Preventing recurrent coronary heart diseaseBMJ 1998; 316 doi: https://doi.org/10.1136/bmj.316.7142.1400 (Published 09 May 1998) Cite this as: BMJ 1998;316:1400
We need to attend more to implementing evidence based practice
- Trudy van der Weijden, Senior researcher (, )
- Richard Grol, Professor
- Centre for Quality of Care Research, Department of General Practice, Universities of Maastricht amd Nijmegen, the Netherlands
General practice pp 1430, 1434
Following the publication of several recent large studies (4S, CARE, and WOSCOPS), there is little doubt about the importance of prevention in patients with coronary heart disease, though controversy still exists about its value in patients without symptoms. General practitioners are in a favourable position to take on the task of secondary prevention, since most have a continuing relationship with their patients, and these patient contacts offer opportunities for measuring cardiovascular risk factors. Nevertheless, preventive care in general practice is haphazard, 1 2 and in this issue Campbell et al confirm this shortfall (p 1430).3 The question that therefore arises is how to implement the new evidence on preventing coronary heart disease effectively in general practice.
An audit in 95 practices in the Netherlands showed that many general practitioners had a critical attitude towards integrating prevention into practice4 and that few practices were sufficiently well organised to provide effective preventive services. Thus, efforts to implement prevention should be directed both at individual general practitioners and at the organisation of services. A controlled trial in these 95 practices studied the effects on the organisation of cardiovascular preventive care of visits to practices by facilitators, who trained practice nurses to set up preventive clinics.5 Compared with practices which just received feedback on their preventive care, the intervention practices improved care significantly, both the way it was organised and the recording of cardiovascular risk factors.
In a well designed and encouraging study in 19 general practices in Scotland, again by Campbell et al (p 1434),6 implementation of preventive care for patients with coronary heart disease was also achieved by such an organisational measure. Almost 2000 patients were identified, and 71% agreed to be randomised. Half were invited to attend nurse led prevention clinics (attendance rate 82%); the other half received usual care. Within a year the intervention group showed important benefits. Nevertheless, some questions emerge from this study.
Firstly, are the benefits of the drug interventions—aspirin, β blockers, angiotensin converting enzyme inhibitors—additive? Since the interaction between these drugs is not established, it is possible that current candidates for cholesterol lowering drugs might not need such treatment if they had received adequate alternative intervention targeted at other risk factors.7 Secondly, the fact that nearly a third of the patients did not wish to participate in the trial, and that 18% of invited patients did not attend the clinic, is worrying. There might be a selection bias in favour of more motivated patients, and patients of lower socioeconomic status—already a vulnerable group8—might be overrepresented among non-participants. A complementary strategy through case finding seems necessary to reach all patients with coronary heart disease. The literature on implementing evidence based change tells us that multifaceted interventions, targeted at specific obstacles to change, are effective in inducing change. 9 10 Because the traditional ways of organisation within general practice seem to be an obstacle to efficient prevention, giving a central role to nurses or practice assistants may be an effective approach. However, we need to evaluate these new models.
How, therefore, might we evaluate a preventive strategy that combines nurse led clinics and case finding by the general practitioner? Firstly, it is questionable whether randomisation and analysis at patient level, as was done in Campbell and colleagues' study, is adequate in this type of study. Individual professionals' behaviour influences patient management to the extent that patients seen by the same professional cannot be assumed to be independent, and therefore the professional should be the unit of analysis and thus the unit of randomisation. Moreover, patients in the control group may try to cross over to the intervention group, or general practitioners may improve their care for all patients. To compare the effect on different groups of patients, randomisation, or at least equivalence, at the level of the practice or individual practitioner should be achieved.
Secondly, studies should describe variances between and within practices, general practitioner and patient characteristics, and any local problems at the interface between primary and secondary care because these might help clarify suboptimal medical management. Thirdly, implementation methods should be cost effective. The balance between the costs of nurse led clinics and their effects should be determined. Costs are usually influenced by local factors, and therefore the external validity of study findings needs to be discussed. In the study of Campbell et al the high incidence of coronary heart disease in Scotland and the discussion of the WOSCOPS study in the media might have had an extra motivating effect on patients, nurses, and doctors. Evidence based guidelines on preventing coronary heart disease in general practice need to be complemented by evidence based implementation.10