Latent class models have been widely used in longitudinal studies to uncover unobserved heterogeneity in a population and find the characteristics of the latent classes simultaneously using the class allocation probabilities dependent on predictors. However, previous latent class models for longitudinal data suffer from uncertainty in the choice of the number of latent classes. In this study, we propose a Bayesian nonparametric latent class model for longitudinal data, which allows the number of latent classes to be inferred from the data. The proposed model is an infinite mixture model with predictor-dependent class allocation probabilities; an individual longitudinal trajectory is described by the class-specific linear mixed effects model. The model parameters are estimated using Markov chain Monte Carlo methods. The proposed model is validated using a simulated example and a real-data example for characterizing latent classes of estradiol trajectories over the menopausal transition using data from the Study of Women’s Health Across the Nation.
LairdNMWareJH.
Random-effects models for longitudinal data. Biometrics1982; 963–974.
6.
DigglePJ.
An approach to the analysis of repeated measurements. Biometrics1988; 959–971.
7.
TaylorJMCumberlandWSyJ.
A stochastic model for analysis of longitudinal aids data. J Am Stat Assoc1994;
89(427): 727–736.
8.
QuintanaFAJohnsonWOWaetjenLE, et al.
Bayesian nonparametric longitudinal data analysis. J Am Stat Assoc2016;
111(515): 1168–1181.
9.
Tang N, Wu Y and Chen D. Semiparametric Bayesian analysis of transformation linear mixed models. J Multivari Anal 2018; 166: 225–240.
10.
HanJ.
Initial classification of joint data in EM estimation of latent class joint model. J Multivari Anal2009;
100(10): 2313–2323.
11.
NeelonBSwamyGKBurgetteLF, et al.
A Bayesian growth mixture model to examine maternal hypertension and birth outcomes. Stat Med2011;
30(22): 2721–2735.
12.
LinHMcCullochCETurnbullBW, et al.
A latent class mixed model for analysing biomarker trajectories with irregularly scheduled observations. Stat Med2000;
19(10): 1303–1318.
13.
LinHTurnbullBWMcCullochCE, et al.
Latent class models for joint analysis of longitudinal biomarker and event process data: application to longitudinal prostate-specific antigen readings and prostate cancer. J Am Stat Assoc2002;
97(457): 53–65.
14.
Proust-LimaCJolyPDartiguesJF, et al.
Joint modelling of multivariate longitudinal outcomes and a time-to-event: a nonlinear latent class approach. Comput Stat Data Anal2009;
53(4): 1142–1154.
15.
Proust-LimaCTaylorJM.
Development and validation of a dynamic prognostic tool for prostate cancer recurrence using repeated measures of posttreatment PSA: a joint modeling approach. Biostatistics2009;
10(3): 535–549.
16.
CeleuxGForbesFRobertCP, et al.
Deviance information criteria for missing data models. Bayesian Anal2006;
1(4): 651–673.
17.
PlummerM.
Penalized loss functions for Bayesian model comparison. Biostatistics2008;
9(3): 523–539.
18.
YangHO’BrienSDunsonDB.
Nonparametric Bayes stochastically ordered latent class models. J Am Stat Assoc2011;
106(495): 807–817.
19.
NasserinejadKvan RosmalenJde KortW, et al.
Comparison of criteria for choosing the number of classes in Bayesian finite mixture models. PLoS One2017;
12(1): e0168838.
20.
PanJCHuangGH.
Bayesian inferences of latent class models with an unknown number of classes. Psychometrika2014;
79(4): 621–646.
21.
LiuYLiuLZhouJ.
Joint latent class model of survival and longitudinal data: an application to CPCRA study. Comput Stat Data Anal2015;
91: 40–50.
22.
Fahrmeir L, Kneib T, et al. Bayesian smoothing and regression for longitudinal, spatial and event history data. OUP Catalogue, 2011.
23.
ZhangDLinXRazJ, et al.
Semiparametric stochastic mixed models for longitudinal data. J Am Stat Assoc1998;
93(442): 710–719.
24.
MackenzieMLDonovanCMcArdleB.
Regression spline mixed models: a forestry example. J Agric Biol Environ Stat2005;
10(4): 394.
25.
ShiJWangB.
Curve prediction and clustering with mixtures of Gaussian process functional regression models. Stat Comput2008;
18(3): 267–283.
26.
IshwaranHZarepourM.
Exact and approximate sum representations for the Dirichlet process. Can J Stat2002;
30(2): 269–283.
27.
Miller JW and Harrison MT. A simple example of Dirichlet process mixture inconsistency for the number of components. In: Burges CJC, Bottou L, Welling M et al. (eds.) Advances in Neural Information Processing Systems 26. Curran Associates, Inc., 2013. pp. 199–206.
28.
Sowers MFR, Crawford S, Sternfeld B, et al. SWAN: a multicenter, multiethnic, community-based cohort study of women and the menopausal transition. Women’s Health Research Faculty Publications. 505. 2000.
29.
GoldEBCrawfordSLAvisNE, et al.
Factors related to age at natural menopause: longitudinal analyses from swan. Am J Epidemiol2013;
178(1): 70–83.
30.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 1996–1997: baseline dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-08-09. https://doi.org/10.3886/ICPSR28762.v4.
31.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 1997–1999: visit 01 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-08-09. https://doi.org/10.3886/ICPSR29221.v2.
32.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 1998–2000: visit 02 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-08-23. https://doi.org/10.3886/ICPSR29401.v3.
33.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 1999–2001: visit 03 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-08-22. https://doi.org/10.3886/ICPSR29701.v3.
34.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2000–2002: visit 04 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-09-24. https://doi.org/10.3886/ICPSR30142.v3.
35.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2001–2003: visit 05 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-10-18. https://doi.org/10.3886/ICPSR30501.v2.
36.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2002–2004: visit 06 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-11-13. https://doi.org/10.3886/ICPSR31181.v2.
37.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2003–2005: visit 07 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-11-19. https://doi.org/10.3886/ICPSR31901.v2.
38.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2004–2006: visit 08 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-11-20. https://doi.org/10.3886/ICPSR32122.v2.
39.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2005–2007: visit 09 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-11-20. https://doi.org/10.3886/ICPSR32721.v2.
40.
Sutton-Tyrrell K, Selzer F, Sowers MFR, et al. Study of Women’s Health Across the Nation (SWAN), 2006–2008: visit 10 dataset. Ann Arbor, MI: Inter-University Consortium for Political and Social Research, 2018-11-15. https://doi.org/10.3886/ICPSR32961.v2.
41.
TepperPGRandolphJFJrMcConnellDS, et al.
Trajectory clustering of estradiol and follicle-stimulating hormone during the menopausal transition among women in the Study of Women’s Health Across the Nation (SWAN). J Clin Endocrinol Metab2012;
97(8): 2872–2880.
42.
BurgerHGHaleGRobertsonD, et al.
A review of hormonal changes during the menopausal transition: focus on findings from the Melbourne women’s midlife health project. Hum Reprod Update2007;
13(6): 559–565.
