Abstract
With a recognition that the management of headache remains far from ideal, there may be more immediate potential to reduce the burden of illness from developments in the delivery of headache services. There is a paucity of evidence in this area and a danger that an expanding research agenda will be dominated by inappropriate methodological frameworks that have been so successful in developing medical treatments. The prevailing scientific methods are underpinned by statistical approaches that are aggregative in nature and assume independence of system elements, an approach that may have limited utility in the analysis of complex systems such as headache care delivery. This review calls for a shift in headache research resources to organizational development and briefly outlines alternative methodological considerations.
Introduction
The aim of health service research is to produce knowledge that may be applied by policy makers and practitioners to improve the public's health. A number of problems of obtaining evidence to inform service development are recognized and include:
The danger of asserting the hegemony of a particular type of evidence.
Exaggerated claims over what research can deliver.
Difficulties in generalizability across services that differ in organization, process and culture.
The problem of different time tables of research and policy making.
The impact of political agendas and the research process.
From the perspective of the delivery of headache care, three key questions are relevant: what configurations of service delivery work; who do they work for; and are they worth paying for?
The care that headache sufferers receive is generally poor and <50% of patients are satisfied with their current treatment (1–3). To date, the primary research focus has been on the epidemiology, pathophysiology and treatment of headache. Healthcare resources are limited and as current therapeutic developments become increasingly marginal, the development of more effective models of service delivery may offer greater immediate potential to reduce the burden of headache.
Figure 1 demonstrates this argument using the concepts of opportunity cost and marginal analysis.
Healthcare resources are limited and resources invested into one area are at the cost of lost opportunity of benefits in another. For any intervention, benefits do not continue to increase in a simple proportional relationship. As more resources are invested, a point will be reached (A) where further increments of investment produce smaller increments of benefit (the law of diminishing marginal return). The contention is that drug development has reached this point and resources would generate greater marginal benefit if diverted to the development of headache services.
In 1994 an International Headache Society Taskforce was set up to describe the ideal organization for the delivery of headache services (4). A number of general suggestions were made, including the importance of education, research and the development of national treatment guidelines. More recently, the World Health Organization has proposed regionally based demonstration projects in developing countries that make optimal use of available resources (5).
Although there have been a small number of studies comparing the outcomes of different models of care (6, 7), the debate on effective service delivery has been largely rhetorical and without evidence of effectiveness or cost effectiveness. For example, the British Association for the Study of Headache has suggested that intermediate care headache centres staffed by general practitioners with a specialist interest should support front-line general practitioners and that neurologists in specialist secondary-care centres should support these two levels (8), but there is no evidence to support these views.
The objective of this review is to alert the headache research community to the importance of developing a research agenda in headache service delivery where there may be more immediate gains to reduce the burden of illness. Its aim is to offer a broad overview of methodological possibilities and, in particular, of the dangers of imposing research approaches that have worked well in the clinical setting to the more complex environment of healthcare delivery.
Developing an evidence base for headache care delivery
How can evidence be produced on which to base the delivery of effective and cost-effective headache care? A number of approaches are available across a broad range of disciplines that vary in the type of methods used, the role of the researcher and approaches to generalizability, reliability and validity. Two main paradigms are recognized.
The positivists—the truth is out there
The positivists see a generalizable truth that will inevitably yield to investigation. The confident assumption is that more data and increasing technical sophistication will inevitably do the job. Systems are viewed as independent elements that can be understood by a reduction into their component parts and the application of inferential statistics. Interactions between elements are seen as confounding variables that must be minimized.
It was inevitable that the success of medical science would direct positivistic methodologies based on the randomized controlled trial to more complex systems such as healthcare delivery. Thus, in the UK an NHS Research and Development Directorate was established to correct the discrepancy between the ‘technical sophistication of medical interventions and organizational dysfunction’ (9). The dominant approach to the evaluation of complex interventions such as health service delivery was to remain within a phased approach which reflected exactly the phases of drug development and was ultimately based on the randomized controlled trial (10).
However, these increasingly sophisticated approaches have not lived up to their early promise. Research still has little direct influence on healthcare delivery policies (11, 12) and evidence-based technical solutions to organizational problems based on the randomized controlled trial have had little impact at grassroots level (13). For example, doctors seem reluctant to follow research-based guidelines (14), but implement small changes that they consider improve on existing structures (15). These are informed by reiterative negotiation with a variety of sources and informal interactions (16). Healthcare managers faced with competing objectives and uncertain cause and effective relationships find relating ends to means highly problematic and encounter substantial barriers that cannot be overcome by methodological refinements or the collection of greater volumes of data (17).
Interpretative approaches
An alternative interpretive framework suggests that our reality is socially constructed and in part will be constructed by the perspective of the investigator. The focus is on understanding the interactions between organizational members in a system that cannot be understood by a reduction into its component parts or engineered towards defined policy objectives. There is no correct organizational solution towards which research will inevitably converge.
A number of pragmatic, interpretative research methodologies have been developed that reflect the complex interdependencies of the healthcare environment. These are invariably underpinned by qualitative methodologies that deal with information not easily reduced to numbers (18).
For example, action research is an enquiry that describes and interprets situations while implementing changes aimed at improvement and involvement. The process involves identification of a problem, engagement with collaborators and the introduction and monitoring of change to address the problem. Important elements are flexible planning, an iterative cycle, a recognition of the importance of context and simultaneous improvement by producing change (19).
Realistic evaluation seeks to understand how mechanisms interact with contextual factors to create unique patterns of outcomes (20). For example, how do specific general practitioner headache clinics interact with local professional networks, relationships, history and cultural norms to produce specific outcomes? This runs counter to the prevailing view, where trials seek to evaluate whether an outcome was achieved or not, minimizing contextual factors in order to provide generalizable relationships between mechanisms and outcomes. A realistic evaluation can offer far deeper and useful insights into how specific interventions may accomplish their objectives, shifting the focus from ‘what works’ to ‘what works for whom in what circumstances?’ These hypotheses then frame the research strategies to test possible configurations of context, mechanism and outcome to provide results that may be transferable rather than generalizable. Although evidence from randomized controlled trials is not excluded, the importance of qualitative, ethnographic and case study research is elevated and can provide a richer understanding of local contexts and contingencies.
More recently and converging from a number of disciplines, complexity theory is challenging dominant theoretical frameworks that retain the fundamental idea that order needs to be somehow created by external forces (21). It offers an over-arching interpretive model with an emphasis on organizations as coevolving systems, where the focus is on patterns of relationships, how they are sustained and self-organize. These insights lead to a number of important implications for a developing organizational research agenda (22). More radical complexity perspectives focus on the quality of interpersonal relationships in an organization and question whether knowledge can be codified and managed (23).
An important interpretive framework for organizations is the concept of ‘Communities of Practice’ (24). This refers to a description of relating that occurs through particular activities or practices undertaken by a group of people that facilitates their sharing of knowledge and negotiating of meaning amongst them. The emphasis is on the importance of implicit knowledge as the primary source of an organization's innovative potential. Communities of practice are seen as appropriate social structures suitable for developing and sharing knowledge in the organization and consist of three structural elements: domain, e.g. know-how or highly specialized professional expertise; community, the environment in which people interact, learn and build relationship; and practice, the set of ideas, tools and documents that they share. An important feature is the development of design elements that can be of assistance in supporting participation against a background of understanding what emerges is a product of both the design and the participatory process which is owned and enacted by the group itself.
Involving lay people in the development of headache services
The interpretist framework outlined above has led to recognition of the importance of involving lay people (patients, potential patients, carers and members of the public) in the research process. By contributing their unique perspectives, they can make research more relevant to the needs of the health service with skills that compliment researchers as equal partners (25, 26).
In the UK, the Department of Health has emphasized the importance of public and patient involvement in the development of health services and recognizes that methodologies appropriate to the evaluation of patient participation will not always follow the scientific and posivitistic research paradigm (27). In their most challenging form, these developments portray research as a dialogue rather than an expert activity undertaken in a socially constructed framework amidst a web of commitments that bind workers together (28).
Here, the terms researched upon and researcher lose their distinctive meanings. Relationships are egalitarian and there are shared objectives by all participants. The more powerful will strive to take a perspective from below. All participants are empowered to construct concepts and categories, discuss results and determine the cause and outcomes of their research. For example, one project has been published where migraineurs worked with researchers to design the study, undertake the research process and write up the findings (29).
Despite the acknowledged practical difficulties, which include competing agendas, power differentials, loss of researcher independence, the investments of time and commitment with pay-offs unlikely to be realized in the short term and technical language barriers, any developing research programme must recognize the importance of lay people and their contribution to the research process. The challenge is to set aside the maze of disciplinary intellectual commitments, methods and goals that derive identity and integrity from institutional structures and accommodate a broader view of the goals of partnership working.
Conclusion
Demographic, epidemiological and socio-economic developments are driving demands for improvements in the efficiency and quality of healthcare delivery. These tensions have resulted in a search for ways of re-aligning healthcare organizations, reflecting a broader agenda to modernize and improve public services as a whole. From the perspective of headache, the focus remains on clinical research, despite the fact that the delivery of services remains far from ideal. In the medium term, greater benefit may be obtained from shifting research resources to organizational development, recognizing that the development and evaluation of new therapies and diagnostic tools is only part of the answer to better healthcare.
Heath service research is becoming a contested terrain where multiple approaches and perspectives are legitimate and the frontiers between researchers, practitioners, patients and citizens are becoming increasingly blurred. This does not overturn approaches based on inferential statistics, but highlights the importance of matching the research approach to the context and complexity of the environment to which it is applied.
However, there is a real danger that reductionist research frameworks underpinned by the randomized controlled trial and the application of inferential statistics that have been so successful in developing medical treatment for headache will dominate the service delivery research agenda. Prevailing statistical approaches that are aggregative and assume independence of system elements may have limited utility in the analysis of complex systems such as headache care delivery. In view of the complexity of the healthcare environment and widespread differences in healthcare delivery contexts, a generalizable evidence base is not a relevant pursuit. The challenge for the research community is not to produce idealized models that have little relevance to local circumstances, but to deliver insights in partnership with patients that commissioners and providers can adapt to their needs and contexts.
In conclusion, we need to guard against the automatic application to headache care delivery of methodological techniques that have been so successful in the clinical and scientific setting. A limited evidence base underpinned by methodologies that reflect underlying organizational mechanisms is more likely to get us to an approximation of where we want to be rather than an untenable pursuit of rigorous methodological frameworks based upon incorrect assumptions where the thinking is wrong.