Current influenza surveillance in Scotland
The influenza (Flu) surveillance programme in Scotland centres on the complementary systems of the GP flu spotter practices (running since 1972), laboratory reporting of influenza, and information derived from the SERVIS scheme. The last combines clinical and laboratory information (since 2000) from sentinel general practices participating in the Continuous Morbidity Recording (CMR) scheme co-ordinated by the Information and Statistics Division (ISD) of NHS Services Scotland (see SCIEH Weekly Report 13 July 2004 for further summary details).
Additional information on influenza related deaths (should they occur) is also available from the General Register Office for Scotland (http://www.gro-scotland.gov.uk).
Current vaccination policy and problems posed when there is ‘mis-match’ between vaccine and circulating flu virus
In recognition of the increased risk of complication from influenza in all individuals over the age of 65, Scotland and the rest of the UK have routinely offered flu vaccine over the past four flu seasons. Current uptake in this group (72.5%) exceeds the national target (70%). Groups under the age of 65 at increased risk of complication from infection are similarly offered vaccine and this coming season a target has been set for the first time in this group (60% uptake).
The composition of the annual flu vaccine for the coming Northern Hemisphere winter flu season is determined each spring by the World Health Organization (WHO) on the basis of the current flu strains in circulation (http://www.who.int/csr/disease/influenza/vaccinerecommendations1/en/). In most years this vaccine is a very good match to the flu viruses encountered in the community and the population are afforded a good protection against these viruses by their vaccine. However in the event of a new strain emerging or difficulties in production of an effective vaccine to some strain types there can be a mismatch between the vaccine produced and the flu virus to which they are exposed. Under such circumstances the vaccine may provide suboptimal protection.
Commentary on the 2003/04 flu season, recent experience with Highly Pathogenic Avian Influenza (HPAI) and their implication for contingency planning for pandemic flu.
The 2003/04 flu season was characterised by the dominance of the Fujian strain of influenza (Influenza A/Fujian/411/2002) which predominantly affected young and adolescent children (see surveillance report , SCIEH Weekly Report 13 July 2004). While early reports of childhood deaths were received (three children under the age of 16 and one teenager under 18) these were within the range expected from previous flu seasons. Despite a mismatch between the vaccine strain types and the Fujian virus encountered, the vaccinated population did however appear to be reasonably well protected by the vaccine against the Fujian strain. This is most probably because it was related to the influenza strain that was contained in the vaccine for that season (see HPA statement at http://www.hpa.org.uk/infections/topics_az/influenza/Activity0304).
Added to this backdrop, later in the season, was the global uncertainty around the pandemic potential of a highly pathogenic avian influenza (HPAI - H5N1) that affected huge numbers in poultry flocks in a large number of south-east Asian countries (Figure 1). Fortunately this virus has thus far demonstrated only a limited ability to pass to or between human hosts and hence has not as yet developed into a virus that would fulfil all of the anticipated prerequisites for developing into a pandemic strain - the ability to produce community outbreaks of influenza (Figure 2).
With vaccination the mainstay of the global public health effort to prevent cases of influenza and minimise the morbidity and mortality from the complications of flu, new flu viruses always pose a potential threat. The fundamental point here is that if a markedly different flu virus evolves to cause significant human illness there may well be a period of four to six months following the identification of such a novel pandemic strain before a new vaccine can be manufactured in sufficient quantity to protect the population. This explains the importance of the need to have a surveillance system in each country sensitive enough to provide as early a warning of the emergence of a new pandemic flu virus strain as possible and that identifies the age groups likely to be most affected. Vaccine manufacture is initiated by the WHO global influenza reference laboratory network who undertake isolation of new viruses and subsequent candidate vaccine development. Vaccine manufacturers then develop vaccine and then produce millions of vaccine doses over a period of months.
Re-examination of pandemic plans is a routine part of the emergency planning of all public health agencies. However the factors above lent additional impetus for the re-examination of local and international plans for dealing with pandemic influenza by all public health agencies.
While the scope of such plans is enormous this article considers the strengths and limitations of current surveillance schemes and discusses ways in which surveillance of influenza could be strengthened in the future.
Rationale for considering additional surveillance methods to supplement those currently available.
The current surveillance measures provide an early alert to the emergence of influenza-like illness (ILI) presenting to general practitioners across Scotland. The GP flu spotter scheme running since 1972 provides a sensitive means to detect such increase in ILI. However ILI can be caused by a range of pathogens (para-influenza virus, respiratory syncytial virus, human meta-pneumovirus to name a few) so it is important that laboratory information is sought to confirm the diagnosis particularly early in the season. Prior to the year 2000, there was a time-lag before confirmation of influenza (or not) by routine laboratory investigation – usually a gap of one to two weeks. With the development of rapid molecular tests, the SERVIS scheme (introduced in 2000) attempted to address these deficiencies by including routine multiplex PCR testing on submitted samples (up to five per week from each practice). It also provided additional information on the age-band of cases lacking in the GP flu spotter scheme.
Such additional information has proved extremely useful in describing the differential effect of flu in any season – e.g. the increased incidence of Fujian influenza in children and adolescents in the 2003/04 season.
In recognition of the need for there to be a global response to the threat of influenza, a new development proposed for this season is the sharing of molecular typing information from Scottish isolates identified through SERVIS by hosting the laboratory isolate results on the UK sponsored section (shared with the HPA) of the Los Alamos Molecular Database. This will enable any unusual flu virus genetic sequences to be accessible to all vaccine development groups should the need arise to consider these.
While presentations to GPs are well described by the complementary systems above, they tell us little, however, of the impact of influenza on hospital services. There are no direct systems that allow prospective, timely information to be collected for this purpose. On the basis of previous experience during flu epidemics we might reasonably expect that the majority of admissions would be presenting with complications arising from influenza. During such epidemics an indirect means by which this can be gauged is provided by information collated by ISD’s System Watch (http://www.show.scot.nhs.uk/systemwatch). System Watch has for the last two years run a specific system to capture information on the total number of new admissions to hospital. By comparing this information with previous trends for the same time periods from previous years we can then see whether the current impact of influenza is creating increased pressures on in-patient bed numbers.
Options for additional means of surveillance
Hospital based surveillance
Having identified that there is no directly collected timely information on the impact of influenza on hospital admissions we could collect information on this. This would require information from paediatric hospitals and adult hospitals since during different seasons the impact of influenza could be greater on one age group than another. During the 2003/04 season a pilot study of influenza surveillance in paediatric intensive care units in Scotland was undertaken with the aim of describing the severe end of the disease spectrum of influenza in children.
The method used involved the four hospitals with paediatric intensive care facilities in Scotland from October 2003 to January 2004 being contacted every 24-48 hours for information on the number of new admissions with influenza-like illness - anonymised data being collected on admissions with a suspected or proven diagnosis of influenza.
A total of 43 cases with influenza-like illness were reported, of whom 28 were influenza A positive (either immunofluorescence or PCR positives) - 17 of these receiving intensive care (ITU) and 11 high dependency unit (HDU) care. Of those requiring ITU, 10 were female (59%). The median age was 19 months (0-172). Of the 17, eight (47%) had no known risk factors. The main mode of presentation was with respiratory symptoms, although two children fitted, and one had a dysrhythmia. The median length of stay was seven days (1-43) and the total number of bed days within ITU was 156.
Few of the children identified received antivirals as part of their clinical management. This daily real-time report on severe influenza-related illness in the paediatric population depended heavily on the co-operation of the clinical staff and the experience of medical epidemiologists at SCIEH to interpret the information collected.
However, the system only revealed the ‘tip of the iceberg’ since only the most severe cases of influenza-like illness were likely to be admitted to the ITU and children may receive high dependency care in a range of clinical settings.
A more comprehensive approach to surveillance could make use of electronic information in admitted patients or alternatively use sentinel paediatric units to provide enhanced purposeful surveillance information. Such approaches are being discussed as part of the contingency planning for routine surveillance and for pandemic planning.
Adult hospital surveillance
No current system captures information on influenza-like illness presentation in adults either presenting to or admitted to hospital. Discussions with colleagues in NHS Services (ISD) about the flexibility of System Watch have identified a potential application of the system for future development (e.g. selecting out information on the number of admissions to acute medicine or respiratory medicine). The principle attraction of this would be the ability to use routinely collected information for an alternative purpose.
Similarly, it may be that recent changes in the structure of the NHS offer an exciting opportunity to develop current surveillance. Discussion about the use of information from sentinel hospitals using presenting complaint and ‘Hospital at night’ derived information are at a preliminary stage.
Other sources of data
NHS Direct in England aggregates summary call information received into a number of key symptom groups including temperature and cough/cold groups. Using this data in the season 2003/04 valuable information was gained on the age specific pattern and magnitude of problem experienced across the whole of England. In the near future similar information will be available from NHS 24 in Scotland that will add to the picture north of the border.
Local authority education department information on pupil absence from primary and secondary school
For a number of years colleagues in England have described the annual effect of flu on boarding school pupils. This type of information is collected in the recognition that new strains of influenza can affect the young first before spreading out to involve the rest of the community. Scotland currently has no means of comprehensively describing the incidence of influenza like illness in a school population. Reporting depends on the notification of exceptional events to NHS board communicable disease and environmental health consultants by either local education departments or from the general public.
Recent advances in electronic information recording on school absenteeism has resulted in many local authorities being provided with information electronically on a daily basis. SCIEH is currently in discussion with one local authority with a view to undertaking a pilot to evaluate the usefulness of such data as an early warning on the effect on the paediatric and adolescent populations of each new flu virus as each season develops.
Our aim, in partnership with all NHS agencies responsible for the protection of public health, is to continue the effective refining of our surveillance systems to meet the challenge of such emerging infections.