Marc Baguelin

Vaccination against pandemic influenza A/H1N1v in England: a real-time economic evaluation.

Decisions on how to mitigate an evolving pandemic are technically challenging. We present a real-time assessment of the effectiveness and cost-effectiveness of alternative influenza A/H1N1v vaccination strategies. A transmission dynamic model was fitted to the estimated number of cases in real-time, and used to generate plausible autumn scenarios under different vaccination options. The proportion of these cases by age and risk group leading to primary care consultations, National Pandemic Flu Service consultations, emergency attendances, hospitalisations, intensive care and death was then estimated using existing data from the pandemic. The real-time model suggests that the epidemic will peak in early November, with the peak height being similar in magnitude to the summer wave. Vaccination of the high-risk groups is estimated to prevent about 45 deaths (80% credibility interval 26-67), and save around 2900 QALYs (80% credibility interval 1600-4500). Such a programme is very likely to be cost-effective if the cost of vaccine purchase itself is treated as a sunk cost. Extending vaccination to low-risk individuals is expected to result in more modest gains in deaths and QALYs averted. Extending vaccination to school-age children would be the most cost-effective extension. The early availability of vaccines is crucial in determining the impact of such extensions. There have been a considerable number of cases of H1N1v in England, and so the benefits of vaccination to mitigate the ongoing autumn wave are limited. However, certain groups appear to be at significantly higher risk of complications and deaths, and so it appears both effective and cost-effective to vaccinate them. The United Kingdom was the first country to have a major epidemic in Europe. In countries where the epidemic is not so far advanced vaccination of children may be cost-effective. Similar, detailed, real-time modelling and economic studies could help to clarify the situation.

Assessment of baseline age-specific antibody prevalence and incidence of infection to novel influenza A/H1N1 2009.

OBJECTIVES The objectives of the H1N1 2009 serological surveillance project were twofold: to document (1) the prevalence of cross-reactive antibodies to H1N1 2009 by age group in the population of England prior to arrival of the pandemic strain virus in the UK and (2) the age-specific incidence of infection by month as the pandemic progressed by measuring increases in the proportion of individuals with antibodies to H1N1 2009 by age. METHODS Residual aliquots of samples submitted to 16 microbiology laboratories in eight regions in England in defined age groups in 2008 and stored by the Health Protection Agency serological surveillance programme were used to document age-stratified prevalence of antibodies to H1N1 2009 prior to the arrival of the pandemic in the UK. Functional antibodies to the H1N1 2009 virus were measured by haemagglutination inhibition (HI) and microneutralisation (MN) assays. For timely measurement of monthly incidence of infection with H1N1 2009 between August 2009 and April 2010, the microbiology serum collections were supplemented by collection of residual sera from chemical pathology laboratories in England. Monthly seroincidence samples were tested by HI only, apart from the final sera collected post pandemic in 2010, which were also tested by MN. Incidence during the pandemic was estimated from changes in prevalence between time points and also by a likelihood-based method. SETTING Eight regions of England. PARTICIPANTS Serum samples from patients accessing health care in England from whom blood samples were taken for unrelated microbiological or chemical pathology testing. INTERVENTIONS None. MAIN OUTCOME MEASURES Baseline age-specific prevalence of functional antibodies to the H1NI 2009 virus prior to the arrival of the pandemic; changes in antibody prevalence during the period August 2009 to April 2010. RESULTS Pre-existing cross-reactive antibodies to H1N1 2009 were detected in the baseline sera and increased with age, particularly in those born before 1950. The prediction of immunological protection derived from the baseline serological analysis was consistent with the lower clinical attack rates in older age groups. The high levels of susceptibility in children < 15 years, together with their mixing within school, resulted in the highest attack rates in this age group. Serological analysis by region confirms that there were geographical differences in timing of major pandemic waves. London had a big first wave among the 5- to 14-year age group, with the rest of the country reducing the gap after the second wave. Cumulative incidence in London remained higher throughout the pandemic in each age group. By the end of the second wave it is estimated that as many as 70% of school-aged children in London had been infected. Taken together, these observations are consistent with observations from previous pandemics in 1918, 1957 and 1968 - that the major impact of influenza pandemics is on younger age groups, with a pattern of morbidity and mortality distinct from seasonal influenza epidemics. CONCLUSIONS Serological analysis of appropriately structured, age-stratified and geographically representative samples can provide an immense amount of information to set in context other measures of pandemic impact in a population, and provide the most accurate measures of population exposure. National scale seroepidemiology studies require cross-agency coordination, multidisciplinary working, and considerable scientific resource. FUNDING The National Institute for Health Research Health Technology Assessment programme and the Health Protection Agency.

Age-Specific Incidence of A/H1N1 2009 Influenza Infection in England from Sequential Antibody Prevalence Data Using Likelihood-Based Estimation

Estimating the age-specific incidence of an emerging pathogen is essential for understanding its severity and transmission dynamics. This paper describes a statistical method that uses likelihoods to estimate incidence from sequential serological data. The method requires information on seroconversion intervals and allows integration of information on the temporal distribution of cases from clinical surveillance. Among a family of candidate incidences, a likelihood function is derived by reconstructing the change in seroprevalence from seroconversion following infection and comparing it with the observed sequence of positivity among the samples. This method is applied to derive the cumulative and weekly incidence of A/H1N1 pandemic influenza in England during the second wave using sera taken between September 2009 and February 2010 in four age groups (1–4, 5–14, 15–24, 25–44 years). The highest cumulative incidence was in 5–14 year olds (59%, 95% credible interval (CI): 52%, 68%) followed by 1–4 year olds (49%, 95% CI: 38%, 61%), rates 20 and 40 times higher respectively than estimated from clinical surveillance. The method provides a more accurate and continuous measure of incidence than achieved by comparing prevalence in samples grouped by time period.

Health and economic impact of the seasonal influenza vaccination programme in England

BACKGROUND The seasonal influenza vaccination programme in England targets individuals over 65 years old and in clinical risk groups. METHODS A model of influenza transmission and disease was fitted to weekly primary care consultations due to influenza in a typical pre-pandemic season (2006/2007). Different scenarios were constructed about influenza severity and how well vaccines match circulating strains to assess the impact and cost-effectiveness of the current vaccination programme. RESULTS A well-matched vaccine may reduce the incidence of laboratory-confirmed influenza illness from 8.2% (95% range 4.3-13%) to 5.9% (95% range 2.9-9.7%), with 56-73% of this due to indirect protection. The programme is likely to be cost-effective unless both low severity and poor matching is assumed. CONCLUSION The current seasonal influenza vaccination programme appears to substantially reduce disease burden and provides good value for money.

Temporal Changes in Ebola Transmission in Sierra Leone and Implications for Control Requirements: a Real-time Modelling Study

Background: Between August and November 2014, the incidence of Ebola virus disease (EVD) rose dramatically in several districts of Sierra Leone. As a result, the number of cases exceeded the capacity of Ebola holding and treatment centres. During December, additional beds were introduced, and incidence declined in many areas. We aimed to measure patterns of transmission in different regions, and evaluate whether bed capacity is now sufficient to meet future demand. Methods: We used a mathematical model of EVD infection to estimate how the extent of transmission in the nine worst affected districts of Sierra Leone changed between 10th August 2014 and 18th January 2015. Using the model, we forecast the number of cases that could occur until the end of March 2015, and compared bed requirements with expected future capacity. Results: We found that the reproduction number, R, defined as the average number of secondary cases generated by a typical infectious individual, declined between August and December in all districts. We estimated that R was near the crucial control threshold value of 1 in December. We further estimated that bed capacity has lagged behind demand between August and December for most districts, but as a consequence of the decline in transmission, control measures caught up with the epidemic in early 2015. Conclusions: EVD incidence has exhibited substantial temporal and geographical variation in Sierra Leone, but our results suggest that the epidemic may have now peaked in Sierra Leone, and that current bed capacity appears to be sufficient to keep the epidemic under-control in most districts.

Modelling the impact of local reactive school closures on critical care provision during an influenza pandemic

Despite the fact that the 2009 H1N1 pandemic influenza strain was less severe than had been feared, both seasonal epidemics of influenza-like-illness and future influenza pandemics have the potential to place a serious burden on health services. The closure of schools has been postulated as a means of reducing transmission between children and hence reducing the number of cases at the peak of an epidemic; this is supported by the marked reduction in cases during school holidays observed across the world during the 2009 pandemic. However, a national policy of long-duration school closures could have severe economic costs. Reactive short-duration closure of schools in regions where health services are close to capacity offers a potential compromise, but it is unclear over what spatial scale and time frame closures would need to be made to be effective. Here, using detailed geographical information for England, we assess how localized school closures could alleviate the burden on hospital intensive care units (ICUs) that are reaching capacity. We show that, for a range of epidemiologically plausible assumptions, considerable local coordination of school closures is needed to achieve a substantial reduction in the number of hospitals where capacity is exceeded at the peak of the epidemic. The heterogeneity in demand per hospital ICU bed means that even widespread school closures are unlikely to have an impact on whether demand will exceed capacity for many hospitals. These results support the UK decision not to use localized school closures as a control mechanism, but have far wider international public-health implications. The spatial heterogeneities in both population density and hospital capacity that give rise to our results exist in many developed countries, while our model assumptions are sufficiently general to cover a wide range of pathogens. This leads us to believe that when a pandemic has severe implications for ICU capacity, only widespread school closures (with their associated costs and organizational challenges) are sufficient to mitigate the burden on the worst-affected hospitals.

Transmission of equine influenza virus during an outbreak is characterized by frequent mixed infections and loose transmission bottlenecks

The ability of influenza A viruses (IAVs) to cross species barriers and evade host immunity is a major public health concern. Studies on the phylodynamics of IAVs across different scales - from the individual to the population - are essential for devising effective measures to predict, prevent or contain influenza emergence. Understanding how IAVs spread and evolve during outbreaks is critical for the management of epidemics. Reconstructing the transmission network during a single outbreak by sampling viral genetic data in time and space can generate insights about these processes. Here, we obtained intra-host viral sequence data from horses infected with equine influenza virus (EIV) to reconstruct the spread of EIV during a large outbreak. To this end, we analyzed within-host viral populations from sequences covering 90% of the infected yards. By combining gene sequence analyses with epidemiological data, we inferred a plausible transmission network, in turn enabling the comparison of transmission patterns during the course of the outbreak and revealing important epidemiological features that were not apparent using either approach alone. The EIV populations displayed high levels of genetic diversity, and in many cases we observed distinct viral populations containing a dominant variant and a number of related minor variants that were transmitted between infectious horses. In addition, we found evidence of frequent mixed infections and loose transmission bottlenecks in these naturally occurring populations. These frequent mixed infections likely influence the size of epidemics.

The cost-effectiveness of vaccinating pregnant women against seasonal influenza in England and Wales.

We assessed the cost-effectiveness of vaccinating pregnant women against seasonal influenza in England and Wales, taking into account the timing of vaccination relative to both the influenza season and trimester of pregnancy. Women were assumed to be vaccinated in their second or third trimester. Vaccination between September and December was found to have an incremental cost-effectiveness ratio of £23,000 per quality adjusted life year (QALY) (95% CI £10,000-£140,000) if it is assumed that infants are partially protected through their mothers, and of £28,000 per QALY gained (95% CI £13,000-£200,000) if infants are not protected. If some vaccine protection lasts for a second season, then the ratio is only £15,000 per QALY gained (95% CI £6,000-£93,000). Most of the benefit of vaccination is in preventing symptomatic episodes, regardless of health care resource use. Extending vaccination beyond December is unlikely to be cost-effective unless there is good protection into a second influenza season. Key sources of uncertainty are the cost of vaccine delivery and the quality of life detriment due to a clinically apparent episode of confirmed influenza. The cost of vaccine purchase itself is relatively low.

Different transmission patterns in the early stages of the influenza A(H1N1) v pandemic: A comparative analysis of 12 European countries

Following the emergence of a novel strain of influenza A(H1N1) in Mexico and the United States in April 2009, its epidemiology in Europe during the summer was limited to sporadic and localised outbreaks. Only the United Kingdom experienced widespread transmission declining with school holidays in late July. Using statistical modelling where applicable we explored the following causes that could explain this surprising difference in transmission dynamics: extinction by chance, differences in the susceptibility profile, age distribution of the imported cases, differences in contact patterns, mitigation strategies, school holidays and weather patterns. No single factor was able to explain the differences sufficiently. Hence an additive mixed model was used to model the country-specific weekly estimates of the effective reproductive number using the extinction probability, school holidays and weather patterns as explanatory variables. The average extinction probability, its trend and the trend in absolute humidity were found to be significantly negatively correlated with the effective reproduction number - although they could only explain about 3% of the variability in the model. By comparing the initial epidemiology of influenza A (H1N1) across different European countries, our analysis was able to uncover a possible role for the timing of importations (extinction probability), mixing patterns and the absolute humidity as underlying factors. However, much uncertainty remains. With better information on the role of these epidemiological factors, the control of influenza could be improved.

Capturing the time-varying drivers of an epidemic using stochastic dynamical systems

Epidemics are often modeled using non-linear dynamical systems observed through partial and noisy data. In this paper, we consider stochastic extensions in order to capture unknown influences (changing behaviors, public interventions, seasonal effects, etc.). These models assign diffusion processes to the time-varying parameters, and our inferential procedure is based on a suitably adjusted adaptive particle Markov chain Monte Carlo algorithm. The performance of the proposed computational methods is validated on simulated data and the adopted model is applied to the 2009 H1N1 pandemic in England. In addition to estimating the effective contact rate trajectories, the methodology is applied in real time to provide evidence in related public health decisions. Diffusion-driven susceptible exposed infected retired-type models with age structure are also introduced.