|CM Regan1, W Sopwith1, M Ashton1, Q Syed1, M Painter2, K Mutton3, K Paver3
1 Communicable Disease Surveillance Centre (North West), Chester, United Kingdom
|A local sentinel network of general practitioners has been established in the north west of England for the surveillance of influenza. In the 2001–02 winter, consultation rates for influenza-like-illness (ILI) were low but the surveillance network was able to demonstrate sub-regional variations in the timing of peak influenza activity, and the infection of different age groups. This suggests the network can contribute to better planning to winter pressures on the North West health service.|
T he North West Region of England is geographically and culturally diverse with a population of around 13% of that of England and Wales (6.9 million) in four Counties: Cheshire, Merseyside, Greater Manchester and Lancashire. Health is poorer than the national average, with life expectancy over one year lower than the England and Wales average. The Standardised Mortality Ratio (SMR) for all causes of mortality is 112 for males and 110 for females, the highest ratios of any region in England and Wales (national ratio =100 for both sexes) (1).
The seasonal burden of respiratory infections, particularly amongst the elderly, can put significant pressure on local health services during winter months in England and Wales. In response to such pressure, a pilot sentinel surveillance network was established to monitor the local impact of seasonal influenza in the City of Liverpool in 1997 through consultations in General Practice (2). The surveillance network detected a localised epidemic of influenza A during the winter of 1997–98. This experience led the local Health Authority and the Communicable Disease Surveillance Centre, North West (CDSC-NW) to extend the sentinel practice network across the region in order to better detect the seasonal distribution of influenza, and so advise the local health services on their planning for winter pressures. Eighty five practices covering all the counties in the north west of England were enrolled for the 2000–01 and 2001–02 winter seasons and data from 2001–02 also included results of virological analysis. We report here surveillance data from the second winter season.
Recruitment of General Practices
General practitioners in the North West Region of England were invited to participate in the surveillance scheme by their local Consultants in Communicable Disease Control (CsCDC). Eighty five practices were recruited, their case load covering approximately 10% of the Region’s population (approximately 651 000 people). The number of practices recruited varied by County (Figure 1) according to the county’s population, the highest number coming from urban Greater Manchester (37 practices). The proportion of each county’s population covered by the recruited practices varied between 8% (Lancashire) and 14% (Cheshire).
Clinical case definition and data collection
General practitioners provided weekly data returns for patients consulting with clinically diagnosed influenza and influenza-like illness (ILI). The clinical case definition of influenza and ILI used was the same as that used in the national weekly returns service undertaken by the Royal College of General Practitioners (3). This definition is the presence of four of the following in an epidemic situation, and six of them in a non-epidemic situation: i) cough, ii) rigors or chills, iii) fever, iv) myalgia, v) prostration and weakness, vi) contact with influenza, vii) redness of nasal mucous membrane and throat and viii) sudden onset. The data collected were age of patient and week of first consultation for the case defining illness. Weekly returns from participating sentinel practices were sent to the Infection Control and Surveillance Unit in Manchester Health Authority, where the data were collated.
All data received were entered onto a database and crude rates of consultation for influenza and influenza-like illness (ILI) were calculated using the case load of each surgery. These rates were used as proxy indicators for the incidence of influenza and ILI in the Region. Data were made instantly available on a dedicated influenza surveillance website accessed over the NHS Net (restricted access network for healthcare professionals in the UK). Consultation rates and details of confirmed isolates of influenza virus (reported by PHLS North West) were also published in a weekly regional report distributed by post and email. Data were made available to the general public on the North West Public Health Observatory website (www.nwpho.org.uk). Consultation rates were presented in relation to levels of ‘epidemic’, ‘higher than average’ and ‘normal’ seasonal activity, which were defined using retrospective seasonal data for England and Wales (4).
In 2001–02 the influenza season in the north west was between weeks 52 and 10 with a peak in consultation rates in week 5. The baseline rates fluctuated between 14 and 18 consultations per 100 000 population, and peaked at 50 consultations per 100 000. The time and magnitude of the peak were very similar in the 2000–01 season, and in both years the rates of consultations never rose above the lower level of normal seasonal activity. This picture was very similar in England and Wales as a whole (figure 2).
Isolation of influenza A virus (in 2001-02) peaked between weeks 4 and 7 in the North West (Figure 3), coinciding with the peak in consultation rates. Isolations of virus increased as a proportion of total respiratory specimens taken over these weeks (Figure 4). In total, influenza A was isolated from 51 individuals in the north west, 50% of whom were less than a year old. From nineteen isolates that were characterised, sixteen were antigenically related to an influenza A Panama-like strain (H3N2), and three to an influenza A New Caledonia-like strain (H1N1). The composition of the influenza vaccine for the 2001–02 season included both these strains.
Subregionally, there was variation in the period of peak consultations between counties with peaks in week 2 in Lancashire and Cheshire, weeks 4 and 5 in Greater Manchester and week 6 in Merseyside (figure 5). Rates of consultation were consistently highest in Lancashire throughout the season, even in the baseline period of weeks 40 to 51. Baseline rates of consultation in the other counties were more similar to each other. After Lancashire, Greater Manchester had the highest peak of consultations and Merseyside the lowest (Figure 5).
During the period the influenza activity peaked, there were sharp increases in consultation rates amongst 0-4 and 5-14 year olds in the Region as a whole (Figure 6). Apart from these peak weeks, consultation rates in older age groups (15+) were consistently higher than in younger ones between weeks 44 and 9. This age distribution during periods of peak consultations ‰ ‰ varied between Counties. The highest rates of consultations in Greater Manchester during the peak week of influenza were amongst 0–4 year olds and 5–14 year olds, whereas in Lancashire and Cheshire, the highest rates were amongst those over 15 years old and particularly those between 45 and 64 (Table). In Merseyside, the highest rates were also amongst 45– 64 year olds during the peak week but rates were also relatively high (compared to other age groups) amongst 5–14 year olds.
Table. Summary of County data
Discussion and conclusion
Regional variation in the epidemiology of seasonal influenza has been described in the United Kingdom but accurate assessment is hampered when sentinel surveillance networks cover small populations spread over large areas (2,5). These geographical and demographic variations may indicate important differences in the extent and speed of transmission of influenza within a region, as well as the particular groups of people at risk. An accurate assessment of the local transmission dynamics of influenza is therefore important for those planning health services, particularly in light of the seasonal pressure exerted on hospital beds in Britain in recent winters (6).
The network of sentinel practices we have established in the north west of England has demonstrated variation in both the incidence of seasonal influenza and the timing of peak influenza activity within the region. As with the national (England and Wales) surveillance system, consultation rates for influenza and ILI in general practice were used as an indicator of disease incidence. The validity of this indicator is supported by the co-incident peaks of consultation rates and isolates of influenza A through the winter season in the North West.
The practices recruited covered around 10% of the population of each county and were widely distributed geographically. Each practice was using a common case definition for influenza and ILI and staff were trained in data collection. We are therefore confident that these data are a good representation of influenza transmission. There were, however, consistently higher consultation rates in Lancashire than in the other counties. There may be true differences in the transmission of influenza and ILI in more rural or northerly areas (such as Lancashire) or there may be systematic differences in diagnosis and reporting. Rates were not consistently higher in Cheshire, also a rural county, and colleagues in Lancashire are investigating possible reasons for the difference in reporting.
Several subregional variations were noted in this study. The timing of the main peak of infection differed across the Region and this is significant for health service planning. Information about the transmission of influenza is used by local Winter Pressure Planning Groups to plan health services, and particularly the availability of hospital beds. A system that accurately and rapidly informs such groups of imminent increases in influenza activity at a local level will be important, particularly in years of high activity. Zonal variations in the age group of cases may be particularly useful for targeting appropriate interventions across the Region, such as vaccination, public health advice and prevention of infection.
The different age peaks described in this study may reflect social differences in access to GP services in different zones, as well as differences in social mixing. It is interesting that in the socially comparative zones of Manchester and Merseyside, there are clear differences in the contribution of 0-4 year olds to the seasonal peak of infection. The reason for this difference is not known.
The sharpest increases in consultation rates through the winter season were noted amongst 0–4 and 5–14 year olds in the North West. This pattern was also seen in England and Wales as a whole (data not shown) but there were several peaks amongst these age groups, beginning in week 51, and this may reflect different periods of peak infection in different regions across the country. The contribution of these age groups to the transmission of influenza is unknown but these data suggest that variation exists within the North West Region. It is only Greater Manchester whose peak consultation rate coincides with a peak in 0–4 year old consultations and the contribution of 45–64 year olds to seasonal peaks appears to be more important in other Counties. Subsequent years of surveillance may clarify these results, and particularly the differences between rural and urban areas.
In addition to clarifying the local transmission of influenza, the north west network is able to strengthen national influenza surveillance by validating national trends. Further evaluation of the benefit of the GP sentinel practice network is planned for subsequent seasons to assess its contribution to contingency planning and disaster preparedness of health and social care services in the north west of England.
We are very grateful to all the GP practices contributing to this surveillance scheme and the staff at Manchester Health Protection Unit for the collection and entry of data. Thanks also go to all CsCDC and Directors of Public Health in the North West for their support. We also thank Dr Carol Joseph at the Respiratory Division of CDSC, Colindale and the anonymous Eurosurveillance referees for their helpful comments on this manuscript.
|References1. Department of Health. Compendium of Clinical and Health Indicators 2001 – Clinical and Health Outcomes Knowledge Base. Dataset published online (limited access) http://nww.nchod.nhs.uk. Accessed August 2002
2. Regan CM, Johnstone F, Joseph CA, Urwin M. Local surveillance of influenza in the United Kingdom: from sentinel practices to sentinel cities. Commun Dis Public Health 2002; 5(1): 17-22.
3. Fleming DM, Ross AM, Weekly Returns service annual report for 1990, Birmingham Research Unit, Royal College of General Practitioners, 1991.
4. Dedman DJ, Watson JM. The use of thresholds to describe levels of influenza activity. PHLS Microbiology Digest 1997; 14: 206-8.
5. Fleming DM, Cohen JM. Experience of European collaboration in Influenza surveillance in the winter of 1993-94. J Pub Health Med 1996;(18)2:133-42.
6. Warden J. Health secretary reports on winter crisis. BMJ. 318(7177):145, 1999