Introduction
Evidence-based medicine (EBM) is a concept that has grown rapidly since the publication of an article in JAMA in 1992.1 Sackett et al.2 described EBM as 'the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients'. The aim of this paper is to understand UK doctors' attitudes towards EBM and their self-perceived understanding of specified EBM terms.
Literature review
Previous research investigated doctors' attitudes towards EBM. These publications were retrieved using the search terms perception, attitude, awareness, views, preference, opinion, understanding, self-rating and Evidence Based Medicine in Scopus and Medline. Table 1 reviews the differences and similarities between these published articles, including country the study was based in, data collection tool, year the data were collected, response rate and the specialism of the doctors involved in the research. All of these studies utilised a questionnaire, although only one hosted the survey on the Internet.11 This research has been undertaken in many different countries, which does raise the issue of direct comparisons due to the different medical education systems and types of healthcare provided (insurance-funded compared to publically funded).
Three studies reported responses to the statement that practising EBM improves patient care. The results in two studies were broadly similar with 62.3%7 and 67.8%8 agreeing and 28.8%7 compared to 22.2%8 strongly agreeing with this statement. The Callen, Fennell & McIntosh10 results were lower with 65.9% agreeing/strongly agreeing. There was a higher percentage, 27.4% who were undecided10 compared to 6.8% not sure7 and 8.9% neutral.8 Three questionnaires utilised a visual analogue scale to determine attitudes towards the statement that practising EBM improved patient care. The McColl et al.4 scale range was 0 (strongly disagree) to 100 (strongly agree) with a median response of 70. Poolman et al.12 used the same scale with a mean response of 71. The Sur et al.11 range was 1 (completely disagree) to 10 (completely agree) with a median response of 9. McAlister et al.5 found EBM users agreed with this statement that EBM improves patient outcomes (61.7%) more than did non-EBM users (42%). Research by Ulvenes et al.13 found less than 10% (9.4%) of respondents strongly agreed that EBM improves patient health. More than one in 10 respondents (11.8%) strongly disagreed that it is difficult to use EBM principles in a busy clinical practice.
Veness, Rikard-Bell & Ward7 found 65.3% disagreed or strongly disagreed that EBM is a good concept that fails in practice. Half the respondents (49.7%) did feel that the whole medical information explosion is overwhelming, with a smaller percentage, 38.4%, disagreeing/strongly disagreeing.
Lewis, Urquhart & Rolinson3 found 55% of respondents strongly agreed or agreed with the statement that EBM requires the use of critical appraisal skills to ensure the quality of all the research papers.
Previous research investigated doctors' understanding of specific EBM terms. Table 2 compares the responses for the option understand the term and could explain (it to others) for four EBM terms: publication bias, confidence interval, number needed to treat and systematic review. Several studies have investigated acute-sector doctors' understandings of EBM terms. Responses to the option understand and can explain it to others for the term number needed to treat have generally risen over time 57%,12 46.8%11 and 44.8%9 with a 2003 study of 59% an exception.7 Research in primary practice produced lower responses with only 35.4%4 and 33%10 understanding and able to explain to others.
Young, Glasziou & Ward6 found 16% of respondents believed that they understood and could explain to others the term number needed to treat.
Oliveri, Gluud & Willie-Jorgensen9 asked respondents to evaluate their understanding of EBM terms, but one of these (the relative odds ratio reduction) was a dummy term. However, 4.6% of respondents understood this term and could explain it to others.
Objectives
The three objectives of the research outlined in this paper were:
* To identify the attitudes of doctors in the UK to EBM
* To determine the understanding doctors in the UK have of specified EBM terms
* To compare these results with previous studies undertaken in England to determine if attitudes to EBM have changed
Methodology
The University's Ethical Clearance Checklist was completed prior to the investigation commencing. Consent to participate in the research was implied by the completion of the questionnaire and no incentives were offered for completing the questionnaire.
The questionnaire was derived from questions used in previous published research, either available as a downloadable appendix or received directly from the author(s).4,5,7,11,14,15,16,17,18,19,20 This enabled comparisons from this research to earlier studies. The questionnaire was piloted before the final version was hosted online by Survey Monkey (http://www.surveymonkey.com) from December 2006 to May 2007. The complete questionnaire is included as an appendix (Appendix I), but this paper only considers the results for the sections focusing on EBM (questions 9 and 13).
This paper considers doctors perceived understanding of certain EBM terms. The terms selected were:11,21
* Relative risk/risk ratio - ratio of the risk of an event among an exposed population to the risk among the unexposed
* Absolute risk - the proportion of study participants who experience the harmful outcome in each comparison group
* Systematic review - an explicit method for undertaking a complete literature review and critical appraisal. Appropriate statistical techniques combine valid studies to provide a summary of the medical literature
* Number needed to treat - the number of patients who need to be treated for one of them to benefit
* Confidence interval - quantifies the uncertainty in measurement. A 95% confidence interval is the range of values within which we can be 95% sure that the true value for the whole population lies
* Publication bias - refers to the tendency of researchers, reviewers and editors to submit and accept manuscripts on the direction or strength of the research findings. Positive outcomes are more likely to be published in medical journals than negative
Thirty-six Royal Colleges and British Associations were approached to promote the questionnaire URL. Seven (Royal College of Ophthalmologists, Royal College of Pathologists, British Geriatric Society, British Nuclear Medicine Society, British Association of Audiological Physicians, British Society for Haematology and British Society for Immunology) agreed to promote the URL on their website or via their electronic newsletters.
Potential participants were emailed directly with a link to the questionnaire. The Internet was searched to locate the email addresses for doctors. Certain sites listed contact details, such as BAPRAS, the British Association of Plastic, Reconstructive and Aesthetic Surgeons (http://www.bapras.org.uk/), Expert Search (http://www.expertsearch.co.uk/), BUPA Hospitals (http://www.bupahospitals.co.uk/), Warwick Medical School (http://www2.warwick.ac.uk/fac/med/) and general practitioners (GPs) from a list of all NHS Primary Care Trusts.
The response rate for this questionnaire is difficult to determine exactly as the questionnaire was not just sent out as a direct email. However, there were in total 2351 emails delivered and 636 questionnaires were completed, representing an approximate response rate of 27%.
The questionnaire data were analysed using Excel. Further analysis of the data was undertaken, by comparing responses from GPs and acute-sector doctors and comparing results according to the number of years since medical school graduation.
Respondents were asked to select their medical specialism from a drop-down menu. A total of 636 questionnaires were completed; 353 responses were from acute-sector doctors and 256 from GPs. Sixteen Public Health doctors were not included as their focus is population events, rather than individual patient care. Those responders who selected other and then entered their own option were reviewed; 11 were not considered in the comparisons as their place of work (acute, general practice or other) could not be ascertained. Appendix II provides the breakdown of responses to the second question on medical specialism.
Results
Respondents were asked which year they graduated from medical school. The three decades between 1970 and 1999 account for 89% of the responses, whilst 36% of respondents qualified since the introduction of EBM and personal computers (1990 to 2007). On average the respondents had graduated from medical school 21 years ago. This relatively high figure suggests that the majority of respondents were not junior doctors.
Attitudes towards EBM
The respondents were asked their views and attitudes towards EBM. Respondents were asked whether they agreed or disagreed with specific statements. The results for this question are shown in Table 3.
Respondents were positive about the statements on requiring critical appraisal skills (85.9%) and improving patient outcomes (72.3%). However, 63.6% felt that busy doctors do not have time to practise EBM and 54.4% were overwhelmed by the medical information 'explosion'.
Nearly a third, 31.6%, of respondents were not sure whether EBM is a good concept that fails in practice, whilst 24.7% were not sure if EBM did improve patient outcomes.
The responses to these statements were compared from acute sector-based doctors and those working in general practice. Three statements produced different results for certain statements (Table 4).
The results were compared according to the number of years since graduating from medical school (Fig. 1). Respondents who had graduated from medical school less than 10 years ago were more positive about the practice of EBM and its effects on patient outcomes (82.4% agreed or strongly agreed compared to 66.8% from those who graduated over 21 years ago) and were also more likely to believe that critical appraisal skills were a requirement for EBM (92.9% compared to 81% for those who graduated over 21 years ago). Those who graduated more than 21 years ago were more likely to believe EBM fails in practice (31.6% agreed or strongly agreed compared to 23.5% from those who graduated less than 10 years ago).
EBM terms
Respondents were asked to indicate their reaction to EBM terms. The terms selected were those commonly used in journal papers about EBM, namely, relative risk, absolute risk, systematic review, number need to treat, confidence interval and publication bias. Figure 2 illustrates the responses to these terms.
Over 85% of respondents were either able to explain the EBM term or had some understanding of them. The least understood term was publication bias, but still 87.8% perceived they had some understanding or yes, understand and could explain to others.
Discussion
Attitudes towards EBM
Respondents were generally positive towards the practice of EBM, as nearly three-quarters of respondents selected agree or strongly agree for the statement that EBM improves patient outcomes. Respondents also seemed to be aware of the skills required to undertake EBM as step three is to critically appraise the evidence and 85.9% of respondents agreed or strongly agreed that EBM required critical appraisal skills to assess the quality of research.
Over 60% of doctors felt that busy doctors do not have time to find and critically appraise the relevant research papers. Doctors may be more inclined to access pre-appraised evidence that is more applicable in the patient care setting.
The not sure response can be construed as a 'sit on the fence' response; alternatively this response suggests that the argument for and against the statement has not been decided by these individuals. Nearly one-third of respondents were not sure whether EBM is a good concept that fails in practice, whilst nearly a quarter were not sure if practising EBM improves patient outcomes. These are considerable numbers that have not been persuaded for or against the arguments to practise EBM.
General practitioners were more likely to agree or strongly agree that busy doctors do not have the time to find and critically appraise the relevant research papers (77.4% compared to 54.5% from acute-sector doctors). This may reflect the different work commitments between the two environments. For example, GPs are more likely to be involved with the running and management of their practices, compared to UK acute-sector doctors, who can rely on the support structure of the hospital for more general administrative duties.
Acute-sector doctors were more likely to agree or strongly agree that in most areas of medicine, there is little or no evidence to guide practice (37.2% compared to 21.8% from GPs). This reflects the issue that systematic reviews from resources such as the Cochrane Library cannot effectively address all medical specialisms as the number of topics to cover is vast. These reviews are more likely to cover more generalised conditions and treatments as these have the largest potential audience. Acute-sector doctors were more likely to agree or strongly agree that EBM requires critical appraisal skills to assess the quality of the research (91.6% compared to 77.8% from GPs). This may be due to the promotion in the primary sector of pre-synthesised information ready for use in patient care situations such as NICE guidance (http://guidance.nice.org.uk/) rather than locating the research data which requires critical appraisal, and evaluation, prior to use in patient care.
Respondents who had graduated from medical school recently were more positive about the practice of EBM and its effects on patient outcomes. This is not surprising as those who graduated since 1995 should have been taught, if not at least introduced to, EBM during their training. This background may lead to their more favourable response to the effectiveness of EBM on patient outcomes.
General practitioner results were compared with research undertaken on family physicians in Canada (another publically funded healthcare system) by Tracy et al.8 This research utilised the word positive instead of agree, neutral instead of not sure and negative instead of disagree. Whilst the terms are different the sentiments expressed are similar and therefore basic comparisons are reasonable. The GPs in this research were more likely to agree that there is little or no evidence to guide practice (37.2% compared to 10.8%8). This suggests either UK GPs have a more negative perception about the evidence available or that they lack the necessary skills to adequately and effectively locate the evidence themselves (so are therefore under the impression that the evidence does not exist).
Data from GPs in England were collected by McColl et al.4 To enable comparisons with this research, the responses in this study were converted from strongly agree/agree to positive. Results for the statement that practising EBM improves patient care/outcomes were identical with 70% of respondents in agreement with this statement. There is, however, a difference in wording as McColl et al.4 considered patient care whilst this research asked about patient outcomes.
EBM terms
Nine out of 10 respondents were either able to explain the EBM terms or had some understanding of them (except publication bias). The two terms doctors were most aware of were the 'number needed to treat' (94.4%) and 'systematic review' (94.2%). These are key EBM terms in clinical practice. Jordens et al.22 found that systematic reviews were used more commonly by those doctors familiar with computers. As this research was restricted to computer users only, this may explain the high percentage of doctors aware of this term. The least understood term was publication bias, but still 87.8% perceived they understood the term to some degree. This is probably the most 'academic' term in the selection and therefore would not seem to be the most relevant for practising doctors.
MacLeod and Mant (p. 280)23 suggested that it is 'not necessary to be able to define something to be able to understand it'. This would suggest that 'some understanding' is in fact adequate for using the term in practice. This would suggest that 90% of the respondents would be able to use these terms in practice.
Two EBM terms produced different results, confidence interval and publication bias from GPs and acute-sector doctors. The acute-based doctors understood to some degree, confidence interval (94.5%) and publication bias (92.8%) compared to the GPs understanding of confidence interval (84%) and publication bias (81.2%). Of the six terms identified in this research, these two are probably the most 'academic' and least directly relevant to patient care management. This could explain the approximate 10% difference in response from acute-sector doctors in the more academic hospital settings compared to GP practices. However, research by Upton and Upton found that GPs rated their skills (i.e. their self-perception) lower that acute-sector doctors.24 This may also explain the difference in responses in this research.
Overall, there was a slight decline in the understanding of EBM terms the longer the doctor graduated from medical school. However, even of those respondents who left medical school more than 21 years ago, over 90% felt they understood most of the EBM terms (except publication bias with 85.6%). This illustrates how seriously doctors and their professional bodies view continuing professional development (CPD). For example, the Royal College of Physicians require 50 CPD credits per annum to remain on the professional register (http://www.rcplondon.ac.uk/professional/cpd/CPDforUKphysicians.pdf).
The GP data from this research were directly compared to results from McColl et al.4 The responses to the questionnaire suggest there has been an increase in understanding in all terms since the last research published in this area in 1998.4 This research suggests that GPs in the UK are continuing to develop their knowledge and understanding of EBM terms.
Study limitations
Contact details for doctors were located via the Internet, but this 'hit and miss' strategy missed many potential respondents and many email messages were not deliverable as the contact details were not current. Overall, 25% (733) of the emails sent were not delivered. This was then compounded by a low response rate (27%). The only other research conducted using an online questionnaire by Sur et al.11 also had a low response rate (9%). This response rate means that the sample is not representative and the results may not be relevant to the wider population of UK doctors.
The questionnaire gathered doctors' perceptions of their understanding of EBM terms. Perceptions may not be the same as reality. This research did not investigate doctors' actual practical use of the different EBM terms investigated. Investigating the actual use of EBM terms is not possible using an online questionnaire; such research would need a methodology more directly involved with doctors, such as observation.
Focus groups would have provided the opportunity to identify why doctors expressed the views and attitudes towards EBM they did. This research quantified their views, but the questionnaire format does not allow the reasons or drivers for these responses to be investigated.
Self-assessments are subject to bias. Respondents may overestimate or underestimate their skill level or knowledge in an area, due to overconfidence or a lack of self-esteem.
Another disadvantage is that respondents are a self-selected group who are not likely to be representative, especially in their attitudes towards EBM.
Conclusion
Respondents were generally positive towards the practice of EBM. However, a minority remain unconvinced as to the benefits of practising EBM.
Respondents appeared to be aware of the skills required to practise EBM. The majority of respondents were either able to explain (or had some understanding of) all but one of the EBM terms. There has been a noticeable increase in understanding of EBM terms by GPs in the UK since research published in 1998. These results are self-ratings by doctors so may potentially not reflect the real world situation they experience, but they reflect doctors' perceptions that they have a good understanding of EBM.
Future research using methodologies such as observation could evaluate the actual use of EBM terms by doctors, whilst research using interviews and focus groups could identify the reasons for doctors' attitudes and views towards EBM.
References