On the specification of prior distributions for variance components in disease mapping models
Keywords:Hierarchical models, Spatial epidemiology, Generalised inverse Gaussian distribution, Intrinsic conditional autoregressive models
AbstractIn this paper, we consider the problem of specifying priors for the variance components in the Bayesian analysis of the Besag-York-Mollié model, a model that is popular among epidemiologists for disease mapping. The model encompasses two sets of random effects: one spatially structured to model spatial autocorrelation and the other spatially unstructured to describe residual heterogeneity. In this model, prior specification for variance components is an important problem because these priors maintain their influence on the posterior distributions of relative risks when mapping rare diseases. We propose using generalised inverse Gaussian priors, a broad class of distributions that includes many distributions commonly used as priors in this context, such as inverse gammas. We discuss the prior parameter choice with the aim of balancing the prior weight of the two sets of random effects on total variation and controlling the amount of shrinkage. The suggested prior specification strategy is compared to popular alternatives using a simulation exercise and applications to real data sets.
O. BARNDORFF-NIELSEN (1977). Exponentially decreasing distributions for the logarithm of particle size. Proceedings of the Royal Statistical Society, series A, 353, pp. 401–419.
L. BERNARDINELLI, D. CLAYTON, C. MONTOMOLI, M. G. M, M. SONGINI (1995). Bayesian estimates of disease maps: how important are priors. Statistics in Medicine, 14, pp. 2411–2431.
J. BESAG (1968). Spatial interaction and the statistical analysis of lattice systems (with discussion). Journal of the Royal Statistical Society, series B, 36, pp. 192–236.
J. BESAG, C. KOOPERBERG (1995). On conditional and instrinsic autoregressions. Biometrika, 82, pp. 733–746.
J. BESAG, J. YORK, S. MOLLIÉ (1991). A bayesian image restoration, with two applications in spatial statistics (with discussion). Annals of the Institute of Statistical Mathematics, 43, pp. 1–59.
N. BEST, L. WALLER, A. THOMAS, E. CONLON, R. ARNOLD (1999). Bayesian models for spatially correlated disease and exposure data. In J. Berardo, J. Berger, A. Dawid, A. Smith (eds.), Sixth Valencia International Meeting on Bayesian Statistics, Oxford University Press, London, pp. 131–156.
B. BIBBY, M. SØRENSEN (2003). Generalized hyperbolic and inverse Gaussian distributions: limiting cases and approximation of processes. In S. Rachev (ed.), Handbook of Heavy Tailed Distributions in Finance, Elsevier, New York, pp. 211–248.
B. BIBBY, M. SØRENSEN (2004). Hyperbolic processes in finance. In R. Dalang, M. Dozzi, F. Russo (eds.), Seminar on Stochastic Analysis, Random Fields and Applications IV, Progress in Probability, Birkh¨auser Verlag, Berlin, pp. 221–264.
L. EBERLY, B. CARLIN (2000). Identifiability problems and convergence issues for markov chain monte carlo fitting of spatial models. Statistics in Medicine, 19, pp. 2279–2294.
E. FABRIZI, C. TRIVISANO (2012). Bayesian estimation of log-normal means with finite quadratic expected loss. Bayesian Analysis, 4, pp. 975 – 996.
F. GRECO, C.TRIVISANO (2009). A multivariate car model for improving the estimation of relative risks. Statistics in Medicine, 28, pp. 1707–1724.
A. LAWSON (2001). Statistical methods in spatial epidemiology. Wiley, Chichester.
A. LAWSON (2009). Bayesian disease mapping: hierarchical modelling in spatial epidemiology. Chapman and Hall, New York.
A. MOLLIÉ (1999). Bayesian and empirical bayes approaches to disease mapping. In A. Lawson, A. Biggeri, D. Bohning, E. Lesaffre, J. Viel, R. Bertollini (eds.), Disease mapping and risk assessment for public health, Wiley, New York, pp. 15–29.
J. PALACIOS (2001). Closed-form formulas for kirchhoff index. International Journal of Quantitative Chemistry, 81, pp. 135–140.
A. THOMAS, B. O’HARA, U. LIGGES, S. STURZ (2006). Making bugs open. R News, 6, pp. 12–17.
D. L. W. BREYMANN (2010). R: A Language and Environment for Statistical Computing. URL cran.r-project.org/web/packages/ghyp/index.html.
J. WAKEFIELD (2007). Disease mapping and spatial regression with count data. Biostatistics, 8, pp. 158–183.
J. WAKEFIELD, S. MORRIS (2001). The bayesian modeling of disease risk in relation to a point source. Journal of the American Statistical Association, 96, pp. 77–91.