Should Australia develop primary care research networks?

There is general agreement that primary care needs a strong research culture and evidence base if it is to deliver cost-effective healthcare.1-6 Between 2001 and 2004, the Commonwealth Government will spend $50 million on a program of Primary Healthcare Research Evaluation and Development (PHCRED). One component of PHCRED funds university departments of general practice and rural health to " 'build capacity' in the area of primary health care research"7 through research skill development, the development of research networks and actual research.

Research networks developed as a tool to increase the research capacity of primary care practitioners and to promote evidence-based practice.1,6 But have these objectives been met? In this article, I review the major achievements and challenges of research networks overseas and discuss the role of research networks in Australia.

Australian academic general practice is only 25 years old.8 There are 11 departments of general practice, yet the ratio of GPs in academic positions compared with specialist physicians8 and the research output of general practice9 remain low. The General Practice Evaluation Program (1990) provided the first significant injection of funds into primary care research in Australia. In 1992, this was included in the General Practice Strategy, which had the aims of introducing blended payments and establishing Divisions of General Practice.10 Divisions were designed to forge better links between GPs and other healthcare agencies and facilitate alternative practice and payment models.10 Divisions have become stable organisational structures that function well as associations for GPs, yet many struggle to develop a clear role within the wider primary healthcare system.11

Despite major differences in healthcare systems, the United Kingdom influences the reform of general practice in Australia: from the initial modelling of the Royal Australian College of General Practitioners (RACGP) on the Royal College of General Practitioners (RCGP)10 to the impact of the Mant Report6 on the current PHCRED program. Primary care research networks (PCRNs) are a key component of the research capacity building program in the UK; however, developing research networks is not an end in itself. The term "research network" describes various structures with varying goals and outputs.5,12 The UK and the Netherlands have a strong culture of research networks and their experience can guide us.

PCRNs emerged in 1991, at a time when there were no geographical groupings or associations of GPs, when the first Regional Research Fellowship of the RCGP facilitated development of a research group that became NoReN, the first UK PCRN.13 Concurrently, local research networks were funded in Scotland by the Chief Scientist Office to increase the quality, quantity and usefulness of health services research.14 Further networks emerged in an independent and opportunistic fashion. Many lacked specific, measurable objectives. There was no uniform reporting structure or concurrent evaluation.12 In 1996–97, 23 PCRNs were active in the UK, most located within academic departments of general practice, and receiving funding of £0 to £400 000 a year.12 There are now 40 members of the UK Federation of Primary Care Research Networks, all with formal links to an academic department (see Box 1 for an example).15

UK PCRNs aim to promote high-quality research by practitioners, high-quality collaborative projects, and research awareness among practitioners.12 They have brought together healthcare professionals interested in primary care research and have assisted them to gain training and undertake research. Lack of baseline data makes it difficult to know whether these activities represent large or small gains, or to assess how much PCRNs have contributed to increasing the quality, quantity and usefulness of primary care research (the reason for their inception). It is timely that a national review is planned, along with a national framework for accredited research practices.16

The Netherlands has a long history of research networks, with groups such as the Nijmegen Family Practice Academic Network (Box 2), the Netherlands Network of Academic Family Practices, and the Registration Network of Family Practices. These networks are more university-centred, top-down organisations than UK PCRNs, undertaking research headed by academic practitioners. It is argued that this is evidence that the Dutch experience represents the "full circle of changed research culture", as eight of the 25 physicians in the Nijmegen network have achieved an MD or PhD qualification. The Nijmegen network aims to compile long-term individual morbidity and outcome-of-care data to increase the evidence base of primary care, and to inform clinical research. Four practices collect comprehensive data and 10 practices collect data for specific conditions.1

The RACGP has a network of practices that undertake research projects in conjunction with State-based research units. These practices have been involved in large clinical trials supported by RACGP research staff. Currently, GP members of existing networks have primarily collected data for research undertaken through the primary care setting. This role is changing with the emergence of networks such as the University Family Practice Network (UFPN). In these networks, GPs are more involved in the development, conduct and interpretation of research, in addition to their data collection role (Box 3).17

Published studies of research networks generally conclude that they are worthy of dedicated funding,1,3,5 yet it has been difficult to obtain clear and tangible evidence of outputs that would not have been achieved without a formal network. Australia has the opportunity to learn from the experience of research networks in other countries, which have demonstrated several important points.

• Research networks should have explicit aims: International experience highlights the need to be explicit about what we expect a research network to achieve. A network set up to encourage evidence-based practice will have different structures, goals, activities and outputs from a network set up to foster individual GPs undertaking small-scale, practice-based research, or a network set up to undertake large-scale interventional or longitudinal studies.

Australian research networks should state clearly how they contribute to building an evidence base in primary care and improving the health of the population, in addition to how they meet the needs of their members.

• Research networks bring interested people together: The UK networks, like Divisions of General Practice in Australia, brought together GPs who had been working in relative isolation. Today groups of UK practices are involved, to differing degrees, in research in primary care (eg, collaborative projects addressing a primary care priority) in addition to undertaking their own research, and facilitating secondary/tertiary care research through primary care (eg, large clinical trials that recruit patients from all clinical settings).

Australian research networks should avoid duplication of effort by building on the capacity of Divisions, which have a functioning organisational structure and communication strategy, know the interests of their members, and have existing networks with other healthcare professionals and organisations. Interested Divisions could appoint research and development officers, who, in collaboration with an academic department, could recruit and support specially funded "research practices" in a similar way to the UK PCRN.

• Networks should have academic links: Network members need easy access to suitably qualified researchers. Formal links to an academic department of primary care will make this possible, yet such links should be mutually beneficial. Academic staff must remain active in research and publish in peer-reviewed journals. Supporting network members is time-consuming, and has placed pressure on many academic units in the UK.

Australian networks and academic departments should be mindful of potential competing demands. Discussions of expectations of level of practical support, investigator status, authorship and fund allocation are essential. Clear statements of what each partner brings to and receives from the collaboration should be made.

• Networks should complement other research capacity building initiatives: Experience from other countries demonstrates the benefit of collaboration between research networks, academic institutions and research bodies. Australian research networks need to complement the work of academic departments and the current national capacity building program, which has an agreed set of national priorities, a well established Primary Health Care Research and Information Service <www.phcris.org.au> and research funding available via the National Health and Medical Research Council competitive grants process.

Research networks will also need to complement the activities of the forthcoming Primary Health Care Research Institute, the Divisions and the RACGP. Getting effective collaboration between these players will be challenging, but needs to be a priority if we are to build the research receptiveness and capacity of primary care.

• Research networks should be realistic about what they can achieve: Australia has much less funding available than the UK network program. We must be strategic in our approach to network development, for instance by encouraging networks that aim to increase evidence-based practice. Setting up a network that tries to achieve progress in many areas is unlikely to succeed.

• It is worthwhile considering the optimal size of a research network: The Dutch experience shows that small groups of practices with skilled GP researchers can achieve considerable outputs, such as publications and higher degrees. The smaller Netherlands-style network, embedded within a Division of General Practice and supported by an academic department of primary care, is an attractive option for the Australian setting.

Australia also needs a nationally coordinated research practice network that is equipped to undertake large-scale, practice-based research to answer questions of national importance in primary care. This could be stand-alone, linked to an organisation (such as the RACGP or NHMRC), or could draw upon the divisional networks suggested above.

• Evaluation and monitoring are essential: The lack of an evaluation strategy during the development of research networks in the UK has resulted in considerable debate about what the PCRNs have actually achieved.16,18 Any network development in Australia should include, and be funded for, a concurrent evaluation that clearly defines the network's objectives, strategies and activities.

Key performance indicators that measure critical processes and outcomes should be agreed upon from the beginning. Indicators should include a mix of measures, such as the number of practitioners and practices actively contributing to research, the number of practices with robust data collection systems in place, the number of practitioners skilled in evidence-based medicine, the amount of grant income from peer-reviewed competitive funding rounds, the number of practitioners completing research training, and the number of reports published in peer-reviewed journals. The networks will need to implement change in response to the evaluation findings.