The informationist: a prospective uncontrolled study
1 Medical Library, Repatriation General Hospital, Daw Park, 2 Division of Surgery, Repatriation General Hospital, Daw Park, 3 Department of Medicine, Flinders Medical Centre, South Australia
Objective. To determine whether doctors in an Australian tertiary hospital would use an informationist service, and to identify how the service would influence care.
Design. A prospective uncontrolled pilot study July 2002–January 2003.
Setting. A teaching hospital in South Australia.
Study participants. Fourteen doctors working in the selected units.
Intervention. An informationist attended specified medical in-patient ward rounds and clinical meetings in the Respiratory Medicine, Sleep Disorders, and Rheumatology units.
Main outcomes measures. Clinician self-assessed impact of information on a range of outcomes relating to clinical decision-making, clinician education, and avoidance of adverse events.
Results. In 23 weeks, 52 questions were generated by nine of 14 eligible doctors. Forty-eight of 52 (92%) feedback forms were completed, indicating an average of 5.7 impacted outcomes per response. Twenty-five of 48 (52%) provided new information to doctors, and 24/48 (50%) provided at least some information that could be used immediately. Most common contributions of the service to patient care were revision of treatment plan (21/48, 44%) and confirmation of proposed therapy (18/48, 38%). Thirteen of 48 (27%) contributed to avoiding adverse events, and 10/48 (21%) contributed to avoiding additional tests and procedures. Eleven of 11 (100%) doctors who used the service assessed that it contributed or probably contributed to their professional development, with 8/10 (80%) indicating a similar impact on improving clinical outcomes.
Conclusion. Medical staff will use an informationist service, which contributes substantially to a multiplicity of outcomes relating to medical decision-making, clinician education, and clinical outcomes.
Keywords: MeSH: evidence-based medicine, librarians, medical staff, patient care teams, Non MeSH: clinical medical librarians, clinical questions, informationist
Address reprint requests to Ruth M. Sladek, Australian Centre for Evidence Based Clinical Practice, Department of Medicine, Flinders Medical Centre, Flinders Drive, Bedford Park, South Australia 5042. E-mail: Ruth.Sladek{at}fmc.sa.gov.au
Accepted for publication August 13, 2004.
 |
Introduction |
|---|
 Top Introduction MethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
Accessing scientific evidence at the point of care can be difficult for reasons of time, inadequate access to appropriate information technology, and lack of searching skills. It may be that clinical questions remain unanswered because of these factors, despite an increasing number of sources of information that may potentially contribute to improved decision-making. The synergy of clinician interest in improving access to evidence at the point of care and the historical success of clinical medical librarianship programmes culminated in the call from within the ranks of medicine for an informationist as a new health care professional [1]. The informationist would be a new member of the clinical team, with an explicit role to search, filter, and provide the best evidence for clinical decisions. Their professional background would be in either a clinical discipline or information management; however, they would be cross-trained, ensuring mastery of a core curriculum, including an understanding of clinical settings, information management, epidemiology and statistics. Davidoff and Florance [1] conclude that the time has come to introduce informationist projects or pilots and demonstrate meaningful benefits, or else programmes should ‘be given a decent burial’.
Clinical medical librarian services are the most closely aligned with the informationist concept [2]. They encompass a range of activities, but their defining feature is the physical presence of an experienced librarian in various clinical settings, providing case-related information as required and thereby contributing to patient care. No Australian studies of clinical medical librarianship, and only four undocumented instances in Australia have been identified (personal communication). Yet clinical medical librarianship programmes originated in the USA in the early 1970s [3] and more recently, programmes have emerged in the UK as a strategy to support clinical governance [4]. Interest and incidence in the UK appear to be increasing, with the first UK Clinical Librarian Conference attracting 90 delegates in March 2002 [5].
The authors of two recent systematic reviews of clinical medical librarianship all concluded that despite being used and well received by clinicians, the beneficial impact of clinical medical librarianship services on patient care or subsequent impact on practice had been minimally demonstrated, and most research had been descriptive in nature, with a paucity of evaluative research noted [6,7].
In this prospective study, we investigated the informationist role in an Australian acute tertiary hospital, to see whether doctors would use the service, and to identify what outcomes it would impact.
|
Methods |
|---|
|
TopIntroductionMethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
The chief investigator of the study (R.S.) was employed 0.5 full-time equivalent as the informationist. Having formal qualifications in both information management and public health, and extensive experience in teaching, searching, and critical appraisal skills, she met the criteria of one professional pathway for an ‘informationist’ as identified by Davidoff and Florance [1]. She attended 54 consultant ward rounds (at the bedside) and 22 clinical meetings in Respiratory Medicine, including the Sleep Disorders Unit (during 22 July 2002–17 January 2003), and Rheumatology (during 16 September–8 November 2002) at the Repatriation General Hospital, Daw Park, South Australia as part of this pilot study.
A series of interviewer-administered questionnaires with 40 doctors prior to the pilot determined key features of the informationist service. We offered one type of service only: the provision of a written evidence-based summary of the best research evidence in response to case-related clinical questions about in-patients or outpatients. We deliberately summarized the evidence rather than provide photocopies of the underlying articles, as we thought it more likely that doctors would read up to two pages, especially as they were given the option of asking for the primary papers if required. Drug-specific questions were generally excluded as a clinical pharmacist routinely attended the ward rounds, however, they were occasionally included subject to consultation.
A search protocol based on Haynes 4S Model [8] (see Appendix) and a response protocol were developed. Responses included a summary, fuller details of findings specifying levels of evidence, comments, and references, ideally within two pages. All doctors associated with the departments were eligible to use the service. Where questions arose in group settings, responses were provided to all doctors and medical students present. For other questions, only the requesting clinician received the response. Where drug related, draft responses were discussed with a clinical pharmacist.
Responses were randomly selected and independently audited during the pilot, assessing issues such as timeliness, whether the question was answered, and appropriateness of sources. Requesting clinicians were provided with a response form with each response, to assess its impact on various outcomes. Upon completion of the pilot, an anonymous evaluation questionnaire assessing the service overall was forwarded to eligible doctors. These forms were based on a published outcomes measurement toolkit [9] (used with permission) and a version used in the UK [10]. SPSS Version 11.0 was used for data analysis.
|
Results |
|---|
|
TopIntroductionMethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
Table 1 provides a summary of the informationist’s activities. Fourteen doctors were eligible to ask questions. Eighty-five per cent of all questions were from consultants, with 15% from registrars/residents. Ninety-two per cent of questions arose during either ward rounds [38/52 (73%)] or clinical meetings [10/52 (19%)].
|
Five doctors made no requests. Reasons for this were not specifically elicited; however, two of these doctors (one consultant and one registrar/resident) both indicated that they didn’t have any specific questions which needed answering. A third doctor (registrar/resident) indicated that they preferred to ask colleagues and do their own searching.
Nine doctors asked 52 eligible questions related to therapy or management (27/52, 52%), diagnosis (11/52, 21%), aetiology (10/52, 19%), and prognosis (3/52, 6%). The mean number of questions asked was 5.7 per requesting doctor (range: 1–14, median: 4). Table 2 provides examples of these questions. We assume all written responses to these questions were read; however, this was not explicitly confirmed.
|
Forty-eight of 52 (92%) feedback forms were returned. Doctors indicated a mean impact of 5.7 outcomes per response. Twenty-five of the 48 (52%) informationist’s responses provided at least some new information to the requesting doctor, and 24/48 (50%) provided at least some information that could be used immediately. Responses were assessed as contributing to the revision of a treatment plan (21/48, 44%) and confirmation of proposed therapy (18/48, 38%). Thirteen of 48 (27%) contributed to avoiding adverse events, avoiding additional tests and procedures (10/48, 21%), and advice given to patients (8/48, 17%) (Table 3).
|
Twelve of 14 (86%) completed the anonymous evaluation questionnaire (Table 4), with 11/11 (100%) of those who had used the service assessing that it either contributed or probably contributed to their professional development, and improving clinical outcomes 8/10 (80%).
|
Responses to overall impact on specific outcomes such as diagnosis, choice of treatment, choice of tests, choice of drugs, and advice given to patients were less clear, with at least one doctor noting each possible alternative on the five-point Likert scale used. Doctors did not assess the service as overall contributing to avoiding surgery, or decreasing length of stay.
Observations whilst on ward rounds were not recorded other than the following situation, which the informationist thought reflected additional potential uses of the service. (The details were written down and confirmed as accurate by Doctor B.) One response provided information to Doctor A on the evidence to support nebulized antibiotics for bronchiectasis for patient X. The response was given to Doctor B who was uninvolved with the care of that patient. Later, when they were themselves caring for patient Y for whom they were considering aerosol-delivered therapy, the response was discussed at the bedside with the pharmacist, and contributed to the clinical decision-making regarding the therapeutic dosage for this new patient. They also indicated that they would now consider its impact on a third patient, patient Z, whom they were also currently treating. Hence, one response minimally impacted on two doctors and three patients.
Five of 52 (10%) audited responses revealed satisfactory performance against predetermined criteria; however, it was recommended that the search protocol should be revised to explicitly include relevant national guidelines. The authors are also considering other modifications, and a revised protocol is under development.
|
Discussion |
|---|
|
TopIntroductionMethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
In this study we have demonstrated that clinicians will use a clinical informationist: 52 clinical questions were asked in this pilot, and the clinical information provided was perceived to improve care, with 8/10 (80%) doctors who used the service assessing that it contributed to improving clinical outcomes. Several issues arose which require further consideration: who asked the questions and in which settings, the explicit role of the informationist, and issues relating to outcomes measured. Certain difficulties and directions for further research were also noted.
Consultants asked the most questions, and nearly all whilst the informationist was present on ward rounds or in clinical meetings. This supported our findings from our phase I interviews with doctors, reported elsewhere, that engaging consultants may be a key element when introducing such strategies [11].
Whilst the informationist’s time spent on ward rounds rarely produced a large number of questions, doctors infrequently accessed the service in her absence. This supports the value of having an informationist as part of the health care team. It may also be that over time the use of the service could become entrenched so that more questions would still be pursued when the informationist was not present. Consideration should be given to how to integrate future services in ways that the informationist has sufficient presence to be asked questions, whilst maximizing use of time. This productivity issue is not new. Lewis [12] concluded that ‘it could therefore be a matter of striking the right balance between making the librarian accessible and time spent in the clinical setting’.
The informationist contributed to outcomes relating to decision-making and clinical education, consistent with the published literature. Direct comparisons are problematic because previous studies do not identify outcomes in any standardized way, and the standard of evaluation is typically poor; however, effects on patient care, management, diagnosis, treatment, and clinician education have all been reported by various authors [6]. Veenstra [13] found that their service provided information that influenced patient care between 40% and 59% of the time. Scura and Davidoff [14] found evidence to support the view that their service affected treatment in 20% of cases, and increased knowledge of clinicians 86% of the time. Guise et al. [15] reported an evaluation using a Likert scale of 1–10 (1 being low and 10 being high), where the average impact on outcomes as assessed by clinicians was >9 for knowledge gained, >9 for overall usefulness, and nearly 9 for improved patient care. In the most extensive evaluation of any clinical medical librarian service, Booth et al. [16] recently reported that users assessed the impact of information provided on direct patient management as 22/42 (52%), and continuing professional development as 17/42 (40%).
Some hospitals such as Repatriation General Hospital have a long history of clinical pharmacists attending ward rounds; however, other studies do not mention whether they similarly have such services. This would predictably influence the workload and impact of any informationist service. The relatively low reported impact on choice of drugs reflects our general exclusion of drug-specific questions, such as those relating to dosage and interactions.
The demonstrated impact on clinician education is important, given the role of teaching hospitals in supporting medical trainees and the lifelong learning needs of health care professionals. Whilst this study’s protocol prescribed that questions be case related, a service where questions could be asked for other reasons, such as teaching or education, would extend this impact. Assessing whether the clinical course for patients changed as a result of the informationist service would be complex, and is an issue shared by all health care intervention evaluations. However, it seems reasonable to assume that an impact on either clinician education or decision-making has substantial potential to either directly or indirectly influence clinical outcomes.
The informationist service contributed to a multiplicity of outcomes. Whilst one diagnostic test may contribute to the immediate issue of diagnosis for one patient, one informationist’s response may impact on the care of not only the immediate patient, but other patients and other doctors.
The informationist noted several difficulties. At times doctors were apologetic when no questions arose, particularly after longer rounds. Despite reassurances that ‘no questions’ was both acceptable and a meaningful finding for the study, this was sometimes awkward. This may not be a bad thing: discomfort may challenge doctors to ask questions. Other practical issues included the informationist’s workload management, for example, seven questions were asked in one day, and none on another. The delineation between queries which the clinical pharmacist or the informationist might answer, was sometimes unclear. Frequent changes of consultants meant that it was difficult to establish the service—as soon as one consultant had gained familiarity with it, another was then rostered onto ward duty. Once established however, it is unlikely that discontinuity of consultants would exert any impact on use.
The informationist contributed to a wide range of outcomes; however, this was always precipitated by the clinician asking the question in the first place. What leads doctors to ask questions, and strategies to encourage them to ask more questions, requires further research. Future research should also address issues of cost effectiveness, and the range and structure of services offered by an informationist that would most effectively contribute to the desired outcomes. Consideration should be given to extending the service to other clinical staff, for example, nursing and allied health staff. These were excluded from the scope of this pilot study; however, given the multi-disciplinary nature of health care, it is important that all decision-makers have improved access to evidence. Independent evaluation should be designed into future studies.
The search protocol represents a potential strategy to improve searching for answers to clinical questions. Its development and use as a teaching tool in the context of evidence-based practice could usefully be explored.
One of the strengths of this study was the high response rates (feedback forms 92%, and evaluation questionnaires 86%). Compared with many previous papers, ours was designed as a prospective study and not a post-hoc programme evaluation, and we included both a search protocol and response protocol which deliberately integrated an evidence-based approach. We also addressed quality issues by using an independent audit, and routine consultation with pharmacists when appropriate.
This study also has certain limitations. Several factors predispose this study to social desirability bias. These include that the informationist was known to many doctors, the head of the division was a co-investigator, and the chief investigator was the informationist. Self-reporting by clinicians as the evaluation method is also a limitation. Independent evaluation, and a blinded process whereby a third party received and recoded questions and their respective feedback forms, and input the data for analysis, would have reduced the possibility of these biases. The relatively small pool of eligible doctors in only two clinical areas may limit the generalizability of our findings.
In conclusion, this study confirmed that medical staff will use an informationist service in an Australian acute teaching hospital, and that such a service can substantially contribute to a multiplicity of outcomes relating to medical decision-making, clinician education, and avoidance of adverse events. Multiple uses for the same information make the concept of an informationist difficult to dismiss, despite the need to further consider its scope, funding, and cost effectiveness. We consider that it offers an opportunity to explicitly contribute the best available evidence into case-related decision-making, ultimately supporting patient safety and the quality of clinical care.
|
Appendix |
|---|
|
TopIntroductionMethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
Search protocol
|
This protocol is not prescriptive. Depending on the question, any resource may be searched first; however, generally searching progresses from Group A resources through to Group D resources.
|
Acknowledgements |
|---|
|
TopIntroductionMethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
This study was funded by the National Institute of Clinical Studies, Australia’s national agency for closing the gaps between evidence and practice.
|
References |
|---|
|
TopIntroductionMethodsResultsDiscussionAppendixAcknowledgementsReferences |
|---|
- Davidoff F, Florance V. The informationist. Ann Intern Med 2000; 132: 996–998.
[Free Full Text] - Davidoff F. MLA Informationist Taskforce’s web based discussion about the informationist concept [Comments], 9 May 2002. Chicago, Medical Library Association. http://www.mlanet.org/research/informationist/transcript.html Accessed 10 April 2002.
- Cimpl K. Clinical medical librarianship: a review of the literature. Bull Med Libr Assoc 1985; 73: 21–27.[Web of Science][Medline]
- Ward LM, Honeybourne CJ, Harrison J. A clinical librarian can support clinical governance. Br J Clin Governance 2001; 6: 248–251.[CrossRef]
- Ward L, Honeybourne C. The First UK Clinical Librarian Conference. Bibl Med Can 2002; 24: 23–25.
- Winning M, Beverley C. Clinical librarianship: a systematic review of the literature. Health Info Libr J 2003; 20 (Suppl 1): 10–21.
- Wagner KC, Byrd GD. Evaluating the effectiveness of clinical medical librarian programs: a systematic review of the literature. J Med Libr Assoc 2004; 92: 14–33.[Web of Science][Medline]
- Haynes R. Of studies, syntheses, synopses, and systems: the "4S" evolution of services for finding current best evidence. Evid Based Med 2001; 6: 36–38.
[Free Full Text] - Urquart CJ, Hepworth JB. The Value of Information Services to Clinicians: A Toolkit for Measurement. Aberystwyth: University of Wales, 1995.
- University Hospitals of Leicester, UK: http://www.le.ac.uk/li/lgh/library/clhow.htm Accessed 5 September 2003.
- Sladek RM, Pinnock C, Phillips PA. The informationist in Australia: a feasibility study. Health Info Libr J 2004; 21: 94–101.[CrossRef][Medline]
- Lewis S. An investigation into the viability of implementing the clinical librarian concept at Leicester General Hospital NHS Trust. Masters Dissertation, Loughborough University, 1998.
- Veenstra RJ. Clinical medical librarian impact on patient care: a one year analysis. Bull Med Libr Assoc 1992; 80: 19–22.[Web of Science][Medline]
- Scura G, Davidoff F. Case-related use of the medica literature: clinical librarian services for improving care. J Am Med Assoc 1981; 245: 50–52.
[Abstract/Free Full Text] - Guise NB, Kafantaris SR, Miller MD et al. Clinical medical librarianship: the Vanderbilt experience. Bull Med Libr Assoc 1998; 86: 412–416.[Web of Science][Medline]
- Booth A, Sutton A, Falzon L. Evaluation of the Clinical Librarian Project: University of Leicester NHS Trust. Sheffield: University of Sheffield, 2002.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. P. Gurses and Y. Xiao A Systematic Review of the Literature on Multidisciplinary Rounds to Design Information Technology J. Am. Med. Inform. Assoc., May 1, 2006; 13(3): 267 - 276. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||