A critical problem in repeated measurement studies is the occurrence of nonignorable missing observations. A common approach to deal with this problem is joint modeling the longitudinal and survival processes for each individual on the basis of a random effect that is usually assumed to be time constant. We relax this hypothesis by introducing time-varying subject-specific random effects that follow a first-order autoregressive process, AR(1). We also adopt a generalized linear model formulation to accommodate for different types of longitudinal response (i.e. continuous, binary, count) and we consider some extended cases, such as counts with excess of zeros and multivariate outcomes at each time occasion. Estimation of the parameters of the resulting joint model is based on the maximization of the likelihood computed by a recursion developed in the hidden Markov literature. This maximization is performed on the basis of a quasi-Newton algorithm that also provides the information matrix and then standard errors for the parameter estimates. The proposed approach is illustrated through a Monte Carlo simulation study and the analysis of certain medical datasets.
LittleRJARubinDB. Statistical analysis with missing data, Hoboken, NJ: John Wiley Sons, 2002.
2.
DigglePKenwardMG. Informative drop-out in longitudinal data analysis (with discussion). Journal of The Royal Statistical Society, Series C Appl Stat1994; 43: 49–93.
3.
LittleRJA. Pattern-mixture models for multivariate incomplete data. J Am Stat Assoc1993; 88: 125–134.
4.
WuMCCarrollRJ. Estimation and comparison of changes in the presence of informative right censoring by modeling the censoring process. Biometrics1988; 44: 175–188.
5.
WuMCBaileyK. Analysing changes in the presence of informative right censoring caused by death and withdrawal. Stat Med1988; 7: 337–346.
6.
WuMCBaileyK. Estimation and comparison of changes in the presence of informative right censoring: conditional linear model. Biometrics1989; 45: 939–955.
7.
FollmannDWuM. An approximate generalized liner model with random effects for informative missing data. Biometrics1995; 51: 151–168.
8.
HoganJWLairdNM. Mixture models for the joint distribution of repeated measures and event times. Stat Med1997; 16: 239–257.
9.
HoganJWLairdNM. Increasing efficiency from censored survival data by using random effects to model longitudinal covariates. Stat Methods Med Res1998; 7: 28–48.
10.
WulfsohnMSTsiatisAA. A joint model for survival and longitudinal data measured with error. Biometrics1997; 53: 330–339.
11.
HendersonRDigglePDobsonA. Joint modelling of longitudinal measurements and event time data. Biostatistics2000; 1: 465–480.
12.
TsiatisAADavidianM. Joint modeling of longitudinal and time-to-event data: an overview. Stat Sin2004; 14: 809–834.
13.
RizopoulousD. Joint models for longitudinal and time-to-event data with applications in R, Boca Raton, FL: Chapman & Hall/CRC Press, 2012.
14.
McCullochCESearleSRNeuhausJM. Generalized, linear, and mixed models, New York: Wiley, 2008.
15.
CoxDR. Regression models and life-tables. J R Stat Soc Ser B1972; 34: 187–220.
16.
RizopoulosDVerbekeGMolenberghsG. Shared parameter models under random effects misspecification. Biometrika2008; 95: 63–74.
17.
HuangXStefanskiLADavidianM. Latent-model robustness in joint models for a primary endpoint and a longitudinal process. Biometrics2009; 65: 719–727.
18.
SongXDavidianMTsiatisAA. A semiparametric likelihood approach to joint modeling of longitudinal and time-to-event data. Biometrics2002; 58: 742–753.
19.
Proust-LimaCSéneMTaylorJM, et al.Joint latent class models for longitudinal and time-to-event data: a review. Stat Methods Med Res2014; 23: 74–90.
20.
XuJZegerSL. Joint analysis of longitudinal data comprising repeated measures and times to events. Appl Stat2001; 50: 375–387.
21.
TaylorJMGCumberlandWGSyJP. A stochastic model for analysis of longitudinal AIDS data. J Am Stat Assoc1994; 89: 727–736.
22.
LavalleyMPDeGruttolaV. Model for empirical Bayes estimators of longitudinal CD4 counts. Stat Med1996; 15: 2289–2305.
23.
WangYTaylorJMG. Jointly modeling longitudinal and event time data with application to acquired immunodeficiency syndrome. J Am Stat Assoc2001; 96: 895–905.
24.
BarrettJDigglePHendersonR, et al.Joint modelling of repeated measurements and time-to-event outcomes: flexible model specification and exact likelihood inference. J R Stat Soc Ser B2015; 77: 131–148.
25.
BartolucciFFarcomeniA. A discrete time-event history approach to informative drop-out in multivariate latent Markov models with covariates. Biometrics2015; 71: 80–89.
26.
AzzaliniA. The skew-normal distribution and related multivariate families (with discussion). Scand J Stat2005; 32: 159–200.
27.
BartolucciFBacciSPennoniF. Longitudinal analysis of self-reported health status by mixture latent auto-regressive models. J R Stat Soc Ser C2014; 63: 267–288.
28.
BaumLEPetrieTSoulesG, et al.A maximization technique occurring in the statistical analysis of probabilistic functions of Markov chains. Ann Math Stat1970; 41: 164–171.
29.
ZucchiniWMacDonaldIL. Hidden Markov models for time series: an introduction using R, Boca Raton, FL: Chapman & Hall/CRC Press, 2009.
30.
BartolucciFFarcomeniAPennoniF. Latent Markov models for longitudinal data, Boca Raton, FL: Chapman and Hall/CRC Press, 2013.
31.
LambertD. Zero-inflated Poisson regression, with an application to defects in manufacturing. Technometrics1992; 34: 1–14.
32.
MinYAgrestiA. Random effect models for repeated measures of zero-inflated count data. Stat Model2005; 5: 1–19.
33.
HeissF. Sequential numerical integration in nonlinear state space models for microeconometric panel data. J Appl Econ2008; 23: 373–389.
34.
WelchLR. Hidden Markov models and the Baum-Welch algorithm. IEEE Inform Theory Soc Newslett2003; 53: 1–13.
35.
LystigTCHughesJP. Exact computation of the observed information matrix for hidden Markov models. J Comput Graph Stat2002; 11: 678–689.
36.
XuSJonesRHGrunwaldGK. Analysis of longitudinal count data with serial correlation. Biometr J2007; 49: 416–428.
37.
DigglePJHeagertyPLiangK-Y, et al.Analysis of longitudinal data, Oxford: Oxford University Press, 2002.
38.
Akaike H. Information theory and an extension of the maximum likelihood principle. In: Petrov BN and Csaki F (eds) Second International symposium of information theory. Budapest: Akademiai Kiado, 267–281.
39.
SchwarzG. Estimating the dimension of a model. Ann Stat1978; 6: 461–464.
40.
GreenbergERBaronJAStukelTA, et al.A clinical trial of beta carotene to prevent basal-cell and squamous-cell cancers of the skin. The N Engl J Med1990; 323: 789–795.
41.
HasanMTSneddonGMaR. Pattern-mixture zero-inflated mixed models for longitudinal unbalanced count data with excessive zeros. Biometr J2009; 51: 946–960.
42.
MurtaughPADicksonERVan DamGM, et al.Primary biliary cirrhosis: prediction of short-term survival based on repeated patient visits. Hepatology1994; 20: 126–134.
43.
RizopoulosD. Fast fitting of joint models for longitudinal and event time data using a pseudo-adaptive Gaussian quadrature rule. Comput Stat Data Anal2012; 56: 491–501.
