Daniela De Angelis - Academia.edu (original) (raw)

Papers by Daniela De Angelis

Journal of the American Statistical Association, 2014

ABSTRACT In this article, we develop a Bayesian framework for parameter estimation of a computati... more ABSTRACT In this article, we develop a Bayesian framework for parameter estimation of a computationally expensive dynamic epidemic model using time series epidemic data. Specifically, we work with a model for A/H1N1 influenza, which is implemented as a deterministic computer simulator, taking as input the underlying epidemic parameters and calculating the corresponding time series of reported infections. To obtain Bayesian inference for the epidemic parameters, the simulator is embedded in the likelihood for the reported epidemic data. However, the simulator is computationally slow, making it impractical to use in Bayesian estimation where a large number of simulator runs is required. We propose an efficient approximation to the simulator using an emulator, a statistical model that combines a Gaussian process (GP) prior for the output function of the simulator with a dynamic linear model (DLM) for its evolution through time. This modeling framework is both flexible and tractable, resulting in efficient posterior inference through Markov chain Monte Carlo (MCMC). The proposed dynamic emulator is then used in a calibration procedure to obtain posterior inference for the parameters of the influenza epidemic.

AIDS, 2014

HIV is a major public health problem among MSM in the United Kingdom with around 2400 new infecti... more HIV is a major public health problem among MSM in the United Kingdom with around 2400 new infections annually. We quantified the contribution of biological and behavioural factors. Modelling study. A partnership-based model of HIV transmission among UK MSM aged 15-64 years was developed and calibrated to time series HIV prevalence. The calibration was validated using multiple surveillance datasets. Population-attributable fractions were used to estimate the contribution of behavioural and biological factors to HIV transmission over the period 2001-2002, 2014-2015, and 2019-2020. The contribution of most biological and behavioural factors was relatively constant over time, with the key group sustaining HIV transmission being higher-sexual activity MSM aged below 35 years living with undiagnosed HIV. The effect of primary HIV infection was relatively small with 2014-2015 population-attributable fraction of 10% (3-28%) in comparison with other subsequent asymptomatic stages. Diagnosed men who were not on antiretroviral therapy (ART) currently contributed 26% (14-39%) of net infections, whereas ART-treated MSM accounted for 17% (10-24%). A considerable number of new infections are also likely to occur within long-term relationships. The majority of the new HIV infections among MSM in the United Kingdom during 2001-2020 is expected to be accounted for by a small group of younger and highly sexually active individuals, living with undiagnosed HIV in the asymptomatic stage. Bringing this group into HIV/AIDS care by improving testing uptake is a vital step for preventing onward transmission and will determine the success of using ART as prevention.

Epidemics, 2014

Bayesian Inference Multiple sources a b s t r a c t Public health-related decision-making on poli... more Bayesian Inference Multiple sources a b s t r a c t Public health-related decision-making on policies aimed at controlling epidemics is increasingly evidence-based, exploiting multiple sources of data. Policy makers rely on complex models that are required to be robust, realistically approximating epidemics and consistent with all relevant data. Meeting these requirements in a statistically rigorous and defendable manner poses a number of challenging problems. How to weight evidence from different datasets and handle dependence between them, efficiently estimate and critically assess complex models are key challenges that we expound in this paper, using examples from influenza modelling.

Statistical Science, 2013

Statistical Communications in Infectious Diseases, 2012

The Lancet, 2006

Harm-reduction interventions in injection drug use. By - Matthew Hickman, John Macleod, Daniela D... more Harm-reduction interventions in injection drug use. By - Matthew Hickman, John Macleod, Daniela De Angelis.

Journal of Statistical Computation and Simulation, 1992

Estimation of the sampling standard deviation of the variance stabilised correlation coefficient ... more Estimation of the sampling standard deviation of the variance stabilised correlation coefficient by the smoothed bootstrap is reconsidered. Four different strategies for empirical choice of the smoothing parameter are discussed and their performances examined ...

Applied Economics, 1997

This paper presents an overview of the application of the bootstrap to unit root testing. We show... more This paper presents an overview of the application of the bootstrap to unit root testing. We show how a bootstrap unit root test can be set up and discuss several options that have been proposed in the literature. The effects of these options on the performance of the test are analysed.

Statistics in Medicine, 1994

Journal of Hepatology, 2014

Hepatitis C (HCV) related disease in England is predicted to rise, and it is unclear whether trea... more Hepatitis C (HCV) related disease in England is predicted to rise, and it is unclear whether treatment at current levels will be able to avert this. The aim of this study was to estimate the number of people with chronic HCV infection in England that are treated and assess the impact and costs of increasing treatment uptake. Numbers treated were estimated using national data sources for pegylated interferon supplied, dispensed, or purchased from 2006 to 2011. A back-calculation approach was used to project disease burden over the next 30 years and determine outcomes under various scenarios of treatment uptake. 5000 patients were estimated to have been treated in 2011 and 28,000 in total from 2006 to 2011; approximately 3.1% and 17% respectively of estimated chronic infections. Without treatment, incident cases of decompensated cirrhosis and hepatocellular carcinoma were predicted to increase until 2035 and reach 2290 cases per year. Treatment at current levels should reduce incidence by 600 cases per year, with a peak around 2030. Large increases in treatment are needed to halt the rise; and with more effective treatment the best case scenario predicts incidence of around 500 cases in 2030, although treatment uptake must still be increased considerably to achieve this. If the infected population is left untreated, the number of patients with severe HCV-related disease will continue to increase and represent a substantial future burden on healthcare resources. This can be mitigated by increasing treatment uptake, which will have the greatest impact if implemented quickly.

Lancet, 2006

The Annals of Applied Statistics, 2014

Background: Accurate measures of the severity of pandemic (H1N1) 2009 influenza (pH1N1) are neede... more Background: Accurate measures of the severity of pandemic (H1N1) 2009 influenza (pH1N1) are needed to assess the likely impact of an anticipated resurgence in the autumn in the Northern Hemisphere. Severity has been difficult to measure because jurisdictions with large numbers of deaths and other severe outcomes have had too many cases to assess the total number with confidence. Also, detection of severe cases may be more likely, resulting in overestimation of the severity of an average case. We sought to estimate the probabilities that symptomatic infection would lead to hospitalization, ICU admission, and death by combining data from multiple sources.

http://isrctn.org/>, 2000

Changes in hepatitis C virus (HCV) prevalence from 1992 to 2006 were examined by using 24,311 rec... more Changes in hepatitis C virus (HCV) prevalence from 1992 to 2006 were examined by using 24,311 records from unlinked anonymous surveillance of injecting drug users in England and Wales. Bayesian logistic regression was used to estimate annual prevalence, accounting for changing recruitment patterns (age, gender, injecting duration, geographic region, interactions) and the sensitivity and specificity of different oral fluid testing devices. After controlling for these differences, the authors found that the adjusted HCV prevalence decreased from 70% (95% credible interval: 62, 78) in 1992 to 47% (95% credible interval: 43, 51) in 1998 before rising again to 53% (95% credible interval: 48, 58) in 2006. Women injecting drug users had a higher HCV risk than did men (odds ratio ¼ 1.50, 95% credible interval: 1.31, 1.73). Two regions (London and North West) had a markedly higher HCV prevalence than did the rest of England and Wales. Among individuals who had injected for less than 1 year, the adjusted HCV prevalence in 2006 was higher than that in 1992 (28% vs. 19%, respectively). HCV infection can be prevented. The public health challenge in England and Wales is to increase action in order to regain a downward trend in HCV risk and the benefit that has been lost since 1998.

PLoS ONE, 2014

Background: One pathway through which pandemic influenza strains might emerge is reassortment fro... more Background: One pathway through which pandemic influenza strains might emerge is reassortment from coinfection of different influenza A viruses. Seasonal influenza vaccines are designed to target the circulating strains, which intuitively decreases the prevalence of coinfection and the chance of pandemic emergence due to reassortment. However, individual-based analyses on 2009 pandemic influenza show that the previous seasonal vaccination may increase the risk of pandemic A(H1N1) pdm09 infection. In view of pandemic influenza preparedness, it is essential to understand the overall effect of seasonal vaccination on pandemic emergence via reassortment. Methods and Findings: In a previous study we applied a population dynamics approach to investigate the effect of infection-induced cross-immunity on reducing such a pandemic risk. Here the model was extended by incorporating vaccination for seasonal influenza to assess its potential role on the pandemic emergence via reassortment and its effect in protecting humans if a pandemic does emerge. The vaccination is assumed to protect against the target strains but only partially against other strains. We find that a universal seasonal vaccine that provides full-spectrum cross-immunity substantially reduces the opportunity of pandemic emergence. However, our results show that such effectiveness depends on the strength of infection-induced cross-immunity against any novel reassortant strain. If it is weak, the vaccine that induces cross-immunity strongly against non-target resident strains OPEN ACCESS but weakly against novel reassortant strains, can further depress the pandemic emergence; if it is very strong, the same kind of vaccine increases the probability of pandemic emergence. Conclusions: Two types of vaccines are available: inactivated and live attenuated, only live attenuated vaccines can induce heterosubtypic immunity. Current vaccines are effective in controlling circulating strains; they cannot always help restrain pandemic emergence because of the uncertainty of the oncoming reassortant strains, however. This urges the development of universal vaccines for prevention of pandemic influenza. Seasonal Vaccination and Pandemic Influenza Emergence PLOS ONE | Secondary infection in people immune to strain 2 with strain 1 J 1 rJ 1 +1, R 2 rR 2 -1 (1-Q)b 1 R 2 (I 1 +J 1 +w 1 I d )/N Secondary infection in people immune to strain 1 with strain 2 J 2 rJ 2 +1, R 1 rR 1 -1 (1-Q)b 2 R 1 (I 2 +J 2 +J V2 +w 2 I d )/N infection in people vaccinated with strain 2 J V2 rJ V2 +1, VrV-1 (1-Q V )b 2 V(I 2 +J 2 +J V2 +w 2 I d )/N

PLOS Medicine, 2009

British Medical Journal, 2010

Objectives To examine survival and long term cessation of injecting in a cohort of drug users and... more Objectives To examine survival and long term cessation of injecting in a cohort of drug users and to assess the influence of opiate substitution treatment on these outcomes. Design Prospective open cohort study. Setting A single primary care facility in Edinburgh. Participants 794 patients with a history of injecting drug use presenting between 1980 and 2007; 655 (82%) were followed up by interview or linkage to primary care records and mortality register, or both, and contributed 10 390 person years at risk; 557 (85%) had received opiate substitution treatment. Main outcome measures Duration of injecting: years from first injection to long term cessation, defined as last injection before period of five years of non-injecting; mortality before cessation; overall survival. Results In the entire cohort 277 participants achieved long term cessation of injecting, and 228 died. Half of the survivors had poor health related quality of life. Median duration from first injection to death was 24 years for participants with HIV and 41 years for those without HIV. For each additional year of opiate substitution treatment the hazard of death before long term cessation fell 13% (95% confidence interval 17% to 9%) after adjustment for HIV, sex, calendar period, age at first injection, and history of prison and overdose. Conversely exposure to opiate substitution treatment was inversely related to the chances of achieving long term cessation. Conclusions Opiate substitution treatment in injecting drug users in primary care reduces this risk of mortality, with survival benefits increasing with cumulative exposure to treatment. Treatment does not reduce the overall duration of injecting.

Statistics in Medicine, 1997

Back-calculation is a widely used method to estimate HIV incidence rates, and is commonly based o... more Back-calculation is a widely used method to estimate HIV incidence rates, and is commonly based on times of AIDS diagnosis. Following up earlier work, we extend this method to also incorporate knowledge of times of HIV diagnosis (first positive test). This is achieved through the use of a Markov model which describes the progress of an HIV infected person through various stages, and which allows causal connections between events to be explicitly modelled. Estimation is based on maximum likelihood, the likelihood being calculated within a discretized version of the Markov model. The effect of sampling uncertainty and model uncertainty (sensitivity) is evaluated simultaneously by means of a combined bootstrap and simulation procedure. At each replication we resample both the data and the model (from a set of possible models described by randomizing one or more parameters). For instance, uncertain knowledge about the incubation distribution affects the estimates of some parameters, but not others. The Markov approach is applied to the prediction of AIDS incidence for homosexuals in England and Wales up to the year 2000.