Calibrating the effect of covariates on multi-state outcome ridden with measurement errors using Bayesian Hidden Markov regression model


  • Author: Amy Ming-Fang Yen and Hsiu-Hsi Chen
  • Date: 12 April 2019
  • Copyright: Image copyright of Patrick Rhodes

Multistate Markov regression models used for quantifying the effect size of state‐specific covariates pertaining to the dynamics of multistate outcomes have gained popularity. However, the measurements of multistate outcome are prone to the errors of classification, particularly when a population‐based survey/research is involved with proxy measurements of outcome due to cost consideration. Such a misclassification may affect the effect size of relevant covariates such as odds ratio used in the field of epidemiology. The authors of an article published in Statistics in Medicine propose a Bayesian measurement‐error‐driven hidden Markov regression model for calibrating these biased estimates with and without a 2‐stage validation design.

The paper is available via this link with Open Access and the authors explain their findings in further detail below:

Bayesian measurement‐error‐driven hidden Markov regression model for calibrating the effect of covariates on multistate outcomes: Application to androgenetic alopecia

Statistics in Medicine, Volume 37, Issue 21, 20 September 2018, pages 3125-3146

thumbnail image: Calibrating the effect of covariates on multi-state outcome ridden with measurement errors using Bayesian Hidden Markov regression model

While measurement errors have been frequently seen in epidemiology most of statistical methods have been focused on the influence of misclassification on the effect size of covariates with binary (two-state) irreversible outcome, such as healthy/ill, alive/death, and so on. Very few statistical methods have been specified to demonstrate how the misclassification of multistate outcomes (rather than only two-state outcome) affects the effect size of state-specific covariates associated with the corresponding transitions. Yen and Chen (2018) show how the measurement errors of multi-state outcome is amenable to the recursive relationships with emission probabilities for misclassifications and transition probabilities for state transitions under the context of Hidden Markov regression model. They then developed a series of simulation algorithms for assessing different scenarios of misclassification and also applied Bayesian Markov Chain Monte Carlo estimation in conjunction with two-stage sampling design to estimate all the parameters of interest. The proposed method has been applied to the community-based survey of androgenetic alopecia and found that the effect size of the majority of covariate was inflated after calibration regardless of which type of misclassification. While the proposed Bayesian measurement-error-driven Hidden Markov regression model is practicable and effective in calibrating the effects of covariates on multistate outcome the prior distribution on measurement errors accrued from two-stage validation design is strongly recommended.

Related Topics

Related Publications

Related Content

Site Footer


This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.