Project

Back to overview

Multivariate analysis of dependent count data

Applicant Held Leonhard
Number 130002
Funding scheme Project funding (Div. I-III)
Research institution Institut für Epidemiologie, Biostatistik und Prävention Universität Zürich
Institution of higher education University of Zurich - ZH
Main discipline Mathematics
Start/End 01.08.2010 - 31.07.2011
Approved amount 97'433.00
Show all

All Disciplines (3)

Discipline
Mathematics
Methods of Epidemiology and Preventive Medicine
Medical Statistics

Keywords (7)

Bayesian statistics; Infectious disease epidemiology; Markov chain Monte Carlo; Spatial epidemiology; Statistical modelling; Structured additive regression; Gaussian Markov random fields

Lay Summary (English)

Lead
Lay summary
There is a trend in biomedical and public health research towards outcome measures of increasing complexity. This project aims to develop novel statistical methodology for multivariate epidemiological count data. Correlated Gaussian Markov random field models based on a Kronecker product precision matrix will be developed. This framework will provide great flexibility for structured additive regression and will be particularly suitable for longitudinal and space-time data. Algorithmic routines based on both Markov chain Monte Carlo (MCMC) and integrated nested Laplace approximations (INLA) will be developed and compared.
Direct link to Lay Summary Last update: 21.02.2013

Responsible applicant and co-applicants

Employees

Publications

Publication
Statistical modeling of infectious disease surveillance data
Held Leonhard, Paul Michaela (2013), Statistical modeling of infectious disease surveillance data, in M’ikanatha N. M. (ed.), Wiley-Blackwell, Hoboken, NJ, 535-544.
A conditional approach for inference in multivariate age-period-cohort models
Held Leonhard, Riebler Andrea (2012), A conditional approach for inference in multivariate age-period-cohort models, in Statistical Methods in Medical Research, 21(4), 311-329.
Estimation and extrapolation of time trends in registry data - borrowing strength from related populations
Riebler Andrea, Held Leonhard, Rue Håvard (2012), Estimation and extrapolation of time trends in registry data - borrowing strength from related populations, in Annals of Applied Statistics, 6(1), 304-333.
Gender-specific differences and the impact of family integration on time trends in age-stratified Swiss suicide rates
Riebler Andrea, Held Leonhard, Rue Håvard, Bopp Matthias (2012), Gender-specific differences and the impact of family integration on time trends in age-stratified Swiss suicide rates, in Journal of the Royal Statistical Society-Series A, 175(2), 473-490.
Modeling seasonality in space-time infectious disease surveillance data
Held Leonhard, Paul Michaela (2012), Modeling seasonality in space-time infectious disease surveillance data, in Biometrical Journal, 54(6), 824-843.
Heterogeneity in vaccination coverage explains the size and occurrence of measles epidemics in German surveillance data
Herzog Sereina, Paul Michaela, Held Leonhard (2011), Heterogeneity in vaccination coverage explains the size and occurrence of measles epidemics in German surveillance data, in Epidemiology and Infection, 139(4), 505-515.
Modelling seasonal patterns in longitudinal profiles with correlated random walks
Riebler Andrea, Held Leonhard, Rue Håvard (2011), Modelling seasonal patterns in longitudinal profiles with correlated random walks, Proceedings of the 26th International Workshop on Statistical Modelling, Valencia, Spain.
Predictive assessment of a non-linear random effects model for multivariate time series of infectious disease counts
Paul Michaela, Held Leonhard (2011), Predictive assessment of a non-linear random effects model for multivariate time series of infectious disease counts, in Statistics in Medicine, 30(10), 1118-1136.
The analysis of heterogeneous time trends in multivariate age-period-cohort models
Riebler Andrea, Held Leonhard (2010), The analysis of heterogeneous time trends in multivariate age-period-cohort models, in Biostatistics, 11(1), 57-69.

Collaboration

Group / person Country
Types of collaboration
Norwegian University of Science and Technology Norway (Europe)
- Publication

Scientific events

Active participation

Title Type of contribution Title of article or contribution Date Place Persons involved
26th International Workshop on Statistical Modelling (IWSM2011) Talk given at a conference Spatio-temporal disease mapping using INLA 11.07.2011 Valencia, Spain Held Leonhard;
INLA discussion meeting Talk given at a conference I don't remember 26.05.2011 Trondheim, Norway Riebler Andrea;
Stochastic modelling and statistical analysis: a Symposium in honour of Julian Besag Talk given at a conference Bayesian Age-Period-Cohort Analysis 02.04.2011 Bristol, Great Britain and Northern Ireland Held Leonhard;
Workshop on Bayesian Inference for Latent Gaussian Models with Applications Talk given at a conference Gaussian Kronecker product Markov random fields 03.02.2011 Zürich, Switzerland Riebler Andrea;
19th International Conference on Computational Statistics Talk given at a conference Spatio-temporal disease mapping using INLA 26.08.2010 Paris, France Held Leonhard;


Associated projects

Number Title Start Funding scheme
124429 Multivariate analysis of dependent count data 01.04.2009 Project funding (Div. I-III)
137919 Statistical methods for spatio-temporal modelling and prediction of infectious diseases 01.03.2012 Project funding (Div. I-III)
124429 Multivariate analysis of dependent count data 01.04.2009 Project funding (Div. I-III)

Abstract

There is a trend in biomedical and public health research towardsoutcome measures of increasing complexity. This project aims todevelop novel statistical methodology for epidemiological countdata. Firstly, correlated Gaussian Markov random field models will bedeveloped. This class of models will provide great flexibility forstructured additive regression and will be particularly suitable forlongitudinal and space-time data. Algorithmic routines based on bothMarkov chain Monte Carlo (MCMC) and integrated nested Laplaceapproximations (INLA) will be developed and compared. Secondly,statistical methodology for the analysis of space-time data oninfectious diseases will be developed. Such data need specificallytailored models to capture the typical epidemic behaviour. We willextend a recently proposed model framework to allow the infectiousnessparameters to depend on observed covariates and to allow forcorrelated random effects. Weights which represent the influence ofpast counts in neighboring regions will be treated as unknown to obtain data-driven information on the spatio-temporal spread ofinfectious diseases. This will be of particular public healthimportance to provide realistic descriptions and predictions of futureinfectious disease epidemics.
-