Decisions under uncertainty: a computational framework for quantification of policies addressing infectious disease epidemics

Dengue Fever (DF) has been identified by the World Health organization (WHO) as one of the most serious vector-borne infectious diseases in tropical and subtropical areas. During 2007, in particular, there were over 2,000 DF cases in Taiwan, which was the highest number of cases in the recorded history of Taiwan epidemics. Most DF studies have focused mainly on temporal DF patterns and its close association with climatic covariates, whereas they have understated spatial DF patterns (spatial dependence and clustering) and composite space-time effects. The present study proposes a spatio-temporal DF prediction approach based on stochastic Bayesian Maximum Entropy (BME) analysis. Core and site-specific knowledge bases are considered, including climate and health datasets under conditions of uncertainty, space-time dependence functions, and a Poisson regression model of climatic variables contributing to DF occurrences in southern Taiwan during 2007. The results show that the DF outbreaks in the study area are highly influenced by climatic conditions. Furthermore, the analysis can provide the required ''oneweek-ahead'' outbreak warnings based on spatio-temporal predictions of DF distributions. Therefore, the proposed approach can provide the Taiwan Disease Control Agency with a valuable tool to timely identify, control, and even efficiently prevent DF spreading across space-time.

Section: Introductionmentioning

confidence: 99%

A spatio-temporal climate-based model of early dengue fever warning in southern Taiwan

Yang

Yen

et al. 2010

“…Spread of infectious diseases to new geographical areas during the last decades induced several attempts using various approaches to model outbreak dynamics of infectious diseases (Mikler et al 2007;Aldstadt 2007;Choi et al 2008). The present study focused on the question how to transform a deterministic dynamic model into a stochastic one.…”

Section: Discussionmentioning

confidence: 99%

Bayesian analysis of a dynamical model for the spread of the Usutu virus

Reiczigel

Brugger

Rubel

et al. 2009

The Usutu virus is an arbovirus transmitted by mosquitoes and causing disease in birds. The virus was detected in Austria for the first time in 2001, while a major outbreak occurred in 2003. Rubel et al. (2008) developed a nine-compartment deterministic SEIR model to explain the spread of the disease. We extended this to a hierarchical Bayes model assuming random variation in temperature data, in reproduction data of birds, and in the number of birds found to be infected. The model was implemented in R, combined with the FORTRAN subroutine for the original deterministic model. Analysis was made by MCMC using a random walk Metropolis scheme. Posterior means, medians, and credible intervals were calculated for the parameters. The hierarchical Bayes approach proved to be fruitful in extending the deterministic model into a stochastic one. It allowed for Bayesian point and interval estimation and quantification of uncertainty of predictions. The analysis revealed that some model parameters were not identifiable; therefore we kept constant some of them and analyzed others conditional on them. Identifiability problems are common in models aiming to mirror the mechanism of the process, since parameters with natural interpretation are likely to exhibit interrelationships. This study illustrated that Bayesian modeling combined with conditional analysis may help in those cases. Its application to the Usutu model improved model fit and revealed the structure of interdependencies between model parameters: it demonstrated that determining some of them experimentally would enable estimation of the others, except one of them, from available data.

“…This prompted us to consider the application of stochastic processes to model the temporal nature of HBV infection as multistate outcomes. From a methodological viewpoint, in spite of several previous attempts by using stochastic processes to model the dynamics of infectious diseases (Mikler et al 2007;Reiczigel et al 2010), this is the first study, to the best of our knowledge, to develop a five-state stochastic model to estimate the parameters governing the temporal natural history of HBV by taking both vertical and horizontal transmissions based on the two different areas of empirical data prior to the introduction of HBV vaccine in Taiwan. Compared with the two-state survival model, the advantages of using the five-state stochastic model are that it appears to be more flexible to deal with the heterogeneity of HBV infection as it not only elucidated the instantaneous rate of each transition from susceptible through latent period, active viremia, and until the carrier state or the recovery state but also estimated the proportion of vertical and horizontal transmission.…”

Section: Discussionmentioning

confidence: 99%

Stochastic model for hepatitis B virus infection through maternal (vertical) and environmental (horizontal) transmission with applications to basic reproductive number estimation and economic appraisal of preventive strategies

Hung

Wang

Yen

et al. 2013

Elucidating the temporal natural history of Hepatitis B virus (HBV) infection provides not only useful information for assessing the spread of HBV infection but also a bedrock for economic appraisal of population-based preventive strategies (such as universal vaccination). HBV transmissions are transmitted through the maternal (vertical) route and environmental (horizontal) route. As the infection process from susceptible to active virus replication and finally to recovery state or to carrier state through both routes is not directly observable, mathematical models were therefore proposed to quantitatively elucidate the dynamic process of HBV infection with time. We began with a simple two-state model (from susceptible to infection) and further developed a five-state (including susceptible, latent phase, active viral replication, carrier, and recovery states) stochastic process to quantify the dynamics of HBV infection, making allowances for a mixed proportion of vertical and horizontal transmissions. The data used in the estimation, which were collected before implementation of a policy for universal HBV vaccination, were derived from several previous serum prevalence studies, including both low (northern) and high (southern) HBV prevalence areas, in Taiwan. The parameters obtained from stochastic models were converted to compute the basic reproductive number (R 0 ) in each study to indicate the extent of the spread of HBV infection. By linking the temporal natural history of HBV infection with the previously established Markov process for the sequelae of HBV infection (chronic hepatitis infection, liver cirrhosis, and hepatocelluar carcinoma), we developed a Markov cycle decision tree to evaluate the two populationbased preventive strategies related to vaccination and maternal lamivudine use with a cost-effectiveness analysis. In the five-state model, horizontal transmission was found to contribute more to HBV infection than vertical transmission in the northern area (59 vs. 41 %), whereas vertical transmission was more likely than horizontal transmission in the southern area (80 vs. 20 %). After considering the parameters of the two transmission routes, annual infection rate (person-years) for susceptible subjects was higher in the northern area (0.35) than the southern area (0.011). A similar finding was noted for annual rate of the transition from the latent period to active viral replications (1.12 in the northern area and 0.072 in the southern area). Annual rate of the conversion from active viral replication to the carrier state was higher in the southern area (0.184) than the northern area (0.119). Annual recovery rates were 0.014 in the northern area and 0.024 in the southern area for the carrier and 0.048 and 0.088 for transient viremia. The R 0 at age 30 using the parameters obtained from the five-state stochastic model was estimated as 4.88 in the northern and 7.03 in the southern area. Regarding the findings of economic appraisal, both preventive strategies were cost-saving, yielding the n...