Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (9)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Oswald, N.
Right arrow Articles by Bateman, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Oswald, N.
Right arrow Articles by Bateman, H.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Family Practice Vol. 16, No. 4, 414-419
© Oxford University Press 1999

Applying research evidence to individuals in primary care: a study using non-rheumatic atrial fibrillation

Nigel Oswald and Hilarie Bateman

Primary Care Resource Development Centre, Grey Towers Court, Stokesley Road, Nunthorpe, Middlesbrough TS7 0PN and General Practice and Primary Care Research Unit, University of Cambridge, Cambridge, UK.

Oswald N and Bateman H. Applying research evidence to individuals in primary care: a study using non-rheumatic atrial fibrillation. Family Practice 1999; 16: 414–419.

Received 23 September 1998; Revised 22 February 1999; Accepted 29 March 1999.


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Appendix 1
 References
 
Background. There is strong research evidence that anticoagulation of patients with non-rheumatic atrial fibrillation reduces the morbidity and mortality of stroke. This evidence is incompletely applied.

Objectives. We aimed to report consequences associated with the intention to apply evidence about anticoagulation for non-rheumatic atrial fibrillation (NRAF) to individuals in general medical practice.

Methods. The study involved prospective structured reporting of the processes of applying evidence about NRAF to individual patients in six general practices in Cambridge. The subjects were patients identified to have NRAF in these practices. The intervention consisted of a practice-based review of evidence and the construction of a practice-owned protocol. This was followed by a review of individual patients' records according to protocol criteria. The main outcomes were indentification of the characteristics of the patients, quantitation of GPs' intention to change treatment, explicit reporting of the reasons for not anticoagulating individuals and time to achieve the practice protocol.

Results. The data collected confirmed that patients excluded from the authoritative randomized controlled trials predominate among patients cared for in general medical practice. Practitioners overestimated the prevalence of NRAF in their patients and underestimated the extent to which their current practice offered intervention. Practitioners initially overestimated the amount of change required in patient management. In reviewing their patients' records with the intention of following evidence-based practice, practitioners explicitly described and regarded as appropriate their reasons for not prescribing anticoagulation to certain individuals. The review process was time-consuming and will need to be repeated as further evidence emerges.

Conclusion. Evidence of the complexity of applying trial results to general practice patients with NRAF is confirmed and extended.


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Appendix 1
 References
 
Non-rheumatic atrial fibrillation is a common, chronic condition, the prevalence of which rises with age1–3 and for which there is compelling research evidence that anticoagulation with warfarin, and to a lesser extent aspirin, confers substantially reduced risk of stroke and its attendant morbidity and mortality.4 It is considered that anticoagulation should be standard practice.5–7 There are concerns about the extent to which evidence about anticoagulation has penetrated into primary and secondary care3,7,8 and the degree to which clinical care for NRAF falls below the ideal. In primary care practice, there are several features which distinguish the population of patients with known NRAF from the populations selected for randomized study. The majority of patients are considered unsuitable for rigorous studies and are on average significantly older than those who are chosen to participate. For example, in the SPAF study,9 only 7.2% of identified patients (1330 out of 18 376) were considered eligible for inclusion, and during most of the enrolment this excluded patients aged over 75 "because of fear of excess risk of haemorrhage". In the Veterans Affairs Study,10 which included only men, 6.8% of men with atrial fibrillation were entered in the study, and of these only 16.7% were over the age of 75 years.

We have little knowledge of the processes by which primary health care teams (PHCTs) decide to apply evidence to individual patients under their care. In this paper we set out to explore some of the consequences for PHCTs of the decision to review clinical care for NRAF in the light of evidence. We also explored how practitioners decided on appropriate care of individual patients in an unselected set of patients with NRAF, considering that factors other than the dissemination of evidence might influence doctors' decisions.


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Appendix 1
 References
 
Six practices within the vicinity of Cambridge were identified by personal contact. Four were training practices and all were computerized. We used a purposive sample selected on the basis that they were willing to consider change on the basis of the evidence. The 6-month study was to focus on the process of change and the systematic consideration of evidence in relation to individuals. There were no clinical interventions, or comparisons of groups of patients. Analysis was confined to patients with NRAF known to the practices, since in routine clinical practice these are the patients able to benefit from the application of evidence.

The study was prospective and the sequence of activities detailed in the protocol was as follows.

    (i) One practitioner agreed to lead the process within his own practice (all lead practitioners were male).

    (ii) Each group of practitioners estimated the number of patients suffering from NRAF and the proportion of patients whose clinical care they expected would need to be changed to bring them into line with current evidence.

    (iii) Each group would devise a search which was practical in their own context to identify patients affected by NRAF.

    (iv) From the starting point of a brief digest of evidence (Appendix 1) and two main references,4,11 the practices would construct a protocol or modify an existing protocol to be applicable in their own circumstances. The time spent by doctors in doing this was recorded.

    (v) Using their own protocol, practitioners would review the records of all the patients identified as having NRAF and reconsider their current care.

    (vi) Where current care was judged appropriate but did not include warfarinization, the patient's own practitioner would explicitly state the reasons in the case of that individual.

The first step, the estimation of number of patients with NRAF, was taken by all six practices. Protocols were submitted as they were completed and were requested, if not submitted, at 6 months. Analysis of intentions with regard to change of care was requested at 6 months. Practitioners were not pressed to maintain a momentum since the process by which teams advanced towards internally agreed goals was part of the study.


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Appendix 1
 References
 
Initially, all lead practitioners were confident of completing a protocol within 6 months. None of the practices already had a written protocol. The age structure of each of the practice populations and practitioners' estimates of the prevalence of NRAF are indicated in Table 1Go.


View this table:
[in this window]
[in a new window]
  		TABLE 1 Practice details and estimates
 
One practice did not complete a protocol and provided no further data during the 6 months following initiation of the study. Five practices completed a protocol, but one did not complete the process of applying the protocol to patients' records. Further data and comparisons therefore refer to four practices which completed a protocol and used this protocol to review patients' management. The actual prevalence of NRAF summed in all the practices was 1.3% in patients aged 60–74 years (64 patients), 4.4% in those aged 75–84 years (105 patients) and 7.0% in those 85 years or older (67 patients), confirming the predominance of cases of NRAF in patients of an age generally excluded from the large formal studies.

The separate estimated and actual prevalences of NRAF in four practices are indicated in Table 2Go. In each case, practitioners overestimated the prevalence of NRAF. Practitioners believed that their practice cared for between 36 and 63% more patients with NRAF than was actually the case.


View this table:
[in this window]
[in a new window]
  		TABLE 2 Practice actual figures for NRAF
 
The extent to which patients of each practice in three age bands were already receiving anti-thrombotic therapy with either aspirin or warfarin is indicated in Table 3Go. Younger patients were more likely to be treated and more likely to be on warfarin than patients aged over 85 years, confirming previous evidence.3,12


View this table:
[in this window]
[in a new window]
  		TABLE 3 Treatment of NRAF prior to review, by age (4 practices)
 
In total, 236 patients aged 60 years or over were identified, of whom 150 were not anticoagulated with warfarin. Following review of these patients against their protocol, practitioners considered initiating warfarin themselves, or referral for specialist opinion, in 27 patients. They identified explicitly the reasons for not considering the remaining 123 patients for warfarin, and the main reasons are indicated in Table 4Go.


View this table:
[in this window]
[in a new window]
  		TABLE 4 Reasons for excluding patients from anticoagulation
 
In the great majority of cases, the reasons for not anticoagulating with warfarin were based on the conscious medical judgements of GPs or consultants to whom patients had been referred. A significant minority of patients had declined anticoagulation after explanation. Value judgements about a person's quality of life, decisions based on age alone and unclear reasons accounted for only 11.4% of reasons.

The time consumed in the practice by practitioners to generate their protocol is indicated in Table 5Go. This takes no account of the time of other members of the primary health care team or of the reviewing of individual patients.


View this table:
[in this window]
[in a new window]
  		TABLE 5 Time spent by doctors on protocol
 

   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Appendix 1
 References
 
There is widespread discussion of the incomplete application of well-founded research evidence in clinical practice.6,13,14 It may seem self-evident that good-quality research, well disseminated, should be applied in practice and result in better health for patients. Sudlow5 and Sweeney,15 considering atrial fibrillation as a specific example, suggested that there are complex reasons for the apparent failure to apply evidence. Our paper extends the existing research evidence by elucidating some of these reasons, peering into the gap between the research evidence and the reality of routine practice. In particular, we raise questions about the interpretation of risk, the unrecognized penetration of evidence and the importance to practitioners of different types of evidence, including the evidence provided by their personal knowledge of individual patients.

Risk of bleeding
In primary care practice, even minor bleeding episodes are significant to patients and of concern to GPs. The main research studies reported acceptably low levels of major bleeding complications of anticoagulation. However, the definition of ‘minor’ bleeding appeared problematic: in the Boston Area Trial (included in the pooled data in Reference 4) 38 patients were recorded as having "minor" bleeding events among which "four events led to hospitalisation and two to a blood transfusion". Evidence about the risks of bleeding while on warfarin in the community is not yet strong. The high prevalence of minor bleeding is indicated by Sweeney.15 In UK general practice, each of these events is likely to lead to at least one GP consultation and consequent reassessment of treatment, with implications for patients and their carers, nursing staff and laboratories. Fihn,16 using a large population of elderly patients controlled as out-patients in anticoagulant clinics in North America, showed an incidence of bleeding episodes "remarkable enough to report to the provider" of 812 in 3702 patient years of treatment and a relative risk of 4.5 (95% CIs 1.3–15.6) for life-threatening or fatal bleeding in patients aged over 80 years. Gustafferson17 has shown how small a change to the incidence of serious bleeding episodes would negate the economic benefits of intervention, and we do not know what might be the effect on the balance of morbidity. Reference has already been made to the high level of selection and relatively young age of patients in the widely quoted studies. Within our study, warfarin was contraindicated by previous severe bleeding episodes in five patients, and practitioners identified a factor associated with increased risk of haemorrhage in a further 33 patients. This spectrum of evidence should encourage us to consider a cautious and individualized approach to anticoagulation in the elderly. We have yet to develop the methodologies which will allow us to understand how to combine most effectively the results from trials with experience of individuals. Glasziou and Irwig18 recommend doing this "by assessing benefits and risks in each patient". However, we do not know by how much the risks to an 80-year-old are increased, either of a severe haemorrhage or of converting a minor fall into a life-threatening event, even though we may know that that individual seems more or less at risk than the average 80-year-old for whom we care.

Penetration of evidence
In our study we observed that in four practices serving 38 965 patients, none of which was attempting to adhere to any formal protocol, 92% of patients aged under 75 years had proof of prior consideration of anti-thrombotic treatment for their atrial fibrillation by receiving active therapy in the form of warfarin or aspirin. Even in patients aged over 85 years, 70% were receiving treatment. The method of identification of these patients was mainly from repeat prescription records. However, the prevalence figures for identified patients in Table 2Go are, if anything, higher than those reported by Hill2 by population screening in a UK general practice setting, so failure to identify an untreated population is not likely to account for the high intervention figures.

Types of evidence
The GPs involved in our study did not believe that they had applied the evidence as well as they might. Table 1Go shows that the five practices that provided figures estimated that between 33 and 67% of their patients assessed against a practice protocol would prove to need a change in their treatment. Four of the five practices estimated the need at between 50 and 67% of patients. Using their protocols, these practitioners found that the actual proportion that they considered to need change lay between 10 and 33%. This represented a smaller number of patients than GPs originally anticipated partly because they overestimated the number of patients with atrial fibrillation and partly because the process caused them to review their understanding of the evidence in favour of warfarinization. At the start of the study a number of practitioners believed that warfarinization was considered standard treatment for virtually all patients and felt that they should have taken more action. This is implied in their responses in Table 1Go. Once the practitioners had constructed their protocols they found that many of their patients did not fall within groups for which the evidence of best practice was clear. As a result, they felt more able to make use of the evidence of their prior experience of the individual.

Bath7 and Lip8 remind us that many specialists also decide not to anticoagulate patients with atrial fibrillation, although the emphasis of the papers suggest that this is considered a failing. The influence of local specialists must be taken into account in discussing the value placed by practitioners on evidence presented on paper. In this study, no less than 16 of the subgroup of patients who had already seen a cardiologist or geriatrician had not been recommended for warfarinization. The effect of opinion-formers is recognized in the process of managing change, and their power (right or wrong) to affect local practice is important but rarely discussed in articles about the application of evidence.

This study cannot answer questions about the effectiveness of methods of screening for NRAF,19 nor can it estimate the cost benefit of putting into practice and maintaining an evidence-based protocol. This would require a much larger study in which the intention to alter treatment was studied prospectively, including the analysis of patients in whom the intention was carried out and those in whom, for whatever reason, it was not. Our present study shows that it is superficial to regard the lack of visible application of evidence in general practice as necessarily due to either poor dissemination or lack of knowledge and interest. The GPs in this study had applied evidence to their patients to a greater extent than they themselves realized or would have reported. Only research asking them to consider every known patient in their population revealed this state of affairs. Currently they anticoagulate with warfarin 37% of all their patients with known atrial fibrillation, and this represents 76% of all of those in whom they would consider treatment after reflecting on their protocol. Certainly the 27 patients who were indentified for review deserve to be considered for treatment, but the study has revealed the extent of ‘hidden’ application of evidence in these practices. The understanding of the penetration of evidence into general medical practice requires sophisticated analysis and involves a complex combination of assessments: these include opportunity costs, risk assessment, recognizing the value attached to different types of evidence and doctors' experience of the circumstances of individual patients.


   Appendix 1
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Appendix 1
References
 
Management of non-rheumatic atrial fibrillation (NRAF): Brief evidence from the research
This document accompanies references to some of the many articles describing the original work on stroke prevention in selected patients with NRAF. Each practice has to decide on what action to take on behalf of its own patients. Set out below are some of the main points which appear to be established by the evidence:

  • Warfarin reduces the relative risk of stroke in patients with NRAF. There is no evidence of an upper age limit for this effect.
  • Aspirin reduces the relative risk of stroke in patients with NRAF, but less than does warfarin.
  • Serious complications of warfarin treatment were low in the study populations, but they rise with increasing age.
  • The evidence suggests that anticoagulating 1000 patients with NRAF for 1 year prevents between 15 and 50 episodes of ischaemic stroke or systemic embolism at a cost of 4–6 major bleeding episodes.
  • The prevalence of NRAF rises with age, but there are no reliable local figures for prevalence.
  • Anticoagulation is not indicated for NRAF patients under the age of 60 years with no other risk factors for stroke.
  • Evidence shows that additional risk factors for stroke in NRAF include:
    1. Previous TIA/CVA (17% per year)
    2. Previous arterial embolism
    3. Hypertension
    4. Congestive cardiac failure
    5. Previous thromboembolism
    6. Myocardial infarction or angina
    7. Diabetes

  • The greater the number of risk factors, the greater the risk of stroke.
  • ECHOcardiographic evidence of left ventricular dysfunction or increased left atrial size identifies a group of patients at increased risk of stroke.
  • A suitable INR for anticoagulated patients is 2.0–3.0.
  • Risks and benefits of warfarin or aspirin treatment are affected by the circumstances of individual patients.
  • If aspirin is given, 300 mg daily is more effective than 75 mg daily.
  • This digest of evidence is itself derived from a document produced by Dr Nigel Starey and a working group from the North West Anglia Health Commission.


   Acknowledgments
 
We are very grateful to the six lead practitioners, Dr John Benson, Dr Richard Davies, Dr Tony Flinn, Dr Michael Grande, Dr John Perry and Dr Mike Redwood for their hard work. We are also grateful to their partners and primary health care team colleagues. Mrs Julie Oakley typed the manuscript with great patience. The Scientific Board, Royal College of General Practitioners provided funding for this research.


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Appendix 1
 References
 
1 Wolf PA, Abbott RD, Kannal WB. Atrial fibrillation: a major contributor to stroke in the elderly. Arch Int Med 1987; 147: 1561–1564.[Abstract]

2 Hill JD, Mottram EM, Killeen PD. Study of the prevalence of atrial fibrillation in general practice patients over 65 years of age. J R Coll Gen Pract 1987; 37: 172–173.[ISI][Medline]

3 Lip GH, Golding DJ, Masood Nazir, Beevers DG, Child DL, Fletcher RI. A survey of atrial fibrillation in general practice. Br J Gen Pract 1997; 47: 285–289.[ISI][Medline]

4 Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of anti-thrombotic therapy in atrial fibrillation. Analysis of pooled data from five randomised controlled trials. Arch Int Med 1994; 154: 1449–1457.[Abstract]

5 Sudlow M, Rodgers H, Kenny RA, Thomson R. Population based study of the use of anticoagulants among patients with atrial fibrillation in the community. Br Med J 1997; 314: 1529–1530.[Free Full Text]

6 Steffensen FH, Olesen F, Sørensen HT. Implementation of guidelines on stroke prevention. Fam Pract 1995; 12: 269–273.[Abstract/Free Full Text]

7 Bath PMW, Prasad A, Brown MM, MacGregor GA. Survey of use of anticoagulation in patients with atrial fibrillation. Br Med J 1993; 307: 1045.

8 Lip GYH, Zarifis J, Watson RDS, Beevers DG. Physician variation in the management of patients with atrial fibrillation. Heart 1996; 75: 200–205.[Abstract/Free Full Text]

9 Stroke prevention in atrial fibrillation investigators. Stroke prevention in atrial fibrillation study: final results. Circulation 1991; 84: 527–539.[Abstract/Free Full Text]

10 Ezekowitz MD, Bridgers SL, James KE et al. Warfarin in the prevention of stroke associated with non rheumatic atrial fibrillation. N Engl J Med 1992; 327: 1406–1412.[Abstract]

11 More RS, Chauhan A. Anti-thrombotic therapy for non rheumatic atrial fibrillation. A review. Q J Med 1996; 89: 409–414.[Abstract]

12 McCrory DC, Matchar DB, Samsa G, Sanders LL, Pritchett EL. Physician attitudes about anti-coagulation for non valvular atrial fibrillation in the elderly. Arch Int Med 1995; 155: 277–281.[Abstract]

13 Haines A, Jones R. Implementing the findings of research. Br Med J 1994; 308: 1488–1492.[Free Full Text]

14 Walshe K, Ham C. Acting on the Evidence. University of Birmingham: NHS Confederation, 1997.

15 Sweeney KG, Gray DP, Steele R, Evans P. Use of warfarin in non-rheumatic atrial fibrillation: a commentary from general practice. Br J Gen Pract 1995; 45: 153–158.[ISI][Medline]

16 Fihn SD, Callahan CM, Martin DC, McDonell MB, Henikoff JG, White RH. The risk for and severity of bleeding complications in elderly patients treated with warfarin. Ann Int Med 1996; 124: 970–979.[Abstract/Free Full Text]

17 Gustafferson C, Asplund K, Britton M, Norrving B, Olsson B, Marké L-A. Cost effectiveness of primary stroke prevention in atrial fibrillation: Swedish national perspective. Br Med J 1992; 305: 1457–1460.

18 Glasziou PP, Irwig LM. An evidence based approach to individualising treatment. Br Med J 1995; 311: 1356–1359.[Free Full Text]

19 Sudlow M, Rodgers H, Kenny RA, Thomson R. Identification of patients with atrial fibrillation in general practice: a study of screening methods. Br Med J 1998; 317: 327–328.[Free Full Text]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (9)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Oswald, N.
Right arrow Articles by Bateman, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Oswald, N.
Right arrow Articles by Bateman, H.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?