Hidden Markov models are stochastic models in which the observations are assumed to follow a mixture distribution, but the parameters of the components are governed by a Markov chain which is unobservable. The issues related to the estimation of Poisson-hidden Markov models in which the observations are coming from mixture of Poisson distributions and the parameters of the component Poisson distributions are governed by an m-state Markov chain with an unknown transition probability matrix are explained here. These methods were applied to the data on Vibrio cholerae counts reported every month for 11-year span at Christian Medical College, Vellore, India. Using Viterbi algorithm, the best estimate of the state sequence was obtained and hence the transition probability matrix. The mean passage time between the states were estimated. The 95% confidence interval for the mean passage time was estimated via Monte Carlo simulation. The three hidden states of the estimated Markov chain are labelled as ‘Low’, ‘Moderate’ and ‘High’ with the mean counts of 1.4, 6.6 and 20.2 and the estimated average duration of stay of 3, 3 and 4 months, respectively. Environmental risk factors were studied using Markov ordinal logistic regression analysis. No significant association was found between disease severity levels and climate components.
BoxGEPJenkinsGMReinselGC. Time series analysis, Hoboken, NJ: John Wiley and Sons, Inc., 2008.
2.
CoorayTMJA. Applied time series analysis and forecasting, New Delhi: Narosa Publishing House, 2008.
3.
MacDonaldILZucchiniW. Hidden Markov and other models for discrete-valued time series, New York: CRC Press, 2000.
4.
ZucchiniWMacDonaldIL. Hidden Markov models for time series: an introduction using R, New York: CRC Press, 2009.
5.
AliMKimDYunusMet al.Time series analysis of cholera in Matlab, Bangladesh, during 1988-2001. J Health Popul Nutr2013; 31: 11–19.
6.
SebastianTAnandanSJeyaseelanVet al.Role of seasonality and rainfall in vibrio cholerae infections: a time series model for 11 years surveillance data. Clin Epidemiol Glob Health2015; 3: 144–148.
7.
ZegerSL. A regression model for time series of counts. Biometrika1988; 75: 621–629.
8.
KuhnLDavidsonLDurkinM. Use of Poisson regression and time series analysis for detecting changes over time in rates of child injury following a prevention program. Am J Epidemiol1994; 140: 943–955.
9.
QuddusM. Time series count data models: an empirical application to traffic accidents. Accid Anal Prev2008; 40: 1732–1741.
10.
ChristensenADethlefsenCLundbye-ChristensenSet al.Modeling gradually changing seasonal variation in count data using state space models: a cohort study of hospitalization rates of stroke in atrial brillation patients in Denmark from 1977 to 2011. BMC Med Res Methodol2012; 12: e174–e174.
11.
ChandySRamanathanKManoharanAet al.Assessing effect of climate on the incidence of dengue in Tamil nadu. Indian J Med Microbiol2013; 31: 283–286.
12.
SiriyasatienPPhumeeAOngrukPet al.Analysis of significant factors for dengue fever incidence prediction. BMC Bioinformatics2016; 17: 1–9.
13.
BrandtPWilliamsJFordhamBet al.Dynamic modeling for persistent event-count time series. Am J Pol Sci2000; 44: 823–843.
14.
BrandtPWilliamsJ. A linear Poisson autoregressive model: the Poisson AR(p) Model. Pol Anal2001; 9: 164–184.
Olteanu M, Garca-Garaluz E, Atencia M, et al. Modelling dengue epidemics with autoregressive switching Markov models. In: BioInspired systems: computational and ambient intelligence: 10th international work-conference on artificial neural networks, 2009, vol. 1, pp.886–892
17.
StratYLCarratF. Monitoring epidemiologic surveillance data using hidden Markov models. Stat Med1999; 18: 3468–3478.
18.
GreenPRichardsonS. Hidden markov models and disease mapping. J Am Stat Assoc2002; 97: 1055–1070.
19.
CooperBLipsitchM. The analysis of hospital infection data using hidden Markov models. Biostatistics2004; 5: 223–237.
20.
WallMMLiR. Multiple indicator hidden Markov model with an application to medical utilization data. Stat Med2009; 28: 293–310.
21.
ShirleyKESmallDSLynchKGet al.Hidden Markov models for alcoholism treatment trial data. Ann Appl Stat2010; 4: 366–396.
22.
RafeiAPashaEOrakR. Tuberculosis surveillance using a hidden Markov model. Iran J Public Health2012; 41: 87–96.
23.
PanSWuHYenAMet al.A Markov regression random effects model for remission of functional disability in patients following a first stroke: a Bayesian approach. Stat Med2007; 26: 5335–5353.
24.
ChamroukhiFSameAGovaertGet al.Time series modeling by a regression approach based on a latent process. Neural netw2009; 22: 593–602.
25.
JeyaseelanVSebastianTLakshmananJet al.Longitudinal data analysis of mean passage time among malnutrition states: an application of Markov chains. J Appl Stat2016; 43: 2729–2739.
26.
AgrestiA. Categorical data analysis, New York: Wiley-Interscience, 2002.
27.
DigglePHeagertyPLiangKet al.Analysis of longitudinal data, New York: Oxford University Press, 1994.
28.
KemenyJGSnellJL. Finite markov chains, New York: Springer, 1983.
29.
FraserA. Hidden Markov models and dynamical systems, Los Alamos, New Mexico: Society for Industrial and Applied Mathematics, 2008.
30.
RabinerLRJuangBH. An introduction to hidden Markov models. IEEE ASSP Magazine1986; 3: 4–16.
31.
Rabiner L. A tutorial on hidden Markov models and selected applications in speech recognition. In: Proceedings of the IEEE, 1989, pp.257–286
32.
AlbertPS. A two-state Markov mixture model for a time series of epileptic seizure counts. Biometrics1991; 47: 1371–1381.
33.
McBrydeEPettittACooperBSet al.Characterizing an outbreak of vancomycin-resistant enterococci using hidden Markov models. J R Soc Interface2007; 4: 745–754.
34.
HunterJJ. Variances of first passage times in a Markov chain with applications to mixing times. Lin Algebra Appl2008; 429: 1135–1162.
35.
EisenbergMKujbidaGTuiteAet al.Examining rainfall and cholera dynamics in haiti using statistical and dynamic modeling approaches. Epidemics2013; 5: 197–207.
36.
BerghFHollowayJPPienaarMet al.A comparison of various modelling approaches applied to cholera case data. ORiON2008; 24: 17–36.
37.
RajendranKSumiABhattachariyaMet al.Influence of relative humidity in vibrio cholerae infection: a time series model. Indian J Med Res2011; 133: 138–145.
38.
EarnestAChenMNgDet al.Using autoregressive integrated moving average (arima) models to predict and monitor the number of beds occupied during a sars outbreak in a tertiary hospital in Singapore. BMC Health Serv Res2005; 5: e36–e36.
39.
HuWTongSMengersenKet al.Weather variability and the incidence of cryptosporidiosis: comparison of time series poisson regression and SARIMA models. Ann Epidemiol2007; 17: 679–688.
40.
ReyburnRKimDEmchMet al.Climate variability and the outbreaks of cholera in Zanzibar, East Africa: a time series analysis. Am J Trop Med Hyg2011; 84: 862–869.
41.
HilbeJ. Negative binomial regression, New York: Cambridge University Press, 2011.
42.
HilbeJ. Modeling count data, New York: Cambridge University Press, 2014.
43.
WatkinsREaglesonSVeenendaalBet al.Disease surveillance using a hidden Markov model. BMC Med Inform Decis Mak2009; 9: e39–e39.
44.
WangPPutermanML. Analysis of longitudinal data of epileptic seizure counts: a two-state hidden Markov regression approach. Biom J2001; 43: 941–962.
45.
DoanTKongDMarshallCet al.Characterising the transmission dynamics of acinetobacter baumannii in intensive care units using hidden Markov models. PLoS ONE2015; 10: e7–e7.
46.
BartolomeoNTrerotoliPSerioG. Progression of liver cirrhosis to HCC: an application of hidden Markov model. BMC Med Res Methodol2011; 11: e38–e38.
47.
ZhangYWuTDaigleBJet al.Identification of disease states associated with coagulopathy in trauma. BMC Med Inform Decis Mak2016; 16: 124–124.
48.
ScottSL. Bayesian methods for hidden Markov models: recursive computing in the 21st century. J Am Stat Assoc2002; 97: 337–351.
