Family Practice Advance Access originally published online on December 6, 2005
Family Practice 2006 23(1):1-7; doi:10.1093/fampra/cmi102
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Editorial |
Primary care epidemiology: its scope and purpose
Department of General Practice and Primary Care, Foresterhill Health Centre, Westburn Road, Aberdeen AB25 2AY, Scotland
Correspondence to Professor Philip Hannaford, Department of General Practice and Primary Care, Foresterhill Health Centre, Westburn Road, Aberdeen AB25 2AY, Scotland; Email: p.hannaford{at}abdn.ac.uk
Received 21 April 2005; Accepted 30 October 2005.
Hannaford PC, Smith BH and Elliott AM. Primary care epidemiology: its scope and purpose. Family Practice 2006; 23: 1–7.
Introduction
Clinical epidemiology is a basic science of medical practice; informing, among other things, diagnostic, prognostic and therapeutic decisions relating to individuals.1 In the 1980s, Mullan coined the term ‘primary care epidemiology’ to describe the application of clinical epidemiology to primary care practice.2 Calls for a new discipline derived from the desire of Mullan, and others, to see an increased use of epidemiological principles to shape the development of community orientated primary care.2–4 To a large extent, these ambitions echoed those of Tudor Hart, who, a decade earlier, had called for a marriage between primary care and epidemiology so that programmes of organised, whole-population anticipatory care could be developed.5 They also built upon the work of other GP researchers who used epidemiological principles to describe clinical events in their practices.6–8 The emphasis on affecting the dynamics of health care within defined (usually single practice) communities required short intervals between the seeking and using of information, with an acknowledged consequence that primary care epidemiology might only be able to answer relatively basic questions.3 We believe, however, that this focus is too narrow for the sector that provides ‘integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practising in the context of family and community’.9 Primary care epidemiology can contribute to wider improvements in health and health care services, through better understanding of disease aetiology, use of health care services and the role of different health care interventions. Primary care is delivered by a wide range of health professionals, including nurses, doctors, care assistants, social workers, mental health specialists, dieticians, pharmacists, dentists, optometrists and other health care professionals, administrators and clerical staff. The practice of each of these professional groups can contribute to, and needs to be informed by, primary care epidemiology.
Key features of primary care epidemiology
Adapting a widely accepted definition of clinical epidemiology,10 we define primary care epidemiology as: the application of epidemiological principles and methods to the study of health problems encountered in primary care, including their aetiology, prevention and diagnosis, and with a view to improving their management. The discipline includes studies of the interface between: primary care and the community/general population; primary and secondary (or tertiary) care; and different members of the primary health care team. Consultations are central components of primary care delivery, therefore much of the work involves studying the determinants and outcomes of consultations in primary care. ‘Determinants’ include the nature of symptoms, signs or illnesses occurring in the community, and factors influencing decisions to consult or, crucially, not to consult. ‘Outcomes’ include duration, severity and impact of symptoms, signs or illnesses. They also include all aspects of primary care management, such as investigations, referrals and treatments. Notable features of primary care epidemiology are the explicit need to study symptoms and symptom complexes as well as formal diagnoses, and for a community/general population perspective.
Researching symptoms
So far, most clinical research has focused on the aetiology and management of defined diseases, or their associated risk factors. Comparatively little research has examined the epidemiology of symptoms themselves,11 even though these subjective ‘falls from usual state of functioning’12 are both very common12 and powerful drivers of health care utilisation.13 Symptoms presented to primary care practitioners are undifferentiated, multifactorial in origin, diverse in spectrum and frequently of short duration. Often only a small proportion of presented symptoms can be attributed to physical or psychological disease, even after detailed investigation. For example, in a study of primary care attendees in North America, 15% had an identifiable organic cause for their presenting symptoms, 10% a psychological explanation and 75% an unknown cause.14 It is not surprising, therefore, that much of the work of primary care practitioners involves the management of symptoms rather than discrete, well-defined diagnoses, especially when treatments are often independent of any specific attributable diagnostic labels used (as, for example, with most back pain). The frequent lack of formal medical diagnosis within primary care does not diminish the sometimes serious clinical, social and economic consequences of symptoms15–17 although it does add to the scientific challenges of researching them.12
The community perspective
A population-based approach is essential if we are to fully understand the frequency of symptoms and illnesses occurring in society, and how people respond to them. Epidemiological studies that examine the occurrence of, or risk factors associated with, disease in selected groups, such as hospital-based cohorts, provide useful information about the prevention or management of disease in the same populations but may provide misleading information about its management in the general (primary care) population where the incidence and prevalence is often lower. The population denominator is crucial for understanding the clinical significance of risk factors in the community, the absolute risk associated with interventions or risk factors, and the efficient targeting of preventive and other resources. It also informs advice given to individual patients/families during the consultation. Most symptoms however, and many diseases experienced in the community are managed without the involvement of health care professionals.18–20 Data from health care service usage provide only a small piece of the picture about illnesses managed largely outwith these services. The registered populations of primary health care professionals such as general practitioners in some countries, offer valuable opportunities to understand population denominators, and to study how people use clinical services.
Purpose of primary care epidemiology
Like all branches of epidemiology, a fundamental purpose of primary care epidemiology is to contribute to improved health and health care. In order to make this contribution the discipline needs to answer a wide variety of questions through descriptive and analytical studies conducted at different levels of the health care system (Table 1). Hence, primary care epidemiology informs clinical decisions about: individual patients seen in primary care; the registered populations for which practitioners are responsible (where applicable); and, by extrapolation from well designed studies, the general population. It does this by:
- Improving understanding of patterns and clinical significance of common symptoms and conditions seen in primary care.
- Providing information that can optimise the efficient use of primary care services.
- Providing a framework for the design and targeting of feasible and acceptable interventions.
|
Improving understanding of patterns and clinical significance of common symptoms and conditions seen in primary care
Primary care epidemiology is needed to describe: the incidence, prevalence, severity and natural history (duration, remission and recurrence) of symptoms and signs, and of defined illnesses occurring in the community; how these problems vary among different groups within the community (e.g. by age, gender, socio-economic status, ethnicity, place of residence); and how these problems cluster or relate to each other. The long-term continuous nature of primary care, often to whole families, highlights both the need for, and opportunities available, to consider the complex genetic, physical, psychological, social and cultural influences on the onset and natural history of common symptoms and illnesses. Knowledge about the natural history of symptoms can inform the development of effective evidence-based interventions, for instance leaflets about cough that reduce re-consultation rates.21 Deeper understanding of the psycho-social factors associated with beliefs about back pain and its primary care management, has led to public health education campaigns that reduce work-related absenteeism and improve beliefs,22 effects that remain apparent after three years.23
Studies have repeatedly shown that most symptoms experienced in the community are managed without seeking health care (the so-called symptom iceberg phenomenon).24 The size of the visible (presented to health care services) and submerged (non-presented) proportions varies for different symptoms. Although sometimes perceived as a phenomenon only related to mild, self-limiting health problems, accumulating evidence shows that large submerged proportions also occur for symptoms of serious disease such as angina,25,26 chronic back pain27 and asthma.28 The long-term consequences may be important,26 although much more work is needed using a variety of outcome measures such as comorbidity, mortality, quality of life, economic costs and use of health care services for other conditions. This work will identify conditions where earlier intervention may be beneficial, as well as identify situations where the medical profession may best leave well alone.29,30 For example, the introduction of genetic tests without appropriate epidemiological evaluation in primary care may give rise to inappropriate ‘medicalisation’, either because ‘illness’ is diagnosed decades before symptoms appear (e.g. Huntingdon's chorea) or because of incomplete penetrance of genes (e.g. only 1% of homozygotes develop frank haemochromatosis).31
Currently little is known about the predictive value of different symptoms or clusters of symptoms experienced in the community, particularly for serious disease. Given the high frequency of many symptoms, and the low incidence of serious disease in the general population, especially among younger people, positive predictive values of individual symptoms are likely to be very low. For example, rectal bleeding is rarely associated with bowel cancer,32 and headaches with brain tumour.33 Research is needed into whether low predictive values such as these can be improved by looking at clusters of symptoms, in association with individual and family characteristics such as age, gender, family history and comorbidity. Without such information, referral guidelines, such as those for suspected cancer34 are likely to remain inefficient.
Although much of the work of primary care epidemiologists necessarily relates to symptoms, some of it needs to consider conditions seen and managed almost exclusively in primary care, such as guttate psoriasis, whose natural history remains poorly understood. Chronic pain is an example of a common, disabling condition whose epidemiology was, until recently, only understood through hospital clinic attendees.35,36 These patients represent a small, atypical proportion of all those experiencing the problem in the community, with very different levels of pain severity, duration and aetiological factors.37 Recent primary care-based research38–40 has provided better understanding of the prevalence and determinants of chronic pain in the community, offering new opportunities to target potentially useful interventions in primary care (including more efficient secondary care referral).
Providing information that can optimise the efficient use of primary care services
Governments throughout the world are striving to provide health care systems which are safe, patient-centred, timely, effective, efficient and equitable.9 In some countries, such as the Netherlands, Spain, Denmark, Norway and the United Kingdom (UK), primary care has a central referral (gatekeeping) role. Other countries, for instance France, Germany, India, and the United States of America, provide unrestricted (parallel) access to primary and secondary care services. The scope and purpose of primary care epidemiology is significantly influenced by the extent to which primary care provides entry to the health care system, and its comprehensiveness, community/population orientation, and degree of centralisation and computerisation. Some questions that can be answered by primary care epidemiology will be system-specific, for example, who is using health care services, when and why? Others will be more generic, for instance, what is the predictive value of (usually undifferentiated) symptoms first presented to health care professionals (whether they be a nurse, community pharmacist, general practitioner or hospital specialist)? Primary care systems with registered lists of patients are much more able to look at issues at the community level (Table 1) than those without, as the lists provide the necessary denominators (populations at risk) for study.
Whatever the health care system, demand for primary care services is huge. For example, in the UK there are estimated to be more than 250 million consultations with a general practitioner each year41 and hundreds of millions more with other members of the primary health care team. With such large numbers, small shifts in demand (of only a few percentage points) can greatly relieve or strain the system. As well as describing what proportion of people with particular symptoms or clusters of symptoms seek health care, primary care epidemiology provides information about how such usage varies according to the characteristics of individuals and their symptoms, and the availability of different health care services.
Changes in the shape of primary care, and the range of symptoms and illnesses managed by different members of the primary health care team, can influence understanding of: the epidemiology of disease in the community; the need for different health care services; and how primary care epidemiology is done. For example, the UK government has recently encouraged more self-management of self-limiting illness42 and introduced new primary care services such as nurse-led telephone advice lines (in and out of hours) and walk-in clinics.42,43 These changes may have reduced the number of people seeing their general practitioner because of self-limiting illnesses such as upper respiratory tract infections or influenza. As a consequence, surveillance systems or studies based solely on general practitioner consultations will give an erroneous picture of the health burden arising from these conditions. Furthermore, GPs are trained to manage uncertainty in clinical diagnosis. If other members of the primary health care team, with different training in handling uncertainty, manage a larger proportion of patients seen in primary care, demands on GPs and secondary care specialists may change. Primary care epidemiology is needed both for the evaluation of whether primary care services are adequate, appropriate, equitable, effective and efficient,44 and, where primary care is the gatekeeper to specialist services, to help shape secondary care.
Primary care epidemiology can also inform policy makers about the appropriateness and effectiveness of new medical developments. For example, a study of the epidemiology of upper respiratory tract infections showed that patients with symptoms of the influenza virus are more likely to consult their general practitioner earlier than those with other infections, although usually not before a median of 3 days of symptoms starting.45 In order to provide effective new treatments that have to be initiated within 1–2 days of symptom onset,46 patients with symptoms suggestive of influenza will need to be seen in general practice earlier (risking the swamping of services), or alternative means of providing treatment found.
Providing a framework for the design and targeting of feasible and acceptable interventions
In addition to providing information about the targeting of interventions, primary care epidemiology provides a framework for deciding when to apply them, and how to evaluate them. The point at which health care is sought during an episode of illness, or exacerbation of symptoms, may influence apparent response to treatment. It has been suggested that patients with chronic pain seek help at times when the pain exceeds its usual severity.47 If people seek help when pain severity is at its peak, subsequent follow-up may observe improvement that might have occurred naturally (through the cyclical nature of chronic pain) regardless of the effectiveness of any intervention. On the other hand, if patients seek help when pain severity is increasing, follow-up may indicate a worsening of symptoms in spite, perhaps, of an effective intervention. Research is required to understand the level of ‘regression dilution bias’48 affecting studies of fluctuating symptoms such as pain or fatigue. As well as forming an essential element in the analysis of such studies, knowledge of the epidemiology of symptoms will assist in their planning. For example, a sampling strategy for randomized controlled trials (RCTs) will be more efficient with detailed knowledge of sub-groups that should be studied, such as individuals at greatest risk of developing serious pathology, or at the point on any severity gradient or cycle when treatment is most likely to be required. This information is also required to determine ‘caseness’ in continuous or symptomatically defined conditions, such as hypertension, headache or chronic fatigue syndrome, and thus the point at which an individual is appropriate for recruitment to a RCT.
Outcome measures for use in primary care need to reflect the complexity of conditions under investigation, and may not currently exist. The symptomatic nature of many primary care conditions, requires the development of outcome measures that are patient-centred, and which frequently include social and psychological as well as physical factors. Primary care epidemiology helps with their development by providing information about how much improvement would need to be observed, beyond any natural improvement, for an intervention to be deemed to have produced a clinically significant result. Such information is important for determining the sample size requirements of RCTs.
Challenges and opportunities
Primary care epidemiology faces a number of challenges. These include methodological issues such as the development of valid case identification and outcome measures for use in the community; conflicts between the need for data collection and data protection; problems of sharing information between different primary health care team professionals, and health care sectors; making sense of diagnostic uncertainty; problems of funding, especially for the important context-setting, descriptive work, which may not be perceived as being immediately useful to clinical services; inadequate research capacity; and the need to cross boundaries, with a view to inter-disciplinary working that combines the expertise of generalist and specialist clinicians, epidemiologists, statisticians, and information, laboratory and other scientists.
Primary care epidemiologists also need to rise to the challenges of engaging the public in their work, so that they research with, rather than on the population.49 Related to this is the challenge of respecting data privacy, while gaining appropriate access to personal and medical data held in primary care, both for primary research and for linkage with other datasets. In many countries, recent changes in data protection legislation has made epidemiological data collection increasingly difficult (if not almost impossible in some places). Epidemiologists need to engage with legislators to find ways of conducting this important work, and with the public to ascertain their views on the use of their personal information for medical research. Increasing use of information technology within different parts of primary care, and the development of secure, confidential, data encryption systems, allow large quantities of personal data to be handled anonymously. Separate, parallel developments allow reliable linkage of these data with other information derived from research and routine secondary health care. However, the research needs of clinical information needs to be considered prospectively when designing systems, since retrospective attempts to link data are more complex and may not be successful50.
The challenges notwithstanding, there are a large number of opportunities, including increasing recognition that epidemiology is an important area of clinical research within primary care51 and a growing cadre of researchers undertaking increasing amounts of applied and methodological research. In many countries another fundamental opportunity is the large proportion of the population registered for primary care services (e.g. around 96% in the UK), with their personal and medical details retained, in varying degrees, on computerised databases. This allows the straightforward identification of individuals to approach for participation in research,52 as proposed for the UK Biobank study.53 Furthermore, these databases are relatively stable, and provide the opportunity for long-term longitudinal research,40,54 for family studies,55 for easy identification of control (non-diseased or non-exposed) groups and for assessment of non-response bias using aggregated data.56 Many of these studies require clear definition of the denominator of the study, and clarity about how the relevant data were assembled. In spite of data protection issues, the quality and quantity of clinical information from primary care in many countries has increased greatly, partly because of the ever decreasing costs of computer systems. In addition, routine clinical data collected in primary care has sometimes been enhanced by linkages with secondary care clinical datasets.
Conclusion
Primary care epidemiology is not new, but its scope and potential is wide and increasing, requiring expanded investment of personnel and resources. Nor is it an isolated discipline, working as it does with other epidemiological, scientific and clinical research. However, primary care epidemiology can make a distinct contribution to our understanding of health, illness and health care utilisation. In comparison with other medical disciplines, primary care has often been the poor relation in terms of research input and output,57 even though, in many countries, primary care manages most of the illness presented to health care services. We remain confident that the opportunities for primary care epidemiology are greater than the challenges. With careful development, the full potential of primary care epidemiology to contribute to the delivery of efficient and effective health care services in the community, primary and secondary care environments might be realised.
Acknowledgments
We are grateful to the many colleagues who have helped shape our ideas for this paper, including participants in workshops held in January 2005 (Meeting of Scottish University Departments of General Practice) and July 2005 (Society for Academic Primary Care), and in particular Professor Peter Croft, Dr George Peat and Dr Kate Dunn.
References
1 Sackett DL, Haynes BR, Tugwell P, Guyatt GH. Clinical Epidemiology: a Basic Science for Clinical Medicine, 2nd edn. London: Lippincott, Williams and Wilkins; 1991.
2 Mullan F. Community-oriented primary care: epidemiology's role in the future of primary care. Public Health Rep 1984; 99: 442–445.[ISI][Medline]
3 Mullan F, Nutting PA. Primary care epidemiology: new uses of old tools. Fam Med 1986; 18: 221–225.[Medline]
4 Abramson JH. Application of epidemiology in community oriented primary care. Public Health Rep 1984; 99: 437–441.[ISI]
5 Hart JT. The marriage of primary care and epidemiology: the Milroy lecture, 1974. J R Coll Physicians Lond 1974; 8: 299–314.[Medline]
6 Pickles WN. Epidemiology in Country Practice. Bristol: John Wright and Sons; 1939.
7 Fry J. Common Diseases. Lancaster: MT Press; 1979.
8 Hodgkin K. Towards Earlier Diagnosis. A Guide to Primary Care, 5th edn. Edinburgh: Churchill Livingstone; 1985.
9 Green LA, Phillips RL, Fryer GE. The nature of primary medical care. In Jones R, Britten N, Culpepper L, Gass D, Grol R, Mant D, Silagy C (eds). Oxford Textbook of Primary Medical Care. Oxford: Oxford University Press; 2004, Vol 1, 3–10.
10 Last RJ. A Dictionary of Epidemiology, 4th edn. Oxford: International Epidemiological Association; 2001.
11 Kroenke K. Symptoms and science: the frontiers of primary care research. J Gen Intern Med 1997; 12: 509–510.[CrossRef][ISI][Medline]
12 Kroenke K. Studying symptoms: sampling and measurement issues. Ann Intern Med 2001; 134: 844–853.
13 Komaroff AL. ‘Minor’ illness symptoms: the magnitude of their burden and of our ignorance. Arch Intern Med 1990; 150: 1586–1587.[CrossRef][ISI][Medline]
14 Kroenke K, Mangelsdorff AD. Common symptoms in ambulatory care: incidence, evaluation, therapy, and outcome. Am J Med 1989; 86: 262–266.[CrossRef][ISI][Medline]
15 Kroenke K, Spitzer RL, Williams JBW et al. Physical symptoms in primary care. Predictors of psychiatric disorders and functional impairment. Arch Fam Med 1994; 3: 774–779.[Abstract]
16 Komaroff AL, Fagioli LR, Doolittle TH et al. Health status in patients with chronic fatigue syndrome and in general population and disease comparison groups. Am J Med 1996; 101: 281–290.[CrossRef][ISI][Medline]
17 Speckens AE, Van Hemert AM, Bolk JH, Rooijmans HG, Hengeveld MW. Unexplained physical symptoms: outcome, utilization of medical care and associated factors. Psychol Med 1996; 26: 745–752.[ISI][Medline]
18 Hammond EC. Some preliminary findings on physical complaints from a prospective study of 1,064,004 men and women. Am J Public Health 1964; 54: 11–23.
19 Verbrugge LM, Ascione FJ. Exploring the iceberg. Common symptoms and how people care for them. Med Care 1987; 25: 539–569.[CrossRef][ISI][Medline]
20 Wyke S, Hunt K, Ford G. Gender differences in consulting a general practitioner for common symptoms of minor illness. Soc Sci Med 1998; 46: 901–906.[CrossRef][ISI][Medline]
21 Macfarlane JT, Holmes WF, Macfarlane RM. Reducing re-consultations for acute lower respiratory tract illness with an information leaflet: a randomised controlled study of patients in primary care. Br J Gen Pract 1997; 47: 719–722.[ISI][Medline]
22 Buchbinder R, Jolley D, Wyatt M. Population based intervention to change back pain beliefs and disability: three part evaluation. BMJ 2001; 322: 1516–1520.
23 Buchbinder R, Jolley D. Population based intervention to change back pain beliefs: three year follow up population survey. BMJ 2004; 328: 321.
24 Last J. The iceberg: ‘completing the clinical picture’ in general practice. Lancet 1963; 282: 28–31.[CrossRef]
25 Richards H, McConnachie A, Morrison C, Murray K, Watt G. Social and gender variation in the prevalence, presentation and general practitioner provisional diagnosis of chest pain. J Epidemiol Community Health 2000; 54: 714–718.
26 Owen-Smith V, Hannaford PC, Elliott AM. Increased mortality among women with Rose angina who have not presented with ischaemic heart disease. Br J Gen Pract 2003; 53: 784–789.[ISI][Medline]
27 Croft PR, Macfarlane GJ, Papgeorgiou AC, Thomas E, Silman AJ. Outcome of low back pain in general practice: a prospective study. BMJ 1998; 316: 1356–1359.
28 Frank P, Ferry S, Moorhead T, Hannaford P. Use of a postal questionnaire to estimate the likely under-diagnosis of asthma-like illness in adults. Br J Gen Pract 1996; 46: 295–297.[ISI][Medline]
29 Coyne JC, Klinkman MS, Gall SM, Schwenk TL. Short term outcomes of detected and undetected depressed primary care patients and depressed psychiatric patients. Gen Hosp Psychiatry 1997; 19: 333–343.[CrossRef][ISI][Medline]
30 Mayor S. Pfizer will not apply for a licence for sildenafil for women. BMJ 2004; 328: 542.
31 Melzer D, Zimmern R. Genetics and medicalisation. BMJ 2002; 324: 863–864.
32 Fijten GA, Blijham GH, Knottnerus JA. Occurrence and clinical significance of overt blood loss per rectum in the general population and in medical practice. Br J Gen Pract 1994; 44: 320–325.[ISI][Medline]
33 Boiardi A, Salmaggi A, Eoli M, Lamperti E, Silvani A. Headache in brain tumours: a symptom to reappraise critically. Neurol Sci 2004; 25(Suppl. 3): S143–S147
34 Available at: http://www.show.scot.nhs.uk/sehd/cancerinscotland/Documents/Scottish%20Referral%20Guidelines%20for%20suspected%20cancer.pdf (accessed 8 March 2005).
35 Davies HTO, Crombie IK, Macrae WA, Rogers KM. Pain clinic patients in North Britain. Pain Clinic 1992; 5: 129–135.
36 Becker N, Thomsen AB, Olsen AK, Sjogren P, Bech P, Eriksen J. Pain epidemiology and health related quality of life in chronic non-malignant pain patients referred to a Danish multidisciplinary center. Pain 1997; 73: 393–400.[CrossRef][ISI][Medline]
37 Crombie IK, Davies HTO. Selection bias in pain research. Pain 1998; 74: 1–3.[CrossRef][ISI][Medline]
38 Verhaak PF, Kerssens JJ, Dekker J, Sorbi MJ, Bensing JM. Prevalence of chronic benign pain disorder among adults: a review of the literature. Pain 1989; 37: 215–222.[CrossRef][ISI][Medline]
39 Gureje O, Von Korff M, Simon GE, Gater R. Persistent pain and well-being. A World Health Organization study in primary care. J Am Med Assoc 1998; 280: 147–151.
40 Elliott AM, Smith BH, Penny KI, Smith WC, Chambers WA. The epidemiology of chronic pain in the community. Lancet 1999; 354: 1248–1252.[CrossRef][ISI][Medline]
41 Available at: http://www.rcgp.org.uk/information/publications/information/PDFInfo/03%20-%20APRIL04.pdf (accessed 8 March 2005).
42 Choice and Opportunity. Primary Care: the future. Department of Health, 1996.
43 Advancing self-care. Helping people take care of their own health. Propriety Association of Great Britain self-care review 2003: London.
44 Anonymous. Epidemiology for primary health care. Int J Epidemiol 1976; 5: 224–225.[ISI][Medline]
45 Nicholson KG, Kent J, Hammersley V, Cancio E. Acute viral infections of upper respiratory tract in elderly people living in the community: comparative, prospective, population based study of disease burden. BMJ 1997; 315: 1060–1064.
46 The MIST study group. Randomised trial of efficacy and safety of inhaled zanamivir in treatment of influenza A and B virus infections. Lancet 1998; 352: 1877–1881.
47 Roland M, Morris R. A study of the natural history of low back pain. Spine 1983; 8: 145–150.[CrossRef][ISI][Medline]
48 MacMahon S, Peto R, Cutler J et al. Blood pressure, stroke, and coronary disease. Part 1, prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet 1990; 335: 765–774.[CrossRef][ISI][Medline]
49 Watt G. William Pickles lecture. General practice and the epidemiology of health and disease in families. Br J Gen Pract 2004; 54: 939–944.[ISI][Medline]
50 Sullivan FM, Pell JP, Sweetland M, Morris AD. How could primary care meet the informatics needs of UK Biobank? A Scottish proposal. Inform Prim Care 2003 11: 129–135.[Medline]
51 Medical Research Council. MRC Topic Review: Primary Health Care. London: MRC; 1997.
52 Murray L, Johnston B, Lane A et al. Relationship between body mass and gastro-oesophageal reflux symptoms: the Bristol Helicobacter Project. Int J Epidemiol 2003; 32: 645–650.
53 Available at: www.biobank.ac.uk.
54 Elliott AM, Smith BH, Hannaford PC, Smith WC, Chambers WA. The course of chronic pain in the community: results of a four-year follow-up study. Pain 2002; 99: 299–307.[CrossRef][ISI][Medline]
55 Upton MN, McConnachie A, McSharry C et al. Intergenerational 20-year trends in the prevalence of asthma and hay fever in adults: the Midspan family study surveys of parents and offspring. BMJ 2000; 321: 88–92.
56 Boardman HF. Headache in primary care: epidemiology, management, and use of health care services. (PhD Thesis) Keele: Keele University: 2003.
57 Mant D, Del Mar C, Glasziou P, Knottnerus A, Wallace P, van Weel C. The state of primary-care research. Lancet 2004; 364: 1004–1006.[CrossRef][ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||