Multistate modelling is a strategy for jointly modelling related time-to-event outcomes that can handle complicated outcome relationships, has appealing interpretations, can provide insight into different aspects of disease development and can be useful for making individualized predictions. A challenge with using multistate modelling in practice is the large number of parameters, and variable selection and shrinkage strategies are needed in order for these models to gain wider adoption. Application of existing selection and shrinkage strategies in the multistate modelling setting can be challenging due to complicated patterns of data missingness, inclusion of highly correlated predictors and hierarchical parameter relationships.
In this article, we discuss how to modify and implement several existing Bayesian variable selection and shrinkage methods in a general multistate modelling setting. We compare the performance of these methods in terms of parameter estimation and model selection in a multistate cure model of recurrence and death in patients treated for head and neck cancer. We can view this work as a case study of variable selection and shrinkage in a complicated modelling setting with missing data.
AndersenP and KeidingN (2002) Multi-state models for event history analysis. Statistical Methods in Medical Research, 11, 91–115.
2.
BartlettJW, SeamanSR, WhiteIRCarpenterJR (2014) Multiple imputation of covariates by fully conditional specification: Accomodating the substantive model. Statistical Methods in Medical Research, 24, 462–87.
3.
BeesleyLJ (2018) Missing Data and Variable Selection Methods for Cure Models in Cancer Research. PhD thesis, University of Michigan, Ann Arbor, Michigan.
4.
BeesleyLJ, WolfGTTaylorJMG (2016) Multiple imputation of missing covariates for the Cox proportional hazards cure model. Statistics in Medicine, 35, 4701–17.
5.
BeesleyLJ, MorganTM, SprattDE, SinghalU, FengFY, FurgalACJacksonWC (2019) Individual and population comparisons of surgery and radiotherapy outcomes in prostate cancer using Bayesian multistate models. JAMA Network Open, 2, 1–15.
6.
BeesleyLJTaylorJMG (2018) EM algorithms for fitting multistate cure models. Biostatistics, 20, 416–32.
7.
CarvalhoCMPolsonNGScottJG (2009) Handling sparsity via the horseshoe. In Proceedings of the 12th International Conference on Artificial Intelligence and Statistics, Barcelona, Spain, pages 73-80.
8.
CarvalhoCMPolsonNGScottJG (2010) The horseshoe estimator for sparse signals. Biometrika, 97, 465–80.
9.
ChipmanH, GeorgeEIMcCullochRE (2001) The practical implementation of Bayesian model selection (Technical report). Institute of Mathematical Statistics URL https://repository.upenn.edu/statistics_papers/184/.
10.
CognettiDM, WeberRSLaiSY (2008) Head and neck cancer: An evolving treatment paradigm. Cancer, 113, 1911–32.
11.
ConlonASC, TaylorJMGSargentDJ (2013) Multi-state models for colon cancer recurrence and death with a cured fraction. Statistics in Medicine, 33, 1750–66.
12.
DellaportasPForsterJJNtzoufrasI (2002) On Byesian model and variable selection using MCMC. Statistics and Computing, 12, 27–36.
13.
DuffySTaylorJMG, TerrellJ, IslamM, YuanZ, FowlerK, WolfGTeknosT (2008) IL-6 predicts recurrence among head and neck cancer patients. Cancer, 113, 750–57.
GrauJJ, CuchiA, TraserraJ, AriasC, BlanchJL and EstapeJ (1997) Follow-up sudy in head and neck cancer: Cure rate according to tumor location and stage. Oncology, 54, 38–42.
18.
GreenPJ (1995) Reversible jump Markov Chain Monte Carlo computation and Bayesian model determination. Biometrika, 82, 711–32.
19.
HastieDIGreenPJ (2012) Model choice using reversible jump Markov Chain Monte Carlo. Statistica Neerlandica, 66, 309–38.
HoetingJAMadiganDRafteryAEVolinskyCT (1999) Bayesian model averaging: A tutorial. Statistical Science, 14, 382–401.
22.
KoslovskyMDSwartzMDChanW, Leon-noveloL, WilkinsonAV, KendzorDE, BusinelleMsScienceD (2018) Bayesian variable selection for multistate Markov models with interval-censored data in an ecological momentary assessment study of smoking cessation. Biometrics, 74, 636–44.
23.
LamnisosDGrinJESteelMFJ (2013) Adaptive Monte Carlo for Bayesian variable selection in regression models. Journal of Computational and Graphical Statistics, 22, 729–48.
24.
LittleRJARubinDB (2002) Statistical Analysis with Missing Data, 2nd edition. Hoboken, NJ: John Wiley and Sons.
25.
MostS, PobneckerWTutzG (2016) Variable selection for discrete competing risks models. Qual Quant, 50, 1589–610.
26.
NewcombePJ, AliHR, BlowsFM, ProvenzanoE, PharoahPD, CaldasCRichardsonS (2014) Weibull regression with Bayesian variable selection to identify prognostic tumour markers of breast cancer survival. Statistical Methods in Medical Research, 26, 414–36.
27.
O'HaraRBSillanpaaMJ (2009) A review of Bayesian variable selection methods: What, how and which. Bayesian Analysis, 4, 85-118.
28.
OtavaMShkedyZ, LinD, GöhlmannHWH, BijnensL, TalloenWKasimA (2014) Dose: Response modeling under simple order restrictions using Bayesian variable selection methods. Statistics in Biopharmaceutical Research, 6, 252–62.
29.
PetersonLA, BellileEL, WolfGT, ViraniS, ShumanAG and TaylorJMG (2016) Cigarette use, comorbidities, and prognosis in a prospective head and neck squamous cell carcinoma population. Head and Neck, 38, 1810–20.
30.
TaylorJMG (1995) Semiparametric estimation in failure time mixture models. Biometrics, 51, 899–907.
31.
TibshiraniR (1996) Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society, 58, 267–88.
32.
TroughtonPTGodsillSJ (1997) A reversible jump sampler for autoregressive time series, employing full conditionals to achieve efficient model space moves (Technical report). University of Cambridge, Cambridge, UK.
33.
ZouHHastieT (2005) Regularization and variable selection via the elastic net. Journal of the Royal Statistical Society (Series B), 67, 301–20.
