VOL: 98, ISSUE: 42, PAGE NO: 30
Carl Thompson, DPhil, RN, is research fellow at the Centre for Evidence-based Nursing, Department of Health Sciences, University of York
Dorothy McCaughan, MSc, RN, is research fellow;Nicky Cullum, PhD, RN, is director;Trevor Sheldon, DSc, is head of department;Pauline Raynor, PhD, RN, RHV, is research fellow, all at the Centre for Evidence-based Nursing, Department of Health Sciences, University of York
An evidence-based culture of health service delivery has evolved during the past 12 years from a series of NHS policy initiatives. At the macro-policy level, the development of national service frameworks, the creation of the National Institute for Clinical Excellence (NICE) and the Commission for Health Improvement are expressions of this culture.
At the level of health care organisations clinical governance seeks to systematically improve the quality of professional decisions in the NHS. At the micro level of individual professional judgement and decision-making, all nurses are, as indeed they have been for the past 30 years, required to give due weight to research-based knowledge.
Evidence-based practice and the development of solid and transparent rationales for decisions are not optional extras for doctors, nurses, allied health professionals and managers in the 21st century NHS.
The evidence-based practice culture
Nursing is in the forefront of efforts to develop a research-conscious workforce. With the shift of nurse training into higher education, the profession is working towards achieving the goal of enabling all nurses to access, appraise and, where appropriate, apply research evidence in practice - in short, to become competent research users.
Alongside the educational response new nursing roles, and the specialist knowledge that accompanies them, are proliferating. For example, a typical selection of nursing roles could include, the clinical nurse specialist running a nurse-led haematology clinic; the health visitor promoting the health needs of asylum seekers; the nurse practitioner triaging patients with undifferentiated diagnoses in primary care settings; or the nurse consultant delivering research-based advice to clinicians on a range of complex chronic wound care problems.
While such new roles offer an exciting opportunity for nursing, they also represent a significant challenge for the nurses involved and those who wish to see the decisions of nurses informed by research knowledge.
The starting point for evidence-based practice
Evidence-based practice involves acknowledging the uncertainty that accompanies making choices (decisions) in clinical practice, followed by seeking, appraising and implementing research-based knowledge that will help reduce such uncertainty (Box 1).
Of course, accepting that the choices you face are accompanied by uncertainty usually means accepting that there may be different and, possibly, more effective methods of care than those currently employed.
Kitson (1999) suggests: ‘Perhaps the most important skill for any health care professional to master in their career is the ability to recognise and handle clinical uncertainty: uncertainty as it is manifested in the range of conditions presented by our patients in wards, outpatients, and in the consulting room; but also about one’s own skill, expertise and knowledge base.’
In clinical practice, asking clinical questions derived from the choices faced by clinicians enables uncertainty to be defined and converted to a form suitable for searching for information. These questions force us to seek relevant, focused information.
The acid test for any piece of information is that it should inform: that is, it should reduce the uncertainty of the person seeking the information. The right kind of design for the right kind of clinical question is a basic tenet of evidence-based practice. We know, for example, that questions concerning effectiveness are generally most reliably informed by the results of randomised controlled trials (RCTs) - or, better still, a systematic review of RCTs. Questions about causation and harm will require evidence from cohort or case-control studies, and questions about feelings and perceptions will require evidence gathered through qualitative techniques.
In making decisions, practitioners are acknowledging that they face a choice between two or more discrete alternatives. When the clinician is uncertain, he or she really has only two options:
- Ignore the uncertainty and let chance, and all the biases that come with unaided decision-making, influence the outcome;
- Acknowledge the uncertainty and find information that will reduce it.
However, not all information is of equal value. Personal experience alone can be a poor guide as to what to do in given situation.
There is a large body of work in the field of decision-making, clinical judgement and cognitive psychology that shows how using unaided recall of experience as the sole basis for decision-making adversely affects outcomes (Jones, 1995; Kahneman and Tversky, 1973; Lichtenstein and Fischoff, 1977).
One such bias is called base rate neglect, where individuals fail to take into account the prevalence of conditions in their patient populations when making decisions. Paying attention to base rates is only possible if the person making the decision has a good knowledge base to draw on - one that accurately records the experience with similar patients - and if they can access this knowledge base reliably.
The problem for clinicians is that it is almost impossible to amass enough experience with some types of patient problems for experience to constitute a knowledge database. The effect of hindsight (Jones, 1995) and other biases also mean it is extremely difficult to access such information in a systematic way to underpin ‘good’ decision-making.
However, in spite of the limitations of knowledge from clinical experience, no professional would be able to function as a decision-maker if they did not make use of it. For many common situations, in which decision-makers get good quality and rapid feedback on the success or failure of their decisions, learning from experience is a powerful and appropriate force for influencing decision-making.
For example, you are a staff nurse with 20 years’ experience on an ear, nose and throat ward providing pain relief - from a limited range of options - for patients post-tonsillectomy (who number several hundred a year). The point of evidence-based approaches to decision-making is to recognise that experience is a necessary source of information for making decisions, but is not sufficient by itself.
Experience can provide a false sense of certainty and is of little help when faced with a situation you have not previously encountered. A good decision from an evidence-based perspective is one that successfully integrates the following four elements:
- Professional expertise (‘know-how knowledge’);
- The available resources;
- The patient’s (informed) values;
- The research knowledge (‘know-what’ knowledge).
Consider the following: the research evidence suggests that hydrocolloid dressings (costing £2.25 each) are no more effective than knitted viscose dressings (costing 25p each) for healing venous leg ulcers when used with compression (Nelson et al, 2000).
On the face of it, the expensive hydrocolloid dressing seems a bad clinical choice. However, the 50-year-old patient is a teacher’s assistant who prefers to use these dressings because she can apply them herself under compression stockings. This means that she can carry on working without having to wait for a district nurse and no longer feels like an invalid - she has a semblance of normality in her life.
Your clinical expertise comes into play here, as you have watched her apply the dressing and asked her to report any changes in condition (smell, size, pain and so on) to you as soon as possible. You are aware that while the hydrocolloid dressing is more expensive on a unit cost basis a district nurse’s visit to apply a simple non-adherent dressing - that is no more or less effective - costs considerably more.
The point of this example is to show that clinical decisions are never simply about the evidence. A good clinical decision is one that balances the research, patient preferences and resource awareness with clinical expertise.
What do we know about health care decisions?
The starting point for evidence-based approaches to health care are rooted in the decisions made by the professionals delivering the service. While there is a sizeable body of work describing decision-making processes in nursing, little research has examined the types of decisions nurses make, which makes it difficult to establish whether existing research evidence is fit for the purposes of reducing clinical uncertainty.
Decision-making in nursing is described using a variety of largely interchangeable terms: ‘clinical decision-making’ is the most common; other terms include ‘clinical judgement’; ‘clinical inference’; ‘clinical reasoning’; and ‘diagnostic reasoning’.
Each term describes the choice of a discrete option from a range available to the clinician. Much of the existing literature does not address the issue of decision-making, but rather deals with the much vaguer idea of using research (Estabrooks, 1999). This does not always involve a choice on the part of the clinician and is not explicitly geared towards the integration of experience, resources, patient values and research knowledge.
Much of the research on decision-making and/or using research in nursing is plagued by small sample sizes (Thompson and Sutton, 1985) and poor response rates (Bostrom and Suter, 1993). Many studies use self-reporting survey methods (Funk et al, 1991; Parahoo 2000; Rodgers, 2000), which means they are limited by people’s tendency to report that they are doing the things they are supposed to (Covell et al, 1985), and assume a self-awareness of research use.
We know from research involving doctors that the range of clinical decisions, questions and forms of uncertainty are finite (Ely et al, 1999). In medicine, decisions can be classified. They can be diagnostic, part of a treatment or intervention, or they may involve delivering prognostic information. What this means is that for doctors an appropriate, or useful, form of evidence is one that reduces the diagnostic, treatment or prognostic uncertainty associated with decisions.
Nurses’ clinical decisions
In order to appreciate the potential for evidence-based practice in nursing we need to understand the nature of the decisions and uncertainties that nurses face. Having this information to hand helps in the development and implementation of decision support tools (such as guidelines). It also helps us to generate research knowledge that is actually useful for clinical practice.
Researchers at the University of York have been exploring the potential for evidence-based decision-making in nursing by examining how acute care nurses use information when making clinical decisions (Boxes 2-3). More recently, the team has used similar methods of data collection and analysis to explore primary care nurses’ use of research information.
An important consideration in the design of the study was that it should seek to describe the decisions nurses actually make in practice, as opposed to the decisions they say they make. The focus of their decisions (pain relief, ward skill mix, nutrition) were not dissimilar to those reported elsewhere (Benner, 2000; Jinks and Hope, 2000). What is newly reported in this study is the detailed description of the types of decisions taken by the nurses. The decisions the nurses actually made can be broadly classified into six categories (Box 3).
Most decisions were concerned with the effectiveness of treatments or interventions. As such, these could easily be translated into focused clinical questions and used as the foundation for evidence-based searching, appraisal and implementation of change where appropriate.
A wealth of information - but is it used?
Since the publication of the NHS research and development strategy (Department of Health, 1993), which promoted decision-making on the basis of sound research, different ways of presenting research information for clinical practice have proliferated. Organisations such as the international Cochrane Collaboration and the NHS Centre for Reviews and Dissemination are making synthesised research accessible via electronic and printed publications, the internet and organisational intranets.
The development of NHSnet and the National Electronic Library for Health strive to bring better quality research-based information to the workforce, while publications such as Clinical Evidence and Evidence-Based Nursing present synopses of preappraised research to involve clinicians in the interpretation of evidence for clinical practice. What remains unclear, however, is what, if any, impact these ways of presenting information are having on clinical decision-makers.
Studies with physicians (Covell et al, 1985) have shown that individuals claim to access research knowledge via professional sources such as journals, while in reality they are more likely to consult colleagues. It is pertinent to note that the truth behind their information use was only exposed by observational research methods rather than self-reported behaviour.
In over 180 hours of observation of nurses’ real-time decision-making, the only text-based resources accessed were the British National Formulary, nursing and medical notes, and local protocols or guidelines. Researchers witnessed nurses using guidelines or protocols just four times, three of which were on coronary care units, although nurses often believed that they were seeing guidelines being implemented in the practice of others. Many felt they had internalised the main points anyway and so did not need to consult primary documentation.
Assessing text-based resources
In our project (Box 2), an audit of ward-based documents and resources was used to describe the information available to nurses, allowing us to cross-reference the sources referred to during interviews. Publication date, reference to research material, the nature of the resource and the clinical focus were all recorded. Analysis of the material available in the acute clinical areas revealed that only around one-third of the 4,000 documents examined made reference to any kind of research evidence.
Of course, large amounts of material on wards could have had an evidence base; however, establishing its provenance and validity, reliability and applicability was often impossible. Indeed, it was not possible to identify the authorship of nearly 1,000 documents. Much of the material was out of date (the average age of textbooks was 11 years), it was badly organised and there was limited storage space to hold it.
Colleagues more accessible than books
Our main finding was that human sources (colleagues) were more accessible and useful than any one kind of information. Apart from immediate colleagues, the people who were viewed as most accessible were those who had a direct connection with ward life: specialist nurses and link nurses who help to bridge the gap between the clinical nurse specialist’s knowledge base (including key technologies, guidelines and protocols) and the day-to-day running of the ward.
Of course, critically appraising the messages for practice from colleagues is difficult and, as a clinician, you are almost wholly reliant on the colleague also internalising any updates of the research base. We found that clinical nurse specialists often stockpiled research-based materials, had extensive clinical, research and commercial networks to draw on, and personal development strategies that included conferences and seminars.
Moreover, they had responsibility for teaching and disseminating research through link nurse structures. The clinical nurse specialists also had the ability to separate good research from bad, although their critical appraisal abilities varied. The problem for nurses is that the critical appraisal of research has not traditionally been taught well - if at all - and it is possible to assume that because a piece of research is published in a journal and or a guideline that it must (by default) be good.
Usefulness of research information
When we asked nurses about the kinds of information they saw as useful, some clear messages emerged. They did not view textbooks or locally compiled information files (the ubiquitous ward folder) as useful resources. This was interesting, since such materials were very much in evidence on the ward and much effort went into compiling them, often as part of the link nurse role.
More worryingly, the internet, online databases and other library-based resources such as the Cochrane Library were also dismissed as useful sources for practice. It was clear that library and support services (which should enable nurses to make the most of the resources available at each site) were perceived as poorly developed.
Nurses were positive about the contribution research could make to practice, but they found it difficult to use it to inform decision-making because of the way in which it is presented. Text-based and electronic sources are not currently much use for nurses engaged in making real decisions in real time and in real clinical practice.
The data from our study revealed that a number of factors hindered the use of research to varying degrees, including: the characteristics of individual nurses (such as a lack of confidence in interpreting and applying research evidence); the organisation (support for implementation, the cultural impact of colleagues); the information available (the amount and nature of statistical material and the language used); and the environment (workload, timescales, personal commitment required).
The major contribution of our study, and what has made it original, is its examination of research use in the context of real-time clinical decision-making, in a period when it is more socially and professionally desirable than ever to be seen to be engaging with research evidence.
Knowledge of the clinical decisions nurses make should enable researchers to focus on areas of nursing care that are of primary relevance to practitioners. It should enable them to respond to the needs of clinicians by providing systematic reviews of evidence where it exists. New research can also be commissioned to fill the evidence gaps.
A greater awareness of the decisions that clinicians face every day in practice should help researchers and commissioners of research to prioritise the questions that need to be answered first.
Our results suggest there is potential in exploring how we can maximise the impact that influential nurse colleagues, such as clinical nurse specialists, can have as opinion leaders and sources of research-based information (Thompson et al, 2001a; Thompson et al, 2001b; McCaughan et al, 2002). These nurses, who in effect act as information intermediaries, should be the focus of targeted attempts to increase critical appraisal, searching, and the other skills associated with evidence-based practice. Link nurses may also have an important role in the dissemination of research, but they must be provided with the resources they need - time, training, organisational and financial support - to carry out this role effectively.
In order to manage the clinical uncertainty encountered in daily decision-making, nurses need to be taught how to rephrase their clinical uncertainties into clinical questions, for which they can then seek, appraise and apply research evidence in practice. One example of how such uncertainty can be represented and used as the basis for evidence-based decision-making is by encouraging nurses to choose a decision they face frequently (for example, what is the best time to give an injection of preoperative low-weight molecular heparin?) and turn it into a clinical question. This could act as the basis for evidence-based searching, appraisal and implementation.
One way of storing the results of these efforts would be as critically appraised topics (CATS), as recommended by Sackett et al (2000). CATS are concise summaries of evidence accompanied by a clinical ‘bottom line’ or take-home message for clinicians. By collecting CATS, a bank of topics can be made available (and updated regularly) to answer frequently occurring questions and help with frequently recurring decisions. The paediatric team at the University of Rochester (2002) offers one example of how CATS can be used.
Knowledge of the kinds of decisions nurses make is also fundamental to the development of decision aids (checklists, decision trees and other support initiatives). It is also a fundamental prerequisite for the development of an evidence base for nursing. The classification of actual decisions (as revealed through observation) into a decision typology provides a framework for the identification and structuring of real - that is clinically significant - questions that require evidence from research.
- The research underpinning this paper was carried out by a team of researchers from the Centre for Evidence-Based Nursing at the University of York, led by Carl Thompson. It was undertaken with financial support from the Department of Health (NHS Research and Development Programme on Methods to Promote the Implementation of Research Findings) and the Medical Research Council.