Amanda Brookman, Clinical Librarian, Library, Audrey Emerton Building, Brighton and Sussex University Hospitals NHS Trust, The Royal Sussex County Hospital, Eastern Road, Brighton BN2 5BE, UK. E-mail: firstname.lastname@example.org
Background: The Clinical Librarian (CL) Service at Brighton was established in 2003 with the aim of providing high-quality evidence to designated teams and fostering an evidence-based culture.
Objective: To evaluate the CL service at Brighton and discuss the implication of the findings.
Methods: A combination of internally collected data (n = 167), and an external evaluation of the service by questionnaires (n = 86) of users and non-users and interviews (n = 9) of users.
Results: Internal data suggest that the service is valued by its users and that patient care and continuing professional development are the most common uses for searches (confirmed by the external study); that searches generally result in some change in knowledge; and that this knowledge is disseminated. The external study found that visibility of the CL was crucial to the effectiveness of the role and that clinicians used the service mostly to get access to a wider range of resources and/or to save time. Users wanted the CL to include evaluative annotation with the results, and for the CL role to become more embedded in the team. Interview results expanded on the issues of integration of the CL and the need for annotation of results.
Conclusions: To be most effective, CLs would be dedicated to one team, but financial constraints make this unlikely. Alternative working patterns are suggested as a possible compromise.
Clinical Librarianship has a history of around 30 years1 and has been the subject of much research, including three recent systematic reviews.2–4 These have highlighted the need for further, more rigorous, research to demonstrate the impact of clinical librarian services on patient care, to investigate in what format clinicians prefer to have information supplied, and to discuss the effectiveness of various models of clinical librarianship and the cost-effectiveness of such roles.
The Clinical Librarian (henceforth CL) Service at Brighton was established in September 2003, to provide access to high-quality evidence for health professionals in their workplace. Its role is to support evidence-based patient care and foster an evidence-based culture in an attempt to bridge the information gap identified by Smith,5 Ely,6 and Davidoff and Florance,7 and based on recent research which indicated the value of the clinical librarian in supporting clinical governance.8 The service is offered to six teams, three in the local acute trust, Brighton and Sussex University Hospitals NHS Trust (BSUH), and three in the community trust, South Downs Health NHS Trust. It involves the CL interacting with his or her teams on a regular basis by attending ward rounds, journal clubs, and various clinical meetings, picking up on questions related to practice, performing the search, and supplying the sifted results with a cover sheet which refers the user to the search strategy, details exclusion criteria, and highlights the best evidence. The CL emails or posts these reports to the recipients with an evaluation form. He or she also receives requests for evidence by telephone, fax and email, and by written queries placed in clearly marked boxes positioned around the wards served. The CL also provides search-skills training to groups or individuals, provides a current awareness service, and has promoted the service using pens and posters and by giving presentations.
In an attempt to fill some of the needs identified by systematic reviews, the CL has collected data from the outset of the service, to evaluate its impact on patient care and evidence-based practice, and to assess its cost-effectiveness. In addition, an external evaluation of the service was carried out by Lovell,9 with the following objectives:
1to gather additional data on the impact of the CL service on clinicians’ time (clinicians here and henceforth denotes all health professionals);
2to compare the information-gathering behaviour of clinicians who do with those who do not use the CL service;
3to determine the reasons why some clinicians do not use the service;
4to investigate the views of clinicians on what level of information evaluation they would like to receive from the CL and how far they would prefer the CL role to be integrated into the clinical team;
5to provide recommendations on the development of the service.
The fourth objective was intended to investigate the viability of two different approaches to providing CL services, the ‘informationist’ or the ‘ivory tower’ approach. The informationist model set out by Davidoff and Florance7 proposes that an information professional becomes a full clinical team member, and the informationist role is itself modelled on the roles of the pharmacist and hospitalist.10,11 The key points of this model are that the informationist has fundamental skills and knowledge in both information retrieval work and in clinical understanding, and the ability to present relevant clinical evidence in a clinical setting. For the model to be most effective, informationists should be trained through an accredited programme and report directly to the clinical team leaders and be funded from their budgets. The ivory tower approach was described by Greenhalgh12 when comparing two special library services with different styles. In this approach, the information professional is based in the library and awaits requests from clinicians, having relatively little personal contact with them.
This article presents the results from both the internal and external evaluations and discusses their implications for the future development of the CL role.
All literature searches sent out by the CL from the start of the service for the two-year period covered by this study contained an evaluation form to be filled out and returned by the enquirer. The librarian made a record in a Microsoft Access database for each search and input additional data on receipt of completed evaluated forms. A reminder was sent at regular intervals to enquirers who had not returned their evaluation form.
The CL used the collected data to generate reports on:
• impact of the service on the user's practice and patient care;
• other uses for the evidence and its perceived usefulness and importance;
• impact of the evidence on the user's knowledge and the extent to which this would be shared;
• the relevance of the evidence and whether the user considered that it would have been obtainable from elsewhere;
• overall user satisfaction with the service;
• comments about various aspects of the service.
The CL also kept records of the time taken to conduct searches and supply material, and of the resources used.
Sampling—questionnaires and interviews. Both questionnaires and interviews were used in the external evaluation. An Excel spreadsheet was compiled listing contact details for each member of staff and the extent of their use of the CL service. A total of 349 questionnaires were distributed covering all users and potential users for the acute trust (BSUH). Two questionnaires were used, for users and non-users of the service, respectively (available from the authors). Of the 349 distributed, 62 questionnaires were sent to users of the service, and 287 to non-users.
Clinicians with an interest in, and frequent use of, the CL service were contacted to see if they would like to take part in an interview, along with respondents who had expressed a willingness to participate in interviews on the returned questionnaire. A total of nine clinicians were interviewed, including three consultants, three nurses (all senior or specialist nurses), two pharmacists and one registrar.
Questionnaire design and analysis. The questionnaire for users of the service asked clinicians why they used the CL; how they felt about the service; where they found out about it; what kind of service they would like to see in the future; and if they also searched the literature themselves. The second questionnaire was for clinicians who did not use the service and asked respondents why they did not use the CL service; how they went about finding evidence; and what developments of the service could be made to encourage them to use it. Each questionnaire was coded to allow the identification of the clinician it was sent to. Data from returned questionnaires was entered with an identifying number on a separate spreadsheet to preserve the respondents’ anonymity.
Staff from Library Services, eight clinicians (five users of the service and three non-users) and the researcher's supervisor at the University of Brighton piloted the questionnaire. A number of strategies were used to encourage a high response rate, including contacting participants before and after the questionnaire, using personalized letters, and keeping the questionnaire as short as possible.13
Data from completed questionnaires were entered into a spreadsheet that was set up to automatically sum frequencies according to the respondent's (i) frequency of use of the CL, (ii) clinical unit membership and (iii) clinical position. Because of the small sample sizes, and the fact that no hypotheses were being tested, no statistical tests of significance were carried out. The results are therefore indicative in nature.
Interview design. The interviews were designed to explore in more detail some of the issues covered in the questionnaire and allowed open-ended questions to be asked. In particular, interviewees were asked about the impact of the CL, the possible evaluative role of the CL, including what kind of information clinicians would like to receive in the way of digests, summaries, etc., and what future developments of the service clinicians would like to see.
The interview sampling was not representative, but rather focused on more frequent users of the service who could comment on their expectations and experience of using the service. The interviews were semi-structured and allowed the juxtaposition of the qualitative and quantitative data, giving the opportunity to triangulate the data and therefore add to the robustness of the overall procedure.14
Interviews were recorded and the data were coded using a spreadsheet. The coded data were reviewed in detail by the principal investigator, and four emergent themes were identified according to the topics raised. They were: (i) satisfaction with, and use of, the current service; (ii) evaluation and critical appraisal; (iii) widespread support for further integration with clinical units; and (iv) facilitation or an electronic query-reply service.
Internal evaluative data
A total of 411 searches were carried out between September 2003 and August 2005, of which 167 were evaluated by clinicians (giving a response rate of 41%).
What was the service used for? A total of 101 searches were directed at patient care (60% of searches with completed evaluation forms) and, of these, 50 searches resulted in a change of practice (50% of the searches which were directed at patient care). Continuing professional development was the highest use after direct patient care (81 searches, 49% of searches with completed evaluation forms), followed by clinical teaching (55 searches, 33% of searches with completed evaluation forms) (Table 1).
Table 1. Illustrating the differing purposes of clinical librarian searches, and the percentage that either resulted in a change in practice or were rated at a success of 4 or 5 (5 being the most successful)
Purpose of searches
% of searches with completed evaluation forms (n = 167)
No. resulting in a change in practice or rated as 4 or 5
% resulting in a change in practice* or rated as 4 or 5†
Users were asked to rate the success of the search in supporting each use on a scale of 1–5, 5 being the best rating (Table 1). ‘Other’ included (for example) service development, producing patient information, producing guidelines, or preparing a presentation. Percentages of ratings at 4 or 5 for usefulness of searches ranged from 55% (six searches) for audit, to 100% (10 searches) for legal/ethical issues. Eighty-one per cent of searches (65) used for continuing professional development were rated at 4 or 5, while for clinical teaching the figures were 76% (40 searches).
What impact did the service have? Users were asked to rate how important the evidence would be to their practice by selecting the most appropriate answer from: ‘not at all important’, ‘not very important’, ‘important’, or ‘very important’. For 149 (89%) searches the results were rated as being important or very important.
Users were also asked how the evidence from the searches had impacted on their knowledge, choosing from: ‘no impact’, ‘totally new idea’, ‘confirmed prior knowledge’, ‘added to prior knowledge’, ‘refuted prior knowledge’, ‘refuted and gave alternative’, or ‘other.’ The majority of searches with completed evaluation forms resulted in some change of knowledge (Table 2).
Table 2. Figures showing the impact of information resulting from requested searches by the clinical librarian
Impact on knowledge
n = 167 searches with completed evaluation forms.
Totally new idea
Confirmed prior knowledge
Added to prior knowledge
Refuted prior knowledge
Refuted and gave alternative
In order to assess how far knowledge was being disseminated, users were asked for each query if the new knowledge had been circulated to: ‘a few colleagues’, ‘the whole team’, or ‘external professional groups’. In the majority of cases the knowledge was shared with at least a few colleagues and, in a fifth of cases, the knowledge was shared with external groups (Table 3).
Table 3. Figures showing the dissemination of information that resulted from searches by the clinical librarian
Dissemination of knowledge
n = 167 searches with completed evaluation forms.
Shared with a few colleagues
Shared with the whole team
Shared with external professional groups
How satisfied with the service were users? The overall satisfaction rating of the service was that 144 (85%) of searches were given a rating of 4 or 5, on a scale of 1–5, 5 being most satisfied. Eight-seven respondents (52%) answered that they would not have been able to obtain the information from elsewhere, and 91 (55%) that the results were all or mostly relevant.
Comment boxes allowed clinicians to expand on how they used the information obtained and their satisfaction with the service. For example:
• ‘Consideration was being made of a quality of life (QOL) tool in the elderly patients with heart failure. Decided this tool was probably not appropriate from the evidence base. Will continue “to use a disease-specific questionnaire”.’
• ‘What we were doing before was not best practice and there was nothing written down for staff to follow’.
• ‘Information not easily available from elsewhere; 5+ satisfaction.’
• ‘Would have been lengthy search: speeded up process.’
Who used the service? Nurse consultants were the initial contacts in setting up the CL Service in BSUH but overall did not make as many requests (74; 18% of total) as consultants (171; 41%) or nurses (96; 28%). However, as a proportion of the actual number of staff in these categories, the use by nurse consultants, of whom there were three in the teams served, was relatively high, and the use by other nurses was relatively low. This corresponds with the external study's investigations into non-use of the service. Junior doctors made 43 (10%) requests, a quarter of the figure made by consultants, which probably reflects the hierarchical culture rather than need for evidence. Managers and allied health professionals did not make much use of the service; all the requests from the latter staff group came from pharmacists.
A total of 86 of the 349 (25%) clinicians sent the questionnaire responded (34 users, 52 non-users), although not all respondents answered all questions in the questionnaire. The number of users who were sent the questionnaire was 62, therefore the response rate for users reached 55%. This rate was higher than that of the internal evaluation (41%). This is unsurprising; pressures on clinicians’ time would make it less likely that they would return an evaluation form for every search than that they would respond to a single questionnaire. Much research and effort was put into increasing the response rate of the external study, including the provision of a prize draw.
Why do clinicians use/not use the service? The main reasons given by non-users for not using the service was that they were either unaware of it (38%; 15/40 respondents), they never had the time (33%; 13/40) or thought that it did not apply to them (20%; 8/40). Over 90% of the non-users who answered this question were nurses, and so there is no useful data as to why other groups may not use the service—the one consultant who responded to this question replied that s/he was not aware of the service.
Users of the CL service were most likely to have first become aware of it because they saw the CL on the ward (56%; 19/34), and/or it was recommended to them (26%; 9/34). Just one person first became aware of it through the boxes placed on the wards, while two became aware because of library induction. Nobody remarked that they first became aware of the service through the posters or pens advertising the service.
The clinicians who used the service did so mostly to get access to a wider range of resources (79%; 27/34) and/or to save time (65%; 22/34)—often both. Fewer users said that it was because they were not confident in their own searching ability (24%; 8/34), or because they were recommended to use it (21%; 7/34). Only two respondents (6%) said it was out of curiosity, or because of lack of access to a personal computer (PC).
The activity that most respondents stated was most likely to motivate them to use the CL service was his or her presence during journal clubs [26% (22/86) said it might motivate them, 33% (28/86) said it would] and clinical meetings [23% (20/86) might, 35% (30/86) would; Fig. 1]. A number of respondents also said that attendance at ward rounds would motivate them (33%; 28/86), although a relatively high figure (20%; 17/86) specifically stated that presence on ward rounds would not motivate them, making this the most divisive issue as to what is the best use of the CL's time. Attendance at grand rounds was considered the least likely to motivate use of the service [22% (19/86) might motivate, 15% (13/86) would motivate].
What do clinicians use the service for? Users of the CL service use it for a variety of reasons (Fig. 2). The most commonly requested searches related to patient care [41% (14/34) of respondents requested such searches at least once a month] and continuing professional development (35%; 12/34). To a lesser extent, searches relating to teaching (26%; 9/34) and research (15%; 5/34) issues were requested, while the service has also been used for searches relating to case presentations, audits and legal issues.
What level of evaluation and integration would clinicians like? Most respondents stated they would like to see some sort of summary (50%; 43/86) or digest (40%; 34/86) of results, with just 17% (15/86) saying that they would simply like a list of articles (the sum comes to over 100% because some respondents ticked more than one box, with the occasional note stating that different questions need different types of answers). When comparing the response of users and non-users of the service, it revealed that, while users were more likely to want a review and some annotation of the retrieved articles (68%; 23/34), the non-users favoured a complete topic digest (50%; 26/52).
Related to evaluation was the issue of integration of the CL into clinical teams. It is interesting to note that, while only about 10% did not want to see integration (defined as ‘the CL as an embedded member of the team, answering directly to clinical team leaders’), 44% of respondents did, although 42% were ‘unsure’ (some respondents failed to answer this question). More users of the service (50%; 17/34) wanted integration than non-users (40%; 21/52).
What is the searching behaviour of clinicians? Most respondents (including both users and non-users) asked their colleagues when they needed some information [74% (64/86) at least once a month, with 45% (39/86) asking colleagues at least once a week]. Respondents also found information from their own searches [67% (58/86) at least once a month] and from published ‘expert recommendations’ (53%; 46/86). Just one person used the service on average at least once a week.
Over all, 18% (14/79) of respondents (including both users and non-users) said that they did not do any searching themselves, while 54% (43/79) did on average less than 1 h a week, and 24% (19/79) between 1 and 2 h. Therefore, only 4% of respondents (three people) did on average over 2 h per week searching the literature; of these three people, two were non-users of the service, while the third had used the service once. Fifty-six per cent of users said that they did less searching of the literature themselves since using the CL.
Theme 1—satisfaction with, and use of, the current service. Clinicians were generally pleased with the CL and had a good working relationship with him or her. For example, one positive response went:
‘It's changed the way I work. I've become more evidence based in my approach. Because it's easier to ask the questions and get the information and evidence and then move on to thinking about the solution, whereas previously you wouldn't have been able to pick up on all of these questions; you wouldn't have had the time to do that’
Clinicians were asked for examples of how searches they had requested from the CL had changed or contributed to their clinical practice. It was frequently stated that, because the CL was often used to find information for research or for writing guidelines, change of a clinical strategy towards an individual patient was rare. Examples of uses for search results from the CL included:
• guidelines on the use of lactation-stimulating drugs;
• when considering drug ‘off label’ use for rare conditions;
• guidelines on A&E tissue donation;
• retrospective questions on interesting patients after treatment;
• general re-affirmation of current practice;
• for writing literature reviews, etc. for articles to be published;
• general background information for longer term planning projects;
• checking to see if anecdotal evidence could be substantiated.
Theme 2—evaluation and critical appraisal. This topic generated equivocal responses. Often, a reply to whether the CL should evaluate search results would begin ‘in an ideal world, yes.’ For example:
‘It depends on how you look at it. I mean, Yes. The simple answer is yes. If you have a librarian who's qualified enough and a very specific question then to summarize the literature in a way that covers the main topics and conclusions then, yes’
All interviewees agreed that annotation to the results would be useful. This might take the form of listing next to each abstract/reference, notation of what level of evidence it represents (e.g. systematic review, randomized controlled trial), or sample sizes, or if there are is any stated conflict of interest, etc. This would enable the clinician to find the most relevant and robust articles more quickly.
However, before being able to take on a fully evaluative role it was often said that the librarian would need to be tied into working very closely, perhaps being dedicated to, a particular group. Over time, therefore, the librarian would gather better trust, and get to know people and procedures better. For example:
‘… at the moment we don't quite have the trust. It's sort of clinicians saying “well I'm not sure if I quite trust you. It's up to me to dismiss this or that reference rather than anybody else.” I think that trust is probably only going to come through personal involvement, and involvement over a period of time within a specialty’
Those who argued against having the CL evaluating the literature, put forward the argument that to critically appraise the sources it is necessary to have a clinical background and real knowledge of the context of the work. Furthermore, too much of an evaluative role by the librarian may lead to a ‘dulling’ of clinician's skills and clinicians, it was suggested, ought to know how to critically appraise articles themselves.
Theme 3—widespread support for further integration with clinical units. Integration was a topic that was returned to repeatedly during the interviews, and was arrived at in different ways (e.g. Fig. 3).
Clinicians generally like the CL to be physically present as this can help form stronger links to the clinical teams and allow a quicker query response, for example from being on a ward round in the morning, to having relevant papers and answers two hours later, or at least the same day. Then there can be continual feedback, for example:
‘To me, the biggest problem in all of this is having access to the information at the point of when you’re asking the question, and the delay between saying “well this is a critically important question” and then evaluating it and going away and dealing with it. And the problem is because that process can take so long then it gets very removed from what the original problem was.’
Theme 4—facilitation or an electronic query-reply service? The final theme to emerge was whether the librarian should be ‘out there’, among the clinicians on the ward, or if their time would be better spent replying to specific clinical queries via, for example, email. There was support for a more facilitative role, which would include the CL's participation in the whole gamut of clinical enquiry, allowing clinicians to bounce ideas off the librarian. The CL could play a valuable role as a link to the library and remind users of the need for good, structured questions. Better searches are also likely to result from the CL's involvement with the discussion. It was said to be important just to be ‘out there’, for example:
‘In my own experience I think of it (a question) when I see the CL: “Oh, there was a question I was going to ask you.” Whereas if I don't see the CL I tend to end up doing it myself, or it doesn't get done’
However, there was also significant interest in an electronic query-response service. Some clinicians would be interested in simply sending an email to the CL describing a request, while others thought that an online form would be suitable.
Although the total response rates of both the internal evaluations and the external study were below that suggested by Mangione15 to be acceptable for validity, the systematic approach to the collection of the internal data, the higher response rate from users of the service to the external survey, and the triangulation of the results go some way to address the need for robust research highlighted by recent systematic reviews.2–4 The external survey obtained responses from a wider number of service users, as well as surveying the whole pool of possible CL service users in order to assess the impact and potential impact of the service.
The uses of evidence from the searches conducted by the CL, their high evaluative rating, their impact on knowledge of practitioners and how this was shared, suggest that, during the period under discussion, the service supported and fostered evidence-based practice, as the timely dissemination of knowledge based on the latest and best evidence is crucial to this process. These findings are reinforced by those of the external study.
The external study also revealed that the CL service was used primarily because it provided evidence from a wider range of resources than those used by clinicians themselves. To a lesser extent, the service was further used because it saved time, but seldom because the clinicians lacked confidence in their ability to search. This is an issue of quality to consider alongside cost-effectiveness and impact. The external study also assessed the information-seeking behaviour of clinicians and any changes in this following exposure to the CL service. It found that proportionately fewer CL users favoured colleague consultation as a source of information, and that they more often consulted journals and expert recommendation than non-users (data not shown).9 CL users were virtually equally divided as to whether they spent more or less time searching since the inception of the CL service. These figures may suggest that the clinicians developed a willingness to delegate searches, or that they searched more because their awareness of the importance of consulting the evidence had been raised. Time spent on searching by clinicians reported by the external study is still quite low compared with the incidence of information need discovered by earlier research,16 but comments obtained from the interviews and from the internal data suggest that the service has made practice more evidence based.
Data collected on cost-effectiveness cannot be discussed in detail here, but the CL's two-year report17 revealed that about three-tenths of the CL's time has been spent on other library activities, while time spent on searches could result in organizational savings (assuming that the clinicians would have performed the searches and spent the same time doing so). A CL has arguably to remain firmly anchored in the broader library service in order to maintain and develop professional competence, but a reduction in the time spent in ‘routine’ library activities could result in higher organizational savings.
The external study found that there was still a way to go in making all potential users aware of the availability of the service and its applicability to them, particularly nurses, and that some of the promotional methods used by the CL had been unsuccessful. About one-third of non-users cited lack of time as a reason for their non-use, and sometimes even users give lack of time as a reason for not wanting a search carried out. However, this could also be more a case of priorities, as many regular users of the service are very busy, and it represents part of the ongoing challenge of finding the best venues and activities for raising awareness of the existence and value of a CL service.
The crucial question is how to achieve optimum uptake of the CL service. The external survey was inconclusive in its findings on motivators for using it and previous research has indicated that ward rounds are only desirable initially,8 but it is clear that it is important for CLs to be highly visible to their users. There was also a high level of interest in some form of annotation of results, suggesting that saving time by providing some level of appraisal would be a motivator for use of the service. Both of these could be fostered by further integration of the CL, although there was also the suggestion that a remote clinical query answering service would be valuable.
The external study therefore made the following recommendations, based on the data gathered from the interviews and questionnaires:
1that the service should be continued;
2that it should define its target group and users;
3that it should offer annotation and value added results;
4that it should consider an electronic query response service;
5that it should offer a different pattern of working, for example spending whole days dedicated to single units.
Recommendation 5 arose from the questionnaire results and, in particular, the interviews. Firstly, as described, clinicians need to trust the CL before allowing him or her to take on an evaluative role. This trust will mostly come from closer involvement, which in turn would arguably come from eight uninterrupted hours per week (i.e. one working day), than the odd hour here and there spread out during the week. Secondly, the change in working pattern would allow the CL to become fully involved in the clinical process for one day. Questions could be identified during the ward round in the morning, investigated by the CL during the day, and evidence bought to bear by the evening ward round. That would allow continual feedback, and also bring the time between the query and the answer down, and hence make the process more ‘real’ for the clinicians involved—an issue identified in the interviews. It would allow the opportunity for learning, both for the clinicians and the CL. Funding will likely never allow for having a CL for every clinical team, and so the informationist model is probably not viable in the current fiscal environment, but by this compromise it would be possible for a librarian to spend most of each working day with up to five teams, with a minimum of time spent on general library activities.
This would, however, mean sacrificing some useful activities with other teams, such as attending journal clubs or research meetings, when they did not occur on ‘their day’; to a certain extent this is already an issue in Brighton. A less-integrated approach would involve seeking out a mix of best venues and activities, including teaching, for meeting and supporting clinicians from different teams, and not necessarily focusing on ward rounds, although this is thought to be the common factor in describing the CL's role.1,18 This would allow the CL greater flexibility and autonomy in selecting activities with the best potential for achieving visibility, but it could also make it more difficult to develop the level of trust than a more integrated approach.
This research has added systematically collected quantitative data and analytical qualitative data to the evidence of the effectiveness of CL roles, robustly reinforcing the view that they are effective and valued by clinicians.
There is no single model of clinical librarianship which is optimally effective for every setting; the study found support for both the ‘informationist’ and the ‘ivory tower’ models. Funding is not available for the full implementation of the informationist model. Reconfiguration of existing staff to increase the number of CLs is an option, and although it is doubtful whether even this would make the informationist model viable in the NHS, it might allow for the model detailed by recommendation number five, as a greater pool of CLs would overcome potential timetable clashes. Where a service has one or two CLs, the less-integrated approach is more feasible. Whichever approach is adopted, clinicians would value some form of annotation of results. CLs have to balance the need for achieving visibility with the greatest number of people, with the need to produce quick, at least partially appraised results from comprehensive resources.
We should finally like to point out that this study was of a relatively small sample, and there would be a great deal of benefit from carrying out this research on a larger scale to see if the insights reflect those of a larger population.
We would like to thank Ben Skinner, Evidence-Based Knowledge Management Librarian at Brighton and Sussex University Hospitals NHS Trust, for his valuable and constructive criticism during the production of this article.
Implications for Policy
• Clinicians are interested in the clinical librarian service and most would like to see them integrated into the clinical team.
• Trust remains a big issue, and it was suggested that integration with clinical teams would allow this trust to develop.
• The current practice of one clinical librarian being spread thinly across many teams (because of budgetary limitations) requires compromises to be made. Should the clinical librarian dedicate solely to one unit, or try to offer the service to as many people as possible?
Implications for Practice
• It is important for clinical librarians to be seen and physically available on the ward and at other clinical meetings.
• Most clinicians wanted some annotation of results, and/or a brief summary of the key points found. However, this would add to the clinical librarian's workload.