A fully Bayesian approach to inference for Coxian phase-type distributions with covariate dependent mean

Phase-type distributions represent the time to absorption for a finite state Markov chain in continuous time, generalising the exponential distribution and providing a flexible and useful modelling tool. We present a new reversible jump Markov chain Monte Carlo scheme for performing a fully Bayesian analysis of the popular Coxian subclass of phase-type models; the convenient Coxian representation involves fewer parameters than a more general phase-type model. The key novelty of our approach is that we model covariate dependence in the mean whilst using the Coxian phase-type model as a very general residual distribution. Such incorporation of covariates into the model has not previously been attempted in the Bayesian literature. A further novelty is that we also propose a reversible jump scheme for investigating structural changes to the model brought about by the introduction of Erlang phases. Our approach addresses more questions of inference than previous Bayesian treatments of this model and is automatic in nature. We analyse an example dataset comprising lengths of hospital stays of a sample of patients collected from two Australian hospitals to produce a model for a patient's expected length of stay which incorporates the effects of several covariates. This leads to interesting conclusions about what contributes to length of hospital stay with implications for hospital planning. We compare our results with an alternative classical analysis of these data.

[1]  Peter H. Millard,et al.  A continuous time Markov model for the length of stay of elderly people in institutional long‐term care , 2005 .

[2]  P. Green,et al.  On Bayesian Analysis of Mixtures with an Unknown Number of Components (with discussion) , 1997 .

[3]  Anders Krogh,et al.  Automatic generation of gene finders for eukaryotic species , 2006, BMC Bioinformatics.

[4]  Søren Asmussen,et al.  Ruin probabilities , 2001, Advanced series on statistical science and applied probability.

[5]  Ajay Jasra,et al.  Population-Based Reversible Jump Markov Chain Monte Carlo , 2007, 0711.0186.

[6]  P Gastmeier,et al.  The preventable proportion of nosocomial infections: an overview of published reports. , 2003, The Journal of hospital infection.

[7]  Dimitris Bertsimas.  An Analytic Approach to a General Class of G/G/s Queueing Systems , 1990, Oper. Res..

[8]  Hwee Pink Tan,et al.  Admission control for differentiated services in future generation CDMA networks , 2009, Perform. Evaluation.

[9]  Nicholas Graves,et al.  Modeling length of stay in hospital and other right skewed data: comparison of phase-type, gamma and log-normal distributions. , 2009, Value in health : the journal of the International Society for Pharmacoeconomics and Outcomes Research.

[10]  S. Sisson,et al.  Reversible jump Markov chain Monte Carlo , 2010, 1001.2055.

[11]  J. Griffiths.  The Theory of Stochastic Processes , 1967 .

[12]  H. Skaug.  Markov Modulated Poisson Processes for Clustered Line Transect Data , 2006, Environmental and Ecological Statistics.

[13]  Qi-Ming He,et al.  Spectral Polynomial Algorithms for Computing Bi-Diagonal Representations for Phase Type Distributions and Matrix-Exponential Distributions , 2006 .

[14]  Steffen L. Lauritzen,et al.  The estimation of phase-type related functionals using Markov chain Monte Carlo methods , 2003 .

[15]  文魚 長谷川.  D.R. Cox and H.D. Miller: The Theory of Stochastic Processes, Methuen. London, 1965, 398頁, 24×16cm, 4,200円. , 1966 .

[16]  Nicholas Graves,et al.  Effect of Healthcare-Acquired Infection on Length of Hospital Stay and Cost , 2007, Infection Control & Hospital Epidemiology.

[17]  Sandjai Bhulai,et al.  Dynamic thread assignment in web server performance optimization , 2009, Perform. Evaluation.

[18]  P. Green,et al.  Corrigendum: On Bayesian analysis of mixtures with an unknown number of components , 1997 .

[19]  B. Carlin,et al.  Bayesian Model Choice Via Markov Chain Monte Carlo Methods , 1995 .

[20]  Michael P. Wiper,et al.  Bayesian analysis of M/Er/1 and M/H_k/1 queues , 1998, Queueing Syst. Theory Appl..

[21]  C. Robert,et al.  Bayesian inference in hidden Markov models through the reversible jump Markov chain Monte Carlo method , 2000 .

[22]  Elia El-Darzi,et al.  Length of Stay-Based Patient Flow Models: Recent Developments and Future Directions , 2005, Health care management science.

[23]  Sally McClean,et al.  Stochastic models of geriatric patient bed occupancy behaviour , 2000 .

[24]  Sally McClean,et al.  Conditional phase-type distributions for modelling patient length of stay in hospital , 2003 .

[25]  P. Green.  Reversible jump Markov chain Monte Carlo computation and Bayesian model determination , 1995 .

[26]  Ren Asmussen,et al.  Fitting Phase-type Distributions via the EM Algorithm , 1996 .

[27]  Sally McClean,et al.  A queueing model for bed-occupancy management and planning of hospitals , 2002, J. Oper. Res. Soc..

[28]  F G Ball,et al.  Stochastic models for systems of interacting ion channels. , 2000, IMA journal of mathematics applied in medicine and biology.

[29]  Marcel F. Neuts,et al.  Matrix-Geometric Solutions in Stochastic Models , 1981 .

[30]  María Concepción Ausín,et al.  Bayesian prediction of the transient behaviour and busy period in short- and long-tailed GI/G/1 queueing systems , 2008, Comput. Stat. Data Anal..

[31]  Mogens Bladt,et al.  Phase-type distributions and risk processes with state-dependent premiums , 1996 .

[32]  Sally McClean,et al.  Analysing data on lengths of stay of hospital patients using phase‐type distributions , 1999 .

[33]  G. Roberts,et al.  Efficient construction of reversible jump Markov chain Monte Carlo proposal distributions , 2003 .

[34]  Nicholas Graves,et al.  Effect of Pressure Ulcers on Length of Hospital Stay , 2005, Infection Control & Hospital Epidemiology.

[35]  S I McClean,et al.  Markov Chain Modelling for Geriatric Patient Care , 2005, Methods of Information in Medicine.