[Show abstract][Hide abstract] ABSTRACT: Recent work has considered the use of densely-sampled genetic data to
reconstruct the transmission trees linking infectors and infectees in
outbreaks. Because transmission trees from one outbreak do not generalize to
future outbreaks, scientific insights that can inform public health policy are
more likely to be obtained by using genetic sequence data to estimate
transmission parameters more precisely (e.g., covariate effects on
infectiousness and susceptibility). In a survival analysis framework,
transmission parameter estimation is based on sums or averages over possible
transmission trees. By providing partial information about who-infected-whom, a
phylogeny can increase the efficiency of these estimates. The leaves of the
phylogeny represent sampled pathogens, which have known hosts. The interior
nodes represent common ancestors of sampled pathogens, which have unknown
hosts. We show that there is a one-to-one relationship between the possible
assignments of interior node hosts and the transmission trees simultaneously
consistent with the phylogeny and the epidemiologic data on person, place, and
time. We develop algorithms to find the set of possible hosts at each interior
node, to generate all possible transmission trees given these host sets, and to
assign branching times to a phylogeny with known interior node hosts. For any
possible transmission tree, there is at least one assignment of branching times
in the phylogeny that is consistent with the epidemiologic data. Finally, the
host set algorithm can be adapted to account for known branching times in the
phylogeny. A simulation study confirms that a phylogeny substantially increases
the efficiency of estimated hazard ratios for infectiousness and
susceptibility. We use these methods to analyze data from foot-and-mouth
disease virus outbreaks in the United Kingdom in 2001 and 2007.
[Show abstract][Hide abstract] ABSTRACT: Avian influenza A (H7N9), emerged in China in April 2013, sparking fears of a new, highly pathogenic, influenza pandemic. In addition, avian influenza A (H5N1) continues to circulate and remains a threat. Currently, influenza H7N9 vaccines are being tested to be stockpiled along with H5N1 vaccines. These vaccines require two doses, 21 days apart, for maximal protection. We developed a mathematical model to evaluate two possible strategies for allocating limited vaccine supplies: a one-dose strategy, where a larger number of people are vaccinated with a single dose, or a two-dose strategy, where half as many people are vaccinated with two doses. We prove that there is a threshold in the level of protection obtained after the first dose, below which vaccinating with two doses results in a lower illness attack rate than with the one-dose strategy; but above the threshold, the one-dose strategy would be better. For reactive vaccination, we show that the optimal use of vaccine depends on several parameters, with the most important one being the level of protection obtained after the first dose. We describe how these vaccine dosing strategies can be integrated into effective pandemic control plans.
[Show abstract][Hide abstract] ABSTRACT: Background The 2014 epidemic of Ebola virus disease in parts of west Africa defi nes an unprecedented health threat. We developed a model of Ebola virus transmission that integrates detailed geographical and demographic data from Liberia to overcome the limitations of non-spatial approaches in projecting the disease dynamics and assessing non-pharmaceutical control interventions.
[Show abstract][Hide abstract] ABSTRACT: In epidemics of infectious diseases such as influenza, an individual may have one of four possible final states: prior immune, escaped from infection, infected with symptoms, and infected asymptomatically. The exact state is often not observed. In addition, the unobserved transmission times of asymptomatic infections further complicate analysis. Under the assumption of missing at random, data-augmentation techniques can be used to integrate out such uncertainties. We adapt an importance-sampling-based Monte Carlo Expectation-Maximization (MCEM) algorithm to the setting of an infectious disease transmitted in close contact groups. Assuming the independence between close contact groups, we propose a hybrid EM-MCEM algorithm that applies the MCEM or the traditional EM algorithms to each close contact group depending on the dimension of missing data in that group, and discuss the variance estimation for this practice. In addition, we propose a bootstrap approach to assess the total Monte Carlo error and factor that error into the variance estimation. The proposed methods are evaluated using simulation studies. We use the hybrid EM-MCEM algorithm to analyze two influenza epidemics in the late 1970s to assess the effects of age and preseason antibody levels on the transmissibility and pathogenicity of the viruses.
[Show abstract][Hide abstract] ABSTRACT: In this talk, I summarize the process of estimating important transmission parameters and building statistical and mathematical models for the transmission and control of pandemic influenza A (H1N1) 2009. This activity began immediately after notice of spread of the virus in Mexico and California in April, 2009. The time and accuracy of the estimates of the transmission parameters, natural history parameters, and pathogenicity and severity indexes will be reviewed. We will describe how statistical and mathematical models were used to project the likely spread of the pandemic and the effectiveness and timing of control strategies. A description will be given of how mathematical models were used to assess the effectiveness of non-pharmaceutical and pharmaceutical interventions. We will pay particular attention to vaccination strategies, concentrating on when vaccine arrived with respect to the timing of the epidemics in various locals in the US. With the vaccine arriving late, we will describe what information analytic methods provided for the tradeoff between vaccinating high spreading or high risk people first. We will put the mathematical modeling and analysis in context with the current post-pandemic spread of seasonal influenza during the 2010 – 2011 influenza season in the US. Future plans will be discussed for integrating mathematical and statistical modeling of infectious disease spread and control with local and national level infectious disease control efforts.
American Association for the Advancement of Science 2011 Annual Meeting; 02/2011
[Show abstract][Hide abstract] ABSTRACT: Objectives: To project the potential economic impact of pandemic influenza mitigation strategies from a societal perspective in the United States.Methods: We use a stochastic agent-based model to simulate pandemic influenza in the community. We compare 17 strategies: targeted antiviral prophylaxis (TAP) alone and in combination with school closure as well as prevaccination.Results: In the absence of intervention, we predict a 50% attack rate with an economic impact of $187 per capita as loss to society. Full TAP (FTAP) is the most effective single strategy, reducing number of cases by 54% at the lowest cost to society ($127 per capita). Prevaccination reduces number of cases by 48% and is the second least costly alternative ($140 per capita). Adding school closure to FTAP or prevaccination further improves health outcomes but increases total cost to society by approximately $2700 per capita.Conclusion: FTAP is an effective and cost-saving measure for mitigating pandemic influenza.
Value in Health 02/2009; 12(2):226 - 233. DOI:10.1111/j.1524-4733.2008.00437.x · 2.89 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Prophylaxis of contacts of infectious cases such as household members and treatment of infectious cases are methods to prevent the spread of infectious diseases. We develop a method based on maximum likelihood to estimate the efficacy of such interventions and the transmission probabilities. We consider both the design with prospective follow-up of close contact groups and the design with ascertainment of close contact groups by an index case as well as randomization by groups and by individuals. We compare the designs by using simulations. We estimate the efficacy of the influenza antiviral agent oseltamivir in reducing susceptibility and infectiousness in two case-ascertained household trials.
Journal of the Royal Statistical Society Series C Applied Statistics 04/2006; 55(3):317 - 330. DOI:10.1111/j.1467-9876.2006.00539.x · 1.42 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: We derive the nonparametric maximum likelihood estimate (NPMLE) of the cumulative incidence functions for competing risks survival data subject to interval censoring and truncation. Since the cumulative incidence function NPMLEs give rise to an estimate of the survival distribution which can be undefined over a potentially larger set of regions than the NPMLE of the survival function obtained ignoring failure type, we consider an alternative pseudolikelihood estimator. The methods are then applied to data from a cohort of injecting drug users in Thailand susceptible to infection from HIV-1 subtypes B and E.
[Show abstract][Hide abstract] ABSTRACT: We estimate the transmission probability for the human immunodeficiency virus from seroconversion data of a cohort of injecting drug users (IDUs) in Thailand. The transmission probability model developed accounts for interval censoring and incorporates each IDU's reported frequency of needle sharing and injecting acts. Using maximum likelihood methods, the per needle sharing act transmission probability estimate between infectious and susceptible IDUs is 0.008. The effects of covariates, disease dynamics, mismeasured exposure information and the uncertainty of the disease prevalence on the transmission probability estimate are considered.
Journal of the Royal Statistical Society Series C Applied Statistics 02/2001; 50(1):1-14. DOI:10.1111/1467-9876.00216 · 1.42 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Multistate models have been increasingly used to model natural history of many diseases as well as to characterize the follow-up of patients under varied clinical protocols. This modeling allows describing disease evolution, estimating the transition rates, and evaluating the therapy effects on progression. In many cases, the staging is defined on the basis of a discretization of the values of continuous markers (CD4 cell count for HIV application) that are subject to great variability due mainly to short time-scale noise (intraindividual variability) and measurement errors. This led us to formulate a Bayesian hierarchical model where, at a first level, a disease process (Markov model on the true states, which are unobserved) is introduced and, at a second level, the measurement process making the link between the true states and the observed marker values is modeled. This hierarchical formulation allows joint estimation of the parameters of both processes. Estimation of the quantities of interest is performed via stochastic algorithms of the family of Markov chain Monte Carlo methods. The flexibility of this approach is illustrated by analyzing the CD4 data on HIV patients of the Concorde clinical trial.
[Show abstract][Hide abstract] ABSTRACT: The per-sexual-act probability of transmission of the human immunodeficiency virus type 1 (HIV) from an infected person to a susceptible sex partner can be determined from a simple model if the number of contacts each study participant has with infected partners is known. The unusual epidemiologic situation in the emerging HIV epidemic in Thailand allowed this quantity to be estimated from a cross-sectional study of young men conscripted into the Thai military in 1991. Although the simple model does not fit the data, an errors-in-variables approach provides a model with adequate fit. Other sources of lack of fit, including heterogeneity of the transmission probability, are discussed. With adjustment for measurement error, the per-act probability is estimated to be 0.056, an order of magnitude higher than similar estimates in North America. Because data indicate that recently infected persons may be more infectious, and because extensive HIV transmission in Thailand began in 1988, this difference may be due, in part, to a higher proportion of recently infected individuals in the emerging Thai epidemic from 1988 to 1991.
Statistics in Medicine 10/1994; 13(19‐20):2097 - 2106. DOI:10.1002/sim.4780131918 · 2.04 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Field studies of the efficacy of prophylactic vaccines in reducing susceptibility rely on the assumption of equal exposure to infection in the vaccinated and unvaccinated groups. Differential exposure to infection could, however, be the goal of other types of intervention programme, or it could occur secondary to belief in the protective effects of a prophylactic measure, such as vaccination. We call this differential exposure the exposure efficacy, or behaviour efficacy. To study the relative contribution of unequal exposure to infection and differential susceptibility to the estimate of vaccine efficacy, we formulate a simple model that explicitly includes both susceptibility and exposure to infection. We illustrate this on the example of randomized field trials of prophylactic human immunodeficiency virus vaccines. Increased exposure to infection in the vaccinated group may bias the estimated reduction in susceptibility. The bias in the estimate depends on the choice of efficacy parameter, the amount of information used in the analysis, the distribution and level of protection in the population, and the imbalance in exposure to infection. Sufficient increase in contacts in the vaccinated could result in the vaccine being interpreted as having an immunosuppressive effect. Estimates of vaccine efficacy are generally more robust to imbalances in exposure to infection when the detailed history of exposure to infection can be used in the analysis or at high levels of protection. The bias also depends on the relationship between the distribution of vaccine protection and the distribution of behaviour change, which could differ between blinded and unblinded trials.
Statistics in Medicine 02/1994; 13(4):357 - 377. DOI:10.1002/sim.4780130404 · 2.04 Impact Factor