Year: 2009
Type: article
Abstract: Attending the Fifth International Evidence Based Library and Information Practice Conference (EBLIP5) in Stockholm in June 2009 provided cause to reflect upon the fitness for purpose of the original five-stage model of evidence-based practice (EBP).1 For so long regarded as the definitive articulation of the EBP process 'the 5A model', as many proponents have come to popularise it, takes you through the five stages of Ask, Acquire, Appraise, Apply and Assess.2 For the first time came the realisation that, as befits its medical origins, the 5A model is conceived and expounded around the concept of the individual self-directed practitioner. True such an individual may subsequently adopt the role of an 'evidence vector', bringing what he or she has found to the attention of a wider patient care team, but so much of the original concept makes assumptions about individualised decision making. Take for example the PICO question formulation which originally related to Patient–Intervention–Comparison–Outcome but which now also commonly refers to Population–Intervention–Comparison–Outcome.3 It is perhaps a testimony to the intuitiveness of the EBP process that it has been able to evolve almost effortlessly from its original patient-encounter orientation into a vehicle for such group-based endeavours as guideline development,4 systematic review production and the commissioning of population-based health services. Stimulated by several EBLIP5 presentations that focused on EBLIP as a group activity, rather than a solitary pursuit, I started mapping out an alternative version of the 5A process—one that is essentially a corporate and collaborative process. A further limitation of the original 5A process is the implication, founded on assumptions of rational decision making, that it involves a logical and sequential procession through five tasks and their subsequent resolution. As Newman comments: 'The 5 stage model of the evidence-based practice process is a deliberate simplification of a complex iterative process.'5 My own dissatisfaction with this model came from a number of sources. First of all work with a PhD student had emphasised that much managerial, and indeed one might speculate, wider professional decision making, does not culminate in a state of completely fulfilled knowing but in a more realistic state of 'satisficing' (i.e. put crudely, getting by with what one knows at this current point in time).6 Instead of proceeding diligently through subsequent stages of the existing 5A process you might decide to 'park' one not particularly fruitful line of evidence quest to revisit your original long list of uncertainties. You may then decide to pursue another line of enquiry. The iterative nature of this Ask–Acquire–Appraise feedback loop, where Appraise refers not to rigorous checklist-supported critical appraisal of one or two key items of evidence but to a rapid assessment or stock take of the collective value of all evidence found at a particular point in time, seems far more apt for complex managerial problems. Here a manager may seek to optimise a decision against multiple competing outcome measures (embodied in the more holistic concept of Evaluation) rather than to emulate a clinician's single-minded pursuit of one primary outcome measure. Furthermore, a similar iterative loop may exist at a later stage of the 5A process. At this point you Apply the evidence and then Assess the subsequent change. What do you do if the effect is suboptimal, negligible or even harmful? In a managerial context you might not simply abandon the original intervention, go back to the Ask stage of the process and start again. More typically you might seek to fine-tune the intervention and/or modify its surrounding environment in order to realise the intended benefits. Such a process, an ongoing Apply–Assess feedback loop, is a fundamental component of Action Learning and Action Research techniques much beloved of managerial consultants.7 In summary then there are two principal limitations of the current 5A model—first it is oriented to individual, not collective, activity; and, second, it seeks to simplify, and thus preserve the integrity of, the entire EBP process making no allowances for the realities of pragmatism and expediency. What might a more collaborative, more iterative model of EBP look like? It is unwise to throw stones at an existing, well-established 'orthodoxy' unless you are able to offer a semblance of an alternative. Below I present an embryonic alternative EBP model. It should be noted that there remains considerable potential to modify this prototype in recognition that it is a 'work in progress'. Furthermore, I have probably oversimplified the entire process in my desire to conform to the brand of a revised '5A model'. In this instance the 5As stand for Articulate, Assemble, Assess, Agree and Adapt. Let us consider each in turn. It is simplistic to assume that a complex managerial situation will yield a single question as in the classic EBP formulating or framing a question. In fairness this is not even the case for a clinical situation, where it is recognised that it is necessary to prioritise and to focus only on those questions that are most important.8 By contrast, a management problem may be more effectively tackled by achieving a wider, more holistic perspective. Within the context of team working and collaboration it is extremely valuable for a team to arrive at a shared understanding of the problem and to articulate this collectively. Each team member has a contribution to make, which itself needs to be valued and carried forward within the decision-making process. Furthermore, EBP involves the reconciliation of user, managerial and professional perspectives. Rather than simply focusing on effectiveness issues which, within the context of managerial interventions, may be equivocal, a team will seek to keep its initial options as broad as is practicable. Furthermore 'articulate' also hints at the complexity of the originating problem in its alternative sense of 'different segments held together by joints'. A major stage of problem definition might be considered to be mapping of the problem territory or domain. Indeed for complex public health problems, in contrast to classically structured health technology pharmaceutical problems, the National Institute for Health and Clinical Excellence (NICE) suggests the usefulness of starting with a so-called 'logic model' to map the nature of the problem.9 Within EBLIP I have similarly proposed the construction of a RESolutions model to enable the mapping of issues against the evidence base.10 Articulating the Problem is therefore a more appropriate designation than Asking the Question. One of the manifest benefits of focusing the question, using PICO or similar frameworks, is to facilitate the subsequent search process. This is particularly the case where multiple elements (e.g. Patient, Intervention and either Outcome or Comparison components) are present—the so-called Foreground Question.3 It is arguable, however, that for many library-based questions it is sufficient to search simply for the combination of the Setting/Perspective (e.g. University/Undergraduate) and the Intervention (Virtual Chat Reference Services).11 In such a case the wide variety of issues raised, evaluation frameworks used and outcomes measured contributes to a wider more informed perspective. Furthermore, the evidence needs to be Assembled, not simply Acquired, because it is derived from multiple sources not just the published literature. An example from Bond University in Australia included routine data from the local library, unpublished survey data from comparator libraries and ad hoc feedback from users.12 Assembly is also a more democratised process in that any of the team may contribute, rather than delegate Acquisition to a single team member. Such evidence needs to be brought together and then placed in meaningful juxtaposition prior to the next stage when its value is Assessed. With evidence originating from a wide variety of sources, many of them not published, it becomes important to make allowances for their heterogeneity. If we apply a strict 'hierarchy of evidence'-based approach to many library problems we find some narrow, possibly non-critical, areas where there is an accumulation of evidence, while for others we contemplate throwing away potentially informative, but low-quality, studies. If, however, we return to our earlier map analogy, we seek instead to place evidence against the different components of a wider overarching problem. Quite quickly we are able to identify areas that have not been addressed and we may elect to conduct supplementary targeted searches to populate 'uncharted' areas of the problem territory. As an example, consideration of the 'problem' of taking professional staff off of reference desk duties yields a significant number of studies. However, when conducting broad searches of the literature for this topic we identify an isolated low-quality study concerning the need to train up paraprofessional or student volunteers as their replacements.13 Had we limited our assessment to only higher quality studies, we would have missed this important issue. Our team might then decide to conduct more purposive searches to identify whether there are similar additional studies, either of equal or, preferably, superior quality. This Articulate–Assemble–Assess iterative feedback loop is an important characteristic of our alternative EBP model. Assessing the evidence for its quantity and quality is a necessary alternative to a model that assumes that studies that may be formally critically appraised will typically exist. For the large majority of library decisions there is no predetermined threshold of evidence at which point a decision to implement is made. Given that library services are human mediated, a significant contributor to the success of any service change is the motivation, involvement and commitment of the team. A model that allows for agreement of subsequent actions is potentially much stronger than a unilateral decision to apply the evidence. Arguably this stage of achieving consensus, and in so doing giving the team members the opportunity to anticipate and rehearse possible objections, is even more important than settling on the most effective intervention. It is frequently the case that the very stage of Agreeing the Actions will lead to adaptation of the implementation. This may result in a difference between the published account of an intervention and how it is applied locally. Unlike a pharmacological intervention which derives its expectation of success from the fact that most of humankind fall within a basic physiological range, libraries show considerable variation with regard to environment, context and institutional values. Deviation from a reported intervention, tailoring it according to a detailed knowledge of the library's clientele, may lead to an enhanced chance of success. By contrast, flexibility in modifying delivery of a pharmaceutical, although frequently performed, is typically seen negatively as a lack of fidelity to the intervention.14 If a library team frequently revisits what they are trying to achieve, fine tuning as they go along, they will probably prove more successful than if they had undertaken a 'back to the drawing board' search for an alternative intervention or mode of service delivery. This Agree–Adapt iterative feedback loop is another important feature of our alternative EBP model. Evidence-based medicine was originally conceived as a process by which an individual clinician might review and enhance the care of an individual patient. Subsequent developments within EBP more widely have led to more collaborative approaches including those targeted at a population rather than an individual patient. EBLIP is very much characterised by the necessity to practice in teams. Furthermore, it must recognise the complexity of library problems and the fact that several steps of iteration may need to be integrated into the overall process. It would be fitting if, in adapting to the agenda so persuasively rehearsed at the EBLIP5 Conference in Stockholm, EBLIP reshapes not only its own future configuration but also the EBP process as conceived by other team-based professions.
Author: Andrew Booth
Institution: University of Sheffield
Language: English
Subfield: General Health Professions
Field: Health Professions
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2006168014
