Relative survival represents the preferred framework for the analysis of population cancer survival data. The aim is to model the survival probability associated with cancer in the absence of information about the cause of death. Recent data linkage developments have allowed for incorporating the place of residence into the population cancer databases; however, modeling this spatial information has received little attention in the relative survival setting. We propose a flexible parametric class of spatial excess hazard models (along with inference tools), named “Relative Survival Spatial General Hazard,” that allows for the inclusion of fixed and spatial effects in both time-level and hazard-level components. We illustrate the performance of the proposed model using an extensive simulation study, and provide guidelines about the interplay of sample size, censoring, and model misspecification. We present a case study using real data from colon cancer patients in England. This case study illustrates how a spatial model can be used to identify geographical areas with low cancer survival, as well as how to summarize such a model through marginal survival quantities and spatial effects.
AllemaniCWeirHKCarreiraH, et al. Global surveillance of cancer survival 1995–2009: analysis of individual data for 25 676 887 patients from 279 population-based registries in 67 countries (CONCORD-2). Lancet2015; 385: 977–1010.
2.
EstèveJBenhamouECroasdaleM, et al.Relative survival and the estimation of net survival: elements for further discussion. Stat Med1990; 9: 529–538.
3.
PermeMPStareJEstèveJ. On estimation in relative survival. Biometrics2012; 68: 113–120.
4.
ElettiAMarraGQuaresmaM, et al. A unifying framework for flexible excess hazard modeling with applications in cancer epidemiology. J R Stat Soc: Ser C2022; 71: 1044–1062.
5.
ExarchakouARachetBBelotA, et al. Impact of national cancer policies on cancer survival trends and socioeconomic inequalities in England, 1996–2013: population based study. BMJ2018; 360: 1–10.
QuaresmaMCarpenterJRTurculetA, et al. Variation in colon cancer survival for patients living and receiving care in London, 2006–2013: does where you live matter?J Epidemiol Community Health2022; 76: 196–205.
8.
LiYRyanL. Modeling spatial survival data using semiparametric frailty models. Biometrics2002; 58: 287–297.
9.
ZhouHHansonT. A unified framework for fitting Bayesian semiparametric models to arbitrarily censored survival data, including spatially referenced data. J Am Stat Assoc2018; 113: 571–581.
10.
KleinJPVan-HouwelingenHCIbrahimJG, et al. Handbook of Survival Analysis. Boca Raton, FL: CRC Press, 2014.
11.
BanerjeeSWallMMCarlinBP. Frailty modeling for spatially correlated survival data, with application to infant mortality in Minnesota. Biostatistics2003b; 4: 123–142.
12.
CarlinBPBanerjeeS. Models for Spatio-Temporally Correlated Survival Data. In Bayesian Statistics 7: Proceedings of the Seventh Valencia International Meeting p. 45, Oxford University Press, 2003.
13.
LiLHansonTZhangJ. Spatial extended hazard model with application to prostate cancer survival. Biometrics2015; 71: 313–322.
14.
ChenYQJewellNP. On a general class of semiparametric hazards regression models. Biometrika2001; 88: 687–702.
15.
BasakPLineroASinhaD, et al. Semiparametric analysis of clustered interval-censored survival data using soft Bayesian additive regression trees (SBART). Biometrics2022; 78: 880–893.
16.
RubioFJDrikvandiR. MEGH: a parametric class of general hazard models for clustered survival data. Stat Methods Med Res2022; 31: 1603–1616.
17.
CharvatHRemontetLBossardN, et al. A multilevel excess hazard model to estimate net survival on hierarchical data allowing for non-linear and non-proportional effects of covariates. Stat Med2016; 35: 3066–3084.
18.
CrambSMMengersenKLLambertPC, et al. A flexible parametric approach to examining spatial variation in relative survival. Stat Med2016; 35: 5448–5463.
19.
CrambSMoragaPMengersenK, et al. Spatial variation in cancer incidence and survival over time across Queensland, Australia. Spat Spatiotemporal Epidemiol2017; 23: 59–67.
20.
YuBTiwariR. A Bayesian approach to mixture cure models with spatial frailties for population-based cancer relative survival data. Can J Stat2012; 40: 40–54.
21.
Etezadi-AmoliJCiampiA. Extended hazard regression for censored survival data with covariates: a spline approximation for the baseline hazard function. Biometrics1987; 43: 181–192.
22.
BuckleyJJamesI. Linear regression with censored data. Biometrika1979; 66: 429–436.
23.
ChenYQWangMC. Analysis of accelerated hazards models. J Am Stat Assoc2000; 95: 608–618.
24.
MoragaP. Geospatial health data: Modeling and visualization with R-INLA and Shiny. Boca Raton, FL: CRC Press, 2019.
25.
R Development Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing Vienna, Austria, 2022. ISBN 3-900051-07-0.
26.
CarpenterBGelmanAHoffmanMD, et al. Stan: a probabilistic programming language. J Stat Softw2017; 76: 1–32.
27.
Freni-SterrantinoAVentrucciMRueH. A note on intrinsic conditional autoregressive models for disconnected graphs. Spat Spatiotemporal Epidemiol2018; 26: 25–34.
28.
MorrisMWheeler-MartinKSimpsonD, et al. Bayesian hierarchical spatial models: implementing the Besag York Mollie model in Stan. Spat Spatiotemporal Epidemiol2019; 31: 100301.
29.
BesagJ. Spatial interaction and the statistical analysis of lattice systems. J R Stat Soc: Ser B (Methodological)1974; 36: 192–225.
30.
BesagJGreenPHigdonD, et al. Bayesian computation and stochastic systems. Stat Sci1995; 10: 3–41.
31.
BesagJYorkJMollieA. Bayesian image restoration, with two applications in spatial statistics. Ann Inst Stat Math1991; 43: 1–20.
32.
MahmoodMAmaralAVRMateuJ, et al. Modeling infectious disease dynamics: integrating contact tracing-based stochastic compartment and spatio-temporal risk models. Spat Stat2022; 51: 100691.
33.
BanerjeeSCarlinBPGelfandAE. Hierarchical modeling and analysis for spatial data. Boca Raton, FL: Chapman and Hall/CRC, 2003a.
34.
RieblerASørbyeSHSimpsonD, et al. An intuitive Bayesian spatial model for disease mapping that accounts for scaling. Stat Methods Med Res2016; 25: 1145–1165.
35.
SørbyeSHRueH. Penalised complexity priors for stationary autoregressive processes. J Time Ser Anal2017; 38: 923–935.
36.
FahrmeirLKneibTLangS, et al. Regression: models, methods and applications. Berlin: Springer, 2013.
37.
DigglePJRibeiroPJ. Model-based Geostatistics. New York: Springer, 2007.
38.
DigglePJMenezesRSuT. Geostatistical inference under preferential sampling. J R Stat Soc: Ser C (Applied Statistics)2010; 59: 191–232.
39.
AmaralAVRKrainskiETZhongR, et al. Model-based geostatistics under spatially varying preferential sampling. J Agric Biol Environ Stat2023; 1–27. https://link.springer.com/article/10.1007/s13253-023-00571-0
40.
RubioFRemontetLJewellN, et al. On a general structure for hazard-based regression models: an application to population-based cancer research. Stat Methods Med Res2019; 28: 2404–2417.
41.
STAN Development Team. RStan: the R interface to STAN. R package version 2.21.3, 2021.
42.
BetancourtMGirolamiM. Hamiltonian Monte Carlo for hierarchical models. Curr Trends Bayesian Methodol Appl2015; 79: 2–4.
43.
ZellnerA. On assessing prior distributions and Bayesian regression analysis with g-prior distributions, 1986.
44.
RueHHeldL. Gaussian Markov random fields: theory and applications. New York: CRC press, 2005.
45.
LunnDSpiegelhalterDThomasA, et al. The BUGS project: evolution, critique and future directions. Stat Med2009; 28: 3049–3067.
46.
GelmanA. Prior distributions for variance parameters in hierarchical models. Bayesian Anal2006; 1: 515–533.
47.
SimpsonRRueHRieblerA, et al. Penalising model component complexity: a principled, practical approach to constructing priors. Stat Sci2017; 32: 1–28.
48.
RubioFJSteelMFJ. Flexible linear mixed models with improper priors for longitudinal and survival data. Electron J Stat2018; 12: 572–598.
49.
AlvaresDRubioFJ. A tractable Bayesian joint model for longitudinal and survival data. Stat Med2021; 40: 4213–4229.
VehtariAGelmanAGabryJ. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Stat Comput2017; 27: 1413–1432.
52.
VehtariAGabryJMagnussonM, et al. loo: efficient leave-one-out cross-validation and WAIC for Bayesian models. R package version 2.5.1, 2022.
53.
VehtariASimpsonDGelmanA, et al. Pareto smoothed importance sampling, 2015.
54.
TaylorBM. Spatial Survival Analysis, pp. 2101–2109. Cham: Springer International Publishing, 2017.
55.
Office for National Statistics. CAL and NCV (Feb 2017) Full Clipped Boundaries in England, 2022. https://geoportal. statistics.gov.uk/datasets/ons::cal-and-ncv-feb-2017-full-clipped-boundaries-in-england/.
RachetBMaringeCWoodsLM, et al. Multivariable flexible modelling for estimating complete, smoothed life tables for sub-national populations. BMC Public Health2015; 15: 1–9.
58.
DattaABanerjeeSHodgesJS, et al. Spatial disease mapping using directed acyclic graph auto-regressive (DAGAR) models. Bayesian Anal2019; 14: 1221.
59.
HendersonRShimakuraSGorstD. Modeling spatial variation in leukemia survival data. J Am Stat Assoc2002; 97: 965–972.
60.
RubioFJ. SimLT. R package version 0.1.0. 2022.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
