general practice:

The use of benchmarking as a management tool is increasing throughout the NHS, and this trend is likely to continue with the new performance framework, which places great emphasis on comparative analysis. Despite this, there appears to be no central benchmarking of primary care.

This was one of the major drivers behind the NHS Benchmarking Club's primary care indicators project. The club - which comprises 40 health authorities, 100 primary care groups and eight associate members (the Audit Commission, the Department of Health, the Welsh Office, four regional offices and a family health services agency) - has completed the first stage of a project designed to benchmark the state of primary care. The project has proved valuable in identifying good practice approaches and highlighting development needs across 23 HAs, covering 2,270 practices (see box, right).

The work is now poised to proceed with the active involvement of a large number of PCGs.

The Benchmarking Club

The NHS Benchmarking Club was set up in 1996 to encourage the use of benchmarking and good practice. Members set the club's work programme through supporting individual projects. Current projects include public health, health improvement programmes, finance and information, HA effectiveness, PCGs, and primary care indicators.

The process involves:

agreeing benchmarking definitions - clarifying the scope of a project, developing a data specification and supporting definitions;

data collection and analysis leading to the production of structured benchmarking comparisons;

discussion of good practice approaches and innovations - identification of good practice approaches, dissemination of findings to encourage local implementation work.

The primary care indicators project

The project decided to focus on general medical services because of the gatekeeper role of GPs and the significance of spend in this sector.

The objectives for the project were:

to develop a toolkit for assessing the state of primary care;

to encourage sharing of good practice approaches between members;

to provide a platform from which members could take forward local implementation work.

The data specification covered around 100 core indicators spread across nine broad categories - designed for ease of collection. The specification required data to be collected from general practice computing systems - HA Exeter family health services system, the Prescription Pricing Authority's computer system PACT, and, to a lesser extent, local systems. Some of the main indicators are outlined in the box below.

Results

Comparative analysis across the HAs revealed large variations in primary care coverage and performance across the NHS. There were large differences in the make-up and coverage of GP practices, prescribing performance, services offered to patients, practice infrastructure, information management and technology, and patient access and involvement.

The reasons for these differences were discussed with participating HAs and GP colleagues. The main factors impacting on primary care development appeared to be historic investment levels in primary care (which varied between HAs) and the extent of innovations and good practice being introduced by both HAs and practices.

HA and practice profiles

Large variations were observed in the average size of practices, with a significant contributory factor being the number of single-handed practices. Rather than an urban-rural split, more evidence was seen of a differential between London and the rest of the country. An average of 29 per cent of practices across the participating HAs were single-handed. This rate increased to around half of all practices in London HAs such as East London and the City or Redbridge and Waltham Forest.

Almost 60 per cent of practices reported access to female GP sessions. But an almost three-fold differential was evident between areas with the greatest coverage (Buckinghamshire and Newcastle, both with around 80 per cent) and the HAs with the lowest coverage (Sunderland at 32 per cent). HAs with low coverage rates were making efforts to improve access to female GPs through initiatives such as the creation of job-share GP posts, and attempts to create structured access to female locums.

IM&T and business management

There was a strong relationship between investment levels and IM&T coverage. For example, Dorset and Stockport HAs (among the biggest spenders) achieved the highest coverage for GP pathology and NHSnet links.

The low number of practices linked to NHSnet - less than a quarter in the best HA and none in many - shows that the IT infrastructure is very poor and that it will be a huge challenge to meet the government's target of all practices being linked to the NHSnet.

Prescribing

This is one of the main areas of interest with the move to PCG-controlled prescribing budgets and peer review of prescribing performance. Given the difficulty of the debate around clinically effective prescribing, the use of indicators on prescribing was limited to analysis of performance on commonly accepted measures - for example, performance against budget, generic prescribing rates, coverage of practice formularies and prescribing rates for certain categories of drugs.

Findings revealed large differences in prescribing habits between HAs. For example, there was a four-fold variation across HAs in the proportion of practices achieving the 60 per cent generic prescribing target.

A strong link was revealed between investment in prescribing adviser support and subsequent performance.

HAs investing the most in prescribing adviser support - Kingston and Richmond, Newcastle, Stockport, and St Helens and Knowsley - all achieved in excess of 80 per cent of practices reaching the 60 per cent target.

This provided a core lesson - now picked up by many HAs and PCGs attempting to improve generic prescribing rates. Wide recognition of the value of prescribing advisers appears to have created a shortage of suitably qualified staff available to work with HAs, PCGs, and individual practices. HAs at the bottom end of the performance scale have experienced particular difficulties in recruiting the necessary staff to improve performance. The percentage of practices with published formularies ranged from none to 78 per cent in South Essex.

Service coverage and performance

Analysis was restricted to those areas where commonly accepted performance measures exist - for example, uptake for screening programmes. This was supplemented by comparative analysis of service coverage from many of the commonly used GP items of service measures, such as comparative rates for night visits, minor surgery procedures and so on. An eight-fold range was evident across the 23 HAs in the per capita rate of minor surgery procedures provided by practices.

Local minor surgery rates were highest in non-metropolitan areas (for example, Buckinghamshire and Shropshire with eight procedures per 1,000 population) - perhaps explained by more distant access to secondary care, as well as the rate of GP accreditation to perform minor surgery and access to suitable facilities in practice premises.

Screening

Analysis of screening programme coverage rates revealed a range in performance between HAs. Coverage rate differences have been well documented in the past and are increasingly the subject of NHS Executive performance management.

One related indicator was used to identify the quality of smear tests across HAs. Comparisons of the number of inadequate smears as a proportion of total smears revealed a noticeable variation between the HAs with the lowest rates (Morecambe Bay and Shropshire at 5 per cent) and the HA with the highest rate (St Helens and Knowsley at 13 per cent). Discussions between project participants suggested that high rates of inadequate smears highlighted a need for practice staff to have training in how best to undertake this examination.

Practice infrastructure

Analysis of indicators on practice staffing and premises again revealed large differences among the 23 HA participants. The availability of cost- rent and improvement grant funding appeared to differ between HAs, reflecting local HA priorities.

Benchmarking comparative numbers and skills of staff working in a practice setting again revealed interesting differences. For example, the use of nurse practitioners in general practice appears limited despite much pushing of this initiative over the past three years. In most HAs, fewer than 5 per cent of practices have access to a nurse practitioner. Even in Dyfed Powys, the best-performing HA, the figure was only 10 per cent.

More basic indicators of the state of practice administration and management again reveal unexpected differences. In about one third of the HAs, more than 90 per cent of practices had a specified practice manager. But this drops to around 50 per cent of practices in HAs where this has not been seen as a priority.

Patient access and involvement

The results of surveying patient access and involvement may serve as useful baseline measures for the improvement in public involvement levels expected from PCGs. The extent of the use of patient participation forums at practice level appears limited, with most HAs having only one or two practices with active participation groups. At best (in Shropshire and Sunderland), the use of participation groups at practice level covers 15-18 per cent of practices.

Access to surgery sessions also shows large differences between HAs. On average, 54 per cent of practices offer 'early surgery sessions' (starting before 9am), 41 per cent offer 'late' sessions (after 6pm) and a surprising 78 per cent offer non-urgent weekend appointments.

The value of benchmarking as a comparative analysis tool comes through again here with neighbouring Dorset and Wiltshire HAs. They share similar demographics but have more than a tenfold difference in service availability for early, late and weekend appointments.

Conclusions

Our research reveals the great diversity of primary care across different areas of the NHS. This creates a need to look at the totality of services and performance - weaknesses or gaps in some areas are often offset by strengths and good practice in others. A standard core general medical service exists, but key areas such as added-value services, access to services and coverage, suggest that the nature of primary care varies in different parts of England and Wales.

Project participants reported that the ability to identify where local services are outliers is perhaps the main short-term benefit of taking part in benchmarking projects. Many participants have identified primary care development needs that are now being targeted through local primary care investment plans.

Further benefits are in identifying good practice. Most of the 23 HA participants were able to demonstrate examples of local innovations that have delivered measurable improvements in service performance and access. These were shared between participants at a series of good practice workshops.

Next steps

Positive feedback from participants has led to the establishment of a second stage of the project, which will take place during the rest of the year. This will widen the scope of the work to cover increased numbers of clinical indicators and also to take account of PCGs.

Targeted clinical indicators will be used to benchmark primary care involvement in the treatment of diseases - initially, asthma, diabetes and coronary heart disease. Benchmarking techniques will be used to establish links between disease incidence, prescribing patterns and hospital admission rates.

The PCG perspective is being taken on board through the development, by a new project team of PCG chief officers and HA primary care managers, of a joint benchmarking data specification to meet the performance management and developmental needs of both HAs and PCGs.

Elaine Rodger is chief executive of Sunderland North primary care group and a steering group member of the NHS Benchmarking Club.

Stephen Watkins isdirector of Lynx VOI strategic and operational consulting.

Key points

A study of general practice involving 23 health authorities and 2,270 practices completed last year, revealed considerable variations across England and Wales.

Single-handed practices account for an average of 29 per cent of all those in the study.

The majority of HAs had no practices linked to NHSnet.

More than half the practices offered early morning surgeries, and many also offered appointments aftersix o'clock.