey’.  This journey is identified from the experiences of both NHS Direct, who provide information services, and patients themselves. 22 in-depth interviews were conducted with patients and information providers at NHS Direct.  The findings identify a health ‘information journey’ for the patient from ‘initiating information requirements’ to the ‘facilitation of that information’ and the need for ‘contextualized information interpretation’ and mediation.  We also discuss how NHS Direct uses information resources to support patients’ needs by specifying elements of the information journey they do and do not support, and how those needs are supported by other professionals.  The study has shown that patients are often confused about how various resources fit into their journey (e.g. can doctors facilitate information retrieval, will they provide barriers to its interpretation). The findings of this paper support health digital library designers in their development of appropriate systems and support for those systems.
 Deferred Action: Information seeking in the National electronic Library of Infection (DA-NeLI)                                                                    

Talk Available
Nandish Patel, Brunel University, London, UK
Recently, we have witnessed an overwhelming amount of information to be available in digital libraries and, in particular, on the Internet, however, the design of search engines and site-specific navigation interfaces lacks the human knowledge searching perspective. Typically, the technical design does not pay attention to human search strategies. Without better understanding and evaluating how humans search for knowledge, most electronic information on Internet is never retrieved, or is retrieved in a misunderstanding or misinterpreting way. This is of particular concern in the development of live critical applications, such as medical digital libraries. Understanding how healthcare professionals search for knowledge in the National electronic Library of Infection (NeLI) www.neli.org.uk and improving the search and navigation design of this national project applying deferred system design principles is the main aim of this research project
 Evaluating the Online Activity and Searching Behaviour of Users of a Medical Digital Library                                                                    

Talk Available
Patty Kostkova, CeRC, City University, London, UK
The National electronic Library of Infection (NeLI) is a UK-based resource, providing free access to current best-available evidence in Infectious Disease for the public and professionals alike. The content of the NeLI is intended to be developed and maintained by an online community of professionals.  The library is currently running as a pilot website and analysis of the behaviour of current users, in particular professionals in the field of infectious disease is necessary to inform development of the new NeLI website.  A web log analysis was undertaken for the period January 2002 to June 2003.  In particular use by hospital/National Health Service (NHS)-based users was evaluated.  Results indicate an increase in activity during the period and an increase in the number of hospital/NHS-based users.  Hospital/NHS-based users were more likely to return to the site, spend more time on the site and to view more pages than other users.  In addition, a free text search facility was added in July 2003 and an analysis of the queries entered into this search has also informed further library development and will be evaluated in future research.  Further research is now necessary to promote the library among professionals and develop the online community.
 Evolving a Digital Library Environment to the Changing Needs of its Users                                                                                                    

Talk Available
Timothy Miles-Board, University of Southampton, UK

A digital archive, together with its users and its contents, does not exist in isolation - there is a cycle of activities which provides the context for the archive's existence. In arguing for the broadening of the traditional view of digital libraries as merely collections towards the processes of collecting and deploying, we have developed an "extended" digital library environment for orthopaedic surgeons which bridges the gap between the undertaking of experimental work and the dissemination of its results through electronic publication. However, in embracing such an approach, we must also consider that the archive should be able to evolve in accordance with the changing needs of its users | we cannot predict in advance the myriad different types of experiment that future users will want to carry out. This paper therefore discusses our recent efforts in addressing this issue through the implementation of a user-oriented Template Generation Toolkit
 Extraction of biomedical facts - a modular Web server at the EBI (Whatizit)

Dietrich Rebholz- Schuhmann , European Bioinformatics Institute, Cambridge, UK

Information extraction (IE) refers to the identification of facts from natural language text. IE becomes increasingly important in the biomedical field, since online journals provide fast access to relevant facts, where the facts represent completary data to the content of bioinformatics databases. This demand leads into the development of technologies, which allow fast and reliable extraction of such facts from scientific text. The Rebholz group at the EBI provides solutions to the

public to extract facts from Medline and any other submitted scientific text. Such facts are: UniProt, UMLS and GO terminology, identification of gene and protein names, mutations and protein-protein interactions. UniProt, UMLS and GO concepts are automatically linked to the original online

data source. The core technology is implemented in single modules and distributed on a cluster of Linux PCs. Access is allowed either via the ’Whatizit’ interface or through direct access to the Web server. Extracted facts as well as the complete Medline abstract are encoded in XML. Full papers can be processed after submission to the server. For single groups like curation teams individualized modules and solutions are

put to the server and take the terminological resources of the team into consideration. Such solutions deal with the extraction of special facts or with an alerting service, whenever Medline abstracts meet a profile of terminological features. Natural language processing is part of the module for the identification of protein-protein interactions.

 Ontological Principles of Disease Management from Public Health Perspective: a Tuberculosis Case Study                                              

Talk Available
Anjana Roy, CeRC & MIM, City University, London, UK

Formal ontological representation of clinical conditions and disease management is a key methodology ensuring that the complex knowledge of disease treatment, control and prevention can be represented, stored and accessed in the most appropriate way to help the medical professionals in their decision making. This is of particular importance for the public health domain where the concern is about the affect of the disease on populations rather than individuals. 

The existing evidence-based knowledge can best be used by professionals if incorporated into care pathways (formal or informal) which relate the sequence of actions necessary for accurate management of diseases to the progression of the illness and treatment.

Therefore, there is a need for an ontological framework to be built around care pathways in order to allow the professionals to access the most relevant information at the time of making a decision. In  this paper we will illustrate a Tuberculosis (TB) care pathway, as developed at City University, and show how a formal ontological representation can, in principle, serve the needs of information retrieval  around this particular disease
 Readability Assessment of British Internet Information Resources on Diabetes Mellitus Targeting Laypersons                                           

Talk Available
Maged N. Kamel Boulos, School for Health, University of Bath, Bath, UK
Sizeable proportions of Western populations have limited language and math skills making it difficult for them to fully and safely understand and act upon online health information. We used the Flesch Reading Ease and Flesch-Kincaid Grade Level formulae to test the readability of 20 representative written Web documents from 15 trustworthy sites/providers of British provenance, providing consumer information on diabetes mellitus. The corresponding average and median UK reading ages of the tested sites/pages were found to be 14.2 and 14.65 years of age respectively (range: 11-16.8 years). The most readable pages in our study (35% of all tested pages from 26.66% of all sites/providers under evaluation) had a Flesch Reading Ease score above 60 (range: 60.4-73.3) and a Flesch-Kincaid Grade Level of 8.0 or lower (range: 6.0-8.0) corresponding to a UK reading age of 11-13 years. These readability levels, even for the most readable pages in our study, are well above the estimated reading age of the UK population in general (nine years). These results echo those of similar studies conducted over the past few years, and point to a chronic and often forgotten information quality problem. Though quick and easy to use, and certainly better than nothing, readability formulae are not without their limitations. These should be recognised, and any health information quality benchmarking scheme or checklist in use should include (or be supplemented with) some explicit testing for content readability. Online health information providers need to ensure that their written material is pitched at a level appropriate to their intended lay audience. Providers should also consider other means of delivering health education and promotion programmes like plain English videos and face-to-face education, so that no one is left behind
 Healthcare Information Giving Services: Technologies and Everyday Practicalities                                                                                  

Talk Available
Roger Slack, University of Edinburgh, Edinburgh, UK
This paper presents findings from observational studies of work practice in two ‘information giving’ services: Scottish Poisons Information Bureau and The Mental Health Helpline as a precursor to informing the design of such services. Our work highlights the interactions that constitute the requesting and giving of information and the role of intermediaries in the delivery of recipient-designed information. We propose a shift of focus from the logic of information in system design to one that encompasses the practicalities of information giving.
 E-Community Care: Bridging the Information Access Gap                                                                                                                         

Talk Available
B.R.M.Manning, European Federation for Medical Informatics
This paper outlines a strategic approach to the development of ‘joined-up’ multi-agency, multi-disciplinary Health and Welfare Contact/Advice Centres, whose aims are to provide ‘one-stop-shop’ access to patient/client services ranging from direct ‘face-to-face’ contact to full e-Government enabled functionality. Limitations of current Customer Relationship Management [CRM] systems design as well as issues of information content management and patient/client dialogues are reviewed in this context. Innovative options to use portal based knowledge content acquisition techniques as the basis for dynamic content/response procedures updating, together with a goal-centred alternative approach to web access based on cognitive process mapping methods to radically improve functionality are discussed
 MeerKat: New generation relational bibliographical reference software for management of information                                                  

overload within Cochrane …. and beyond.                                                                                                                                                Talk Available
Judith Wright, Cochrane Schizophrenia Group, University of Leeds, Leeds, UK

In this paper we describe a relational bibliographical reference database which is used for the management of information within the Cochrane Collaboration. This description is not for 'marketing' purposes but in order to generate debate on a way forward through the citation information overload. MeerKat software is open source and is built in Microsoft Access.  It easily imports tagged text records and merges multiple citations under one record allowing the user to manage not only the citations to all relevant work but also records of each piece of research.  MeerKat can be used to track the use of citations by logging which citations are sent to researchers and recording their   action upon receipt.  It also tracks the inclusion or exclusion of the studies, to which there may be multiple citations, within Cochrane reviews. MeerKat is a next-generation citation manager countering information overload with its capacity to organise one-to-many relationships.  It is open source and is continually evolving. Recently it has had a Study Finder function added (this helps to piece together the sausage [study] from salami publications [citations]) and also a Hard Copy Finder. This makes manual writing of order forms a thing of the past.  Future developments will see inclusion of direct links to PDF and Clustering technology.