Stroke is the second most common cause of death worldwide¹ and the predominant cause of adult neurological disability. There is now strong evidence that stroke prevention is a realistic proposition. The incidence of stroke is falling, despite an ageing population, owing to increased use of secondary preventive treatment and reduction of risk factors.² One key element of stroke prevention is the management of patients with transient ischaemic attacks (TIAs). Since 15% of ischaemic strokes are preceded by a TIA,³ rapid access clinics have been advocated in national guidelines.^4,5 Recent community based data confirm the early risk of stroke following TIA or minor stroke to be as high as 8–12% at 7 days.⁶ Patients suffering a TIA should be seen within the first few days for stroke prevention to be most effective. Despite the strong evidence of the benefits of secondary prevention, TIAs remain under-recognised and undertreated.⁷ Only two thirds of UK hospitals have a neurovascular clinic and only just over half of these can see and investigate patients within 14 days.⁸

In 1999 a rapid access neurovascular clinic was established at Bradford Teaching Hospitals Trust to provide specialist review of all patients who had suffered a TIA. A multidisciplinary group of health professionals from primary and secondary care was convened to develop evidence-based guidelines for the diagnosis, management, and referral of patients with TIA. There were three main goals of the guidelines: (1) to improve recognition of TIA symptoms; (2) to ensure that patients who had suffered a TIA were treated immediately with an antiplatelet agent and given advice not to drive; (3) to ensure that patients with suspected TIA were referred to the rapid access clinic for specialist review of diagnosis, investigations, and management.

A quality improvement strategy to support general practitioners in adopting the guidelines and achieving these goals was developed. It was based on methods that have been demonstrated to be effective in changing professional practice. The quality improvement strategy was pragmatic and based on interventions that were feasible and affordable. In order to evaluate the benefits of the strategy, it was implemented in one half of the local district with deferred implementation in the other half to provide a control comparison.

SETTING

There are three primary care trusts in Bradford. Two (North Bradford and Bradford South and West with respective populations of 98 000 and 147 000) consist mainly of large partnerships (15 and 23 practices, respectively) and their populations are demographically and socioeconomically similar. Each has its own programme for training and development. The third (Bradford City, population 147 000, 42 practices) has a high proportion of single handed practices and a more culturally diverse and materially deprived population. Half of the practices in City had established a practice development programme that was run on a Wednesday and the other half on a Thursday. We randomly selected one of the peripheral South and West primary care trust, and one half of Bradford City practices to collaborate in developing a quality improvement programme and participate in a controlled before and after evaluation.

STRATEGIES FOR CHANGE

Local ownership and involvement of relevant stakeholders was seen as a key step in achieving change (box 1). During the development of the guideline, interviews were undertaken with patients and professionals to identify gaps in current services, barriers to good practice, and educational needs. The TIA evidence base and aspects of cost effectiveness were reviewed by a multiprofessional group who met on five occasions. Draft versions of the guidelines were circulated widely to health professionals in the target practices to promote awareness about the topic and allow feedback on the emerging recommendations and guidance. Details of the implementation programme can be found at www.learnonline.nhs.uk/pace.

An audit of referrals to the rapid access clinic was undertaken to provide practice specific data on numbers of referrals and their appropriateness. A training workshop was run by local stroke specialists in South and West primary care trust and in City for the half of the practices selected for the change strategy. Those who attended included general practitioners, nurses, practice management staff, and pharmacists. These workshops used case histories and an interactive format to ensure relevance and participation and included feedback of the audit results. When representatives from practices were unable to attend the workshops, individual practice visits were offered to present and discuss the emerging guideline. A total of 19 outreach visits (9 in City and 10 in South and West) were arranged.

To remind clinicians about the diagnosis of TIA, criteria for referral and the treatment advice, specific paper and electronic referral forms were developed that provided prompts to reinforce the guidance. Key recommendations were also summarised on coasters that were distributed to general practitioners.

Finally, the importance of building change into the medical services contracts was recognised. For South and West primary care trust the audit results were linked to quality targets in the Personal Medical Services contract agreements.

KEY MEASURES OF IMPROVEMENT

Three main outcome measures were identified: the numbers of referrals to the rapid access clinic and the number of patients with a confirmed diagnosis of TIA, both before (27 months—from the start of data recording at the clinic) and after (19 months—to the end of the funded evaluation) the intervention were measured. Following the intervention the proportion of TIA patients who were receiving an antiplatelet drug (either already prescribed, prescribed after the ischaemic event or contraindicated) when seen in the rapid access clinic was combined with the proportion of drivers who had been advised not to drive into a single outcome measure of overall compliance with best practice recommendations described in the guideline.

EFFECTS OF CHANGE

We found large overall differences in referral rates between City and the other two primary care trusts. The outcomes were therefore evaluated by comparing the two City groups, comparing South and West with North primary care trust, and combining these two comparisons in a single significance test, adjusted for age.

There was a significant difference in the change in referral rate to the rapid access clinic between the practices that were part of the quality improvement programme and those that were not (fig 1, table 1). There was a 41% increase in referrals from trained practices compared with control practices. This difference demonstrated improved access to specialist services for patients in the practices receiving training. Total referrals from the trained South and West primary care trust, which were initially lower than those in the North, rose to the same level after the intervention. Referral rates declined in the control (untrained) but not in the trained City practices. Rates for confirmed TIA diagnosis followed the same pattern, but the difference did not quite reach significance. Referral rates in the North rose in the final 6 months, possibly related to the spread of the guidelines and recommendations beyond the intervention practices.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Monthly appropriate referrals to TIA clinic by primary care trust group (moving average; intervention at 28 months).

Adherence to best practice standards was significantly higher in practices that had received the training programme than in the controls (table 2). The improvement occurred principally for driving advice for the non-City practices and for the antiplatelet advice for the City practices. This may reflect the higher car ownership in the more affluent non-City areas, and the existing high levels of aspirin prescribing that limits scope for improvement.

LESSONS LEARNED

The National Service Framework for Older People stipulates that patients suffering from a TIA should be referred urgently to a rapid response neurovascular clinic using a locally agreed protocol.⁵ This intervention improved the identification and referral of patients who had had a TIA to specialist services where they could be appropriately investigated and treated. In addition, there was an improvement in the early treatment of patients with effective antiplatelet agents and the provision of safe advice about driving. The increase in the number of inappropriate referrals following the change programme was small and non-significant. It was felt that this increase was inevitable due to diagnostic uncertainty and was acceptable to ensure that the right patients received the right care at the right time.

A number of important components for achieving changes in professional practice have been identified.⁹ These include strategic planning, contextual analysis, and use of a multifaceted approach to implementation. The particularly novel aspects of our project included the stepwise and pragmatic intervention to improve quality, and our findings that a strategic approach to change is feasible and can be implemented across a diverse health community. The intervention described recognised the complex influences on professional behaviour and adopted a whole systems approach to change, covering as many aspects as we pragmatically could. Professionals were successfully involved in the collaborative development of evidence-based guidelines and in the implementation of these guidelines using dynamic interventions that were informed by those targeted for change. This tailored and multifaceted change strategy recognised that people have different learning needs. A stepwise approach to meeting these needs allowed concentration on resource intensive one to one outreach visits for those who really wanted them. The importance of building quality improvement into professional contracts was recognised.

There is an imperfect evidence base to inform how to change professional behaviour.^10,11 This rigorously designed and conducted study used a before and after design with control comparisons to provide robust evidence about the effectiveness of an implementation programme. A recent evaluation of the impact of NICE guidance identified factors that were important in successful implementation.¹² Our multifaceted intervention benefited from some of the key components identified from this study and previous reviews (box 2).

CONCLUSIONS

Healthcare organisations have limited resources for education and quality improvement and require better evidence about how to use these wisely. However, finding generalisable evidence that is appropriate to varied local contexts can be difficult. The model of change that is described here can be developed to provide a systematic approach to the implementation of evidence-based practice across the health community. This model combines careful planning with a multifaceted approach to changing practice that can work in very different primary care settings. It is a major advance from the haphazard and poorly coordinated approach to dealing with national and local clinical guidelines that has existed historically, and it provides a structured and strategic framework for getting changes into clinical practice. Further research is required to determine if it is a generalisable and sustainable model for implementing the ever increasing volume of guidance in primary care.