ECDC HEALTH INFORMATION

Pandemic (H1N1) 2009 planning assumptions to end of May 2010 for EU/EEA countries

11 November 2009

What are pandemic planning assumptions?

Planning assumptions represent recommended figures that those preparing services should ensure they can deal with. They represent the ‘reasonable worst case’ that could be expected. They are not predictions, and actually represent a quite unlikely scenario. They represent a group of scenarios that national and local planners should prepare for if they want to have confidence that they will be able to cope with the peak demand in a pandemic. They also ensure a degree of standardisation across large populations. Pandemic planning assumptions have some special features, notably they have to recognise the heterogeneity of influenza transmission – so that even in a small country some areas may be heavily affected but others can get off quite lightly– and that local epidemic waves differ somewhat from what is observed nationally.

Attack rates

The experience in the Southern Hemisphere epidemics and modelling based on this suggests overall infection rates in the range of 10–40% when both symptomatic and asymptomatic infections are combined. Taking the conventional figure of 50% of influenza infections being asymptomatic suggests clinical attack rates in the range of 5–20% [2]. Detailed analysis of the multiple wave epidemic in the United Kingdom is consistent with these assumptions. These relatively low attack rates are consistent with significant background levels of immunity in the adult population that increase with age. Hence clinical attack rates can be expected to be twice as high in people under 16 years than it is in older people. This is consistent with serological information from Australia and (unpublished) surveys in the UK [8]. The limited immunity in younger adults is considered to arise from multiple exposures to the seasonal H1N1 virus, while the higher levels in older adults due to exposure to the pre-1957 seasonal H1N1 virus (which was replaced by the 1957 pandemic virus). Experience in the Southern Hemisphere is that, in most countries where there is good preparation, overall absenteeism is no higher than in the other winters.

People seeking care – Primary care consultations

This is the most difficult parameter to estimate at European level, not least because of the variability across European countries of the arrangements for delivering primary care. An estimate previously cited is that around 15% of those with clinical symptoms will seek primary care [5]. However, the relatively mild nature of infection in many individuals may reduce this proportion. Conversely, a number of countries are expecting primary care to deliver antivirals, which will tend to increase the proportion of people seeking care – although some countries are using call lines or pharmacies to provide antivirals to reduce pressure on primary care. Additionally, a number of countries are providing pandemic vaccines via primary care. For these reasons, no useful EU/EEA figure can be given and countries may seek to estimate their own values. However, the crucial point is that it is important to have primary care and hospitals working together to ensure that hospitals and emergency services are not suddenly overloaded, as has happened in some settings outside Europe. The risk of this rises especially during weekends and holidays, when primary care may reduce the services available. Countries may need to make special provisions for this.

Hospitalisations

Given the difficulty in identifying clinical cases, hospitalisation rates per head of population are a more useful indicator of severity for the current pandemic than case hospitalisation proportions. Analyses of Northern and Southern Hemisphere countries suggest overall hospitalisation rates (per capita) of 1 per 105 to 25 per 105 population) [2], though an important complication is that the criteria for hospitalisation vary from country to country, and that is likely to apply also in Europe. Experience from the UK suggests rather higher rates of hospitalisation, which would lead to overall rates possibly as high as 100 per 105. This higher figure is the planning assumption estimated attack rate combined with data on hospitalisations and may be due to different hospitalisation criteria or definition of a pandemic case. Hence a maximum reasonable estimate would be of the order of 100 per 105 population. An additional justification for this higher planning level than suggested by the experience in the Southern Hemisphere is that there are now some indications from North America (where the pandemic is running ahead of that in Europe) that the Northern Hemisphere winter epidemics could result in more hospitalisations and numbers of severe cases than what was seen in Australia and New Zealand. The United States is seeing higher rates of influenza-like illness than in relation to the most recent influenza season and mortality due to influenza or pneumonia that is above normal variation (though, as yet, well below what was seen in more severe seasonal influenza seasons) [10]. On average, about 10–20% of hospitalisations require advanced critical care including intensive care. However, because of the length of stay in intensive case is generally greater than the length of stay for general influenza hospitalisations, intensive care patients may represent up to 25% of the patients in hospital on any given day [2,7].

Mortality

Again, given the difficulty in identifying clinical cases, mortality rates (deaths per head of population) are a more useful indicator of severity for the current pandemic than the case fatality. Analysis of the Southern Hemisphere epidemics suggests mortality rates of between 0.4 per 105 and 1.4 per 105 population. Taking account of infection attack rates might suggest a maximum reasonable estimate of up to 3 per 105. With the usual assumption of 50% of cases showing clinical symptoms, this corresponds to a case fatality of less than 0.02%. These results are also consistent with analyses from the UK. In England and Wales there has been no detected effect on all-cause mortality at a population level during the epidemic to date [11]. However, it needs to be remembered that this is mortality additional to any underlying mortality that can also rise in the winter due to other viruses, cold weather, etc.

Persons affected – Risk groups

pandemic strain are especially concentrated in certain risk groups. The proportions vary from setting to setting [2]. Considering deaths in the UK 70–80% were persons in the risk groups (people with certain chronic diseases plus pregnant women) [12]. These percentages are lower when considering patients in intensive care and lower still in lower levels of hospital care, though here there is variability that comes from the criteria used for hospitalisation. Considering hospitalised cases in Ireland, 49% of hospitalised cases belonged to a risk group and 58% of those in ICU were risk groups patients [13]. In the Netherlands, so far, 62% of hospitalised cases, 66% of ICU cases and 80% of deaths were in people belonging to a pandemic risk group [14].

Quantitative assessment for young children

Clinical attack rates can be up to twice as high for the age group < 16 years compared with the elderly group, due to a combination of lower levels of prior immunity and higher contact rates in the young [7,10–12]. In England and in the Netherlands, around 20% of patients admitted to hospital were < 5 years old; this percentage has remained stable for several months. In several countries in the Northern and Southern Hemispheres, the rate of hospital admission in the < 5 years age group is at least 2–3 times higher than in the 5+ years age groups when comparing rates per 100 000 population. In UK, numbers of up to 4.5 times higher have been observed for the < 5 year group than of the 5+ year group.

Shape of the pandemic wave

For planning purposes, a local pandemic wave may last around 15 weeks from start to finish, but at a national level this may be longer – how much longer will depend on size, geography and travel patterns in specific countries. For planning purposes, it can be assumed that about 20% of all cases will occur in each of the two weeks either side of the peak. It also needs to be remembered that local peaks can be considerably higher than national peaks. Of course, they can also be lower but that is not relevant to planning assumptions. For an illustration, see Figure 1.

References

1. European Centre for Disease Prevention and Control (ECDC). European pandemic planning assumptions (Report of a meeting, January 2009).
2. World Health Organization (WHO) annotation. Weekly Epidemiological Record (WER) (accepted for publication – November 2009)
3. Norwegian Institute of Public Health. Planning assumptions for Influenza A(H1N1) July 2009.
4. Norwegian Institute of Public Health. Planning assumptions for Influenza A(H1N1) September 2009.
5. UK Cabinet Office and Department of Health. Swineflu UK Planning Assumptions – 16 July 2009.
6. UK Cabinet Office and Department of Health. Swineflu UK Planning Assumptions – 3 September 2009.
7. UK Cabinet Office and Department of Health. UK Swineflu: Guidance for planners 2nd update, 22 October 2009.
8. Greenberg ME, Lai MH, Hartel GF, Wichems CH, Gittleson C, Bennet J. Response after one dose of a monovalent, unadjuvanted influenza A (H1N1) 2009 vaccine – preliminary report. NEJM 10 September 2009 (10.1056/NEJMoa0907413).
9. European Centre for Disease Prevention and Control (ECDC). Revised pandemic 2009 planning assumptions for Europe.
10. Centers for Disease Control and Prevention (CDC). FluView, 2009-2010 Influenza Season Week 43 ending October 31, 2009. http://www.cdc.gov/flu/weekly/
11. UK Health Protection Agency. Weekly National Influenza Report, 05 November 2009 (week 45).
12. European Centre for Disease Prevention and Control (ECDC). Risk Assessment Pandemic (H1N1) 2009, Version 6, 6 November 2009.
13. Cullen G, Martin J, O’Donnell J, Boland M, Canny M, Keane E, et al. Surveillance of the first 205 confirmed hospitalised cases of pandemic H1N1 influenza in Ireland, 28 April – 3 October 2009. Euro Surveill. 2009;14(44):pii=19389.
14. Rijksinstituut voor Volksgezondheit en Milieu (RIVM). Overzicht verspreiding Nieuwe Influenza A (H1N1) in Nederland 6 november 2009.