Group-based trajectory modelling is a model-based clustering technique applied for the identification of latent patterns of temporal changes. Despite its manifold applications in clinical and health sciences, potential problems of the model selection procedure are often overlooked. The choice of the number of latent trajectories (class-enumeration), for instance, is to a large degree based on statistical criteria that are not fail-safe. Moreover, the process as a whole is not transparent. To facilitate class enumeration, we introduce a graphical summary display of several fit and model adequacy criteria, the fit-criteria assessment plot.
Methods
An R-code that accepts universal data input is presented. The programme condenses relevant group-based trajectory modelling output information of model fit indices in automated graphical displays. Examples based on real and simulated data are provided to illustrate, assess and validate fit-criteria assessment plot’s utility.
Results
Fit-criteria assessment plot provides an overview of fit criteria on a single page, placing users in an informed position to make a decision. Fit-criteria assessment plot does not automatically select the most appropriate model but eases the model assessment procedure.
Conclusions
Fit-criteria assessment plot is an exploratory, visualisation tool that can be employed to assist decisions in the initial and decisive phase of group-based trajectory modelling analysis. Considering group-based trajectory modelling’s widespread resonance in medical and epidemiological sciences, a more comprehensive, easily interpretable and transparent display of the iterative process of class enumeration may foster group-based trajectory modelling’s adequate use.
NaginDSTremblayRE. Trajectories of boys’ physical aggression, opposition, and hyperactivity on the path to physically violent and nonviolent juvenile delinquency. Child Dev1999; 70: 1181–1196.
2.
NaginDS. Group-based modeling of development, Cambridge, MA: Harvard University Press, 2005.
3.
JonesBLNaginDSRoederK. A SAS procedure based on mixture models for estimating developmental trajectories. Sociol Methods Res2001; 29: 374–393.
4.
JonesBLNaginDS. Advances in group-based trajectory modeling and an SAS procedure for estimating them. Sociol Methods Res2007; 35: 542–571.
5.
JonesBLNaginDS. A note on a Stata plugin for estimating group-based trajectory models. Sociol Methods Res2013; 42: 608–613.
6.
OsgoodDW. Making sense of crime and the life course. Ann Am Acad Pol Soc Sci2005; 602: 196–211.
7.
PiqueroARTaking stock of developmental trajectories of criminal activity over the life course. In: AMLiberman (ed). The long view of crime: A synthesis of longitudinal research. pp.23–78, 2008.
8.
SaundersJM. Understanding random effects in group-based trajectory modeling: an application of Moffitt’s developmental taxonomy. J Drug Issues2010; 40: 195–220.
LaubJHNaginDSSampsonRJ. Trajectories of change in criminal offending: good marriages and the desistance process. Am Sociol Rev1998; 63: 225–238.
12.
NaginDSOdgersCL. Group-based trajectory modeling in clinical research. Annu Rev Clin Psychol2010; 6: 109–138.
13.
NaginDS. Group-based trajectory modeling: an overview. Ann Nutr Metab2014; 65: 205–210.
14.
LiuSJonesRNGlymourMM. Implications of lifecourse epidemiology for research on determinants of adult disease. Public Health Rev2010; 32: 489–511.
15.
TillingKHoweLDBen-ShlomoY. Commentary: Methods for analysing life course influences on health—untangling complex exposures. Int J Epidemiol2011; 40: 250–252.
16.
SandmanCCordovaCDavisEet al.Patterns of fetal heart rate response at ∼30 weeks gestation predict size at birth. J Dev Orig Health Dis2011; 2: 212–217.
17.
NgJBarrettLMWongAet al.The role of longitudinal cohort studies in epigenetic epidemiology: challenges and opportunities. Genome Biol2012; 13: 246–246.
18.
TuY-KTillingKSterneJAet al.A critical evaluation of statistical approaches to examining the role of growth trajectories in the developmental origins of health and disease. Int J Epidemiol2013; 43: 1664–1665.
19.
ZhengHTuminDQianZ. Obesity and mortality risk: new findings from body mass index trajectories. Am J Epidemiol2013; 178: 1591–1599.
20.
ZiyabAHKarmausWKurukulaaratchyRJet al.Developmental trajectories of body mass index from infancy to 18 years of age: prenatal determinants and health consequences. J Epidemiol Community Health2014; 68: 934–941.
21.
CarterMADuboisLTremblayMSet al.Trajectories of childhood weight gain: the relative importance of local environment versus individual social and early life factors. PLoS One2012; 7: e47065–e47065.
22.
Van RyzinMJChathamMKryzerEet al.Identifying atypical cortisol patterns in young children: the benefits of group-based trajectory modeling. Psychoneuroendocrinology2009; 34: 50–61.
23.
FerroMACamfieldCSLevinSDet al.Trajectories of health‐related quality of life in children with epilepsy: a cohort study. Epilepsia2013; 54: 1889–1897.
24.
MaddoxTMRossCTavelHMet al.Blood pressure trajectories and associations with treatment intensification, medication adherence, and outcomes among newly diagnosed coronary artery disease patients. Circ Cardiovasc Qual Outcomes2010; 3: 347–357.
25.
FranklinJMShrankWHPakesJet al.Group-based trajectory models: a new approach to classifying and predicting long-term medication adherence. Med Care2013; 51: 789–796.
26.
WillkeRJZhengZSubediPet al.From concepts, theory, and evidence of heterogeneity of treatment effects to methodological approaches: a primer. BMC Med Res Methodol2012; 12: 185–185.
27.
LevineSZLeuchtS. Treatment response heterogeneity in the predominant negative symptoms of schizophrenia: analysis of amisulpride vs placebo in three clinical trials. Schizophr Res2014; 156: 107–114.
28.
HsuH-CJonesBL. Multiple trajectories of successful aging of older and younger cohorts. Gerontologist2012; 52: 843–856.
29.
ChenCFLeeWCYangHIet al.Changes in serum levels of HBV DNA and alanine aminotransferase determine risk for hepatocellular carcinoma. Gastroenterology2011; 141: 1240–1248e2.
30.
ZaslavskyOCochraneBBHertingJRet al.Application of person‐centered analytic methodology in longitudinal research: Exemplars from the women’s health initiative clinical trial data. Res Nurs Health2014; 37: 53–64.
31.
M’BailaraKCosnefroyOVietaEet al.Group-based trajectory modeling: a novel approach to examining symptom trajectories in acute bipolar episodes. J Affect Disord2013; 145: 36–41.
32.
HuangDYLanzaHIWright-VolelKet al.Developmental trajectories of childhood obesity and risk behaviors in adolescence. J Adolesc2013; 36: 139–148.
33.
ØstbyeTMalhotraRLandermanLR. Body mass trajectories through adulthood: results from the National Longitudinal Survey of Youth 1979 Cohort (1981–2006). Int J Epidemiol2011; 40: 240–250.
34.
SoicherJEMayoNEGauvinLet al.Trajectories of endurance activity following pulmonary rehabilitation in COPD patients. Eur Respir J2012; 39: 272–278.
35.
Von EhrensteinOSMikolajczykRTZhangJ. Timing and trajectories of fetal growth related to cognitive development in childhood. Am J Epidemiol2009; 170: 1388–1395.
36.
BergerRPBazacoMCWagnerAKet al.Trajectory analysis of serum biomarker concentrations facilitates outcome prediction after pediatric traumatic and hypoxemic brain injury. Dev Neurosci2011; 32: 396–405.
37.
SmithORFKupperNDe JongePet al.Distinct trajectories of fatigue in chronic heart failure and their association with prognosis. Eur J Heart Fail2010; 12: 841–848.
38.
SmithORFKupperNDenolletJet al.Vital exhaustion and cardiovascular prognosis in myocardial infarction and heart failure: predictive power of different trajectories. Psychol Med2011; 41: 731–738.
39.
NiyonkuruCWagnerAKOzawaHet al.Group-based trajectory analysis applications for prognostic biomarker model development in severe TBI: a practical example. J Neurotrauma2013; 30: 938–945.
40.
SchmitzNGariépyGSmithKJet al.Trajectories of self-rated health in people with diabetes: associations with functioning in a prospective community sample. PLoS One2013; 8: e83088–e83088.
41.
BauerDJCurranPJ. Distributional assumptions of growth mixture models: implications for overextraction of latent trajectory classes. Psychol Methods2003; 8: 338–363.
42.
SampsonRJLaubJHEgglestonEP. On the robustness and validity of groups. J Quant Criminol2004; 20: 37–42.
43.
NaginDSTremblayRE. Further reflections on modeling and analyzing developmental trajectories: a response to Maughan and Raudenbush. Ann Am Acad Pol Soc Sci2005; 602: 145–154.
44.
NaginDSTremblayRE. From seduction to passion: a response to Sampson and Laub. Criminology2005; 43: 915–915.
45.
RaudenbushSW. How do we study “what happens next”?Ann Am Acad Pol Soc Sci2005; 602: 131–144.
46.
SampsonRJLaubJH. Seductions of method: rejoinder to Nagin and Tremblay’s developments trajectory groups: fact or fiction. Criminology2005; 43: 905–905.
47.
BauerDJReyesHLM. Modeling variability in individual development: differences of degree or kind?Child Dev Perspect2010; 4: 114–122.
48.
SkardhamarT. Distinguishing facts and artifacts in group-based modeling. Criminology2010; 48: 295–320.
49.
BrameRPaternosterRPiqueroAR. Thoughts on the analysis of group-based developmental trajectories in criminology. Justice Q2012; 29: 469–490.
50.
EgglestonEPLaubJHSampsonRJ. Methodological sensitivities to latent class analysis of long-term criminal trajectories. J Quant Criminol2004; 20: 1–26.
51.
BrameRNaginDSWassermanL. Exploring some analytical characteristics of finite mixture models. J Quant Criminol2006; 22: 31–59.
SterbaSKBauerDJ. Statistically evaluating person-oriented principles revisited. Dev Psychopathol2010; 22: 287–294.
54.
MorinAJMaïanoCNagengastBet al.General growth mixture analysis of adolescents' developmental trajectories of anxiety: the impact of untested invariance assumptions on substantive interpretations. Struct Equ Model Multidisc J2011; 18: 613–648.
55.
SterbaSKBaldasaroREBauerDJ. Factors affecting the adequacy and preferability of semiparametric groups-based approximations of continuous growth trajectories. Multivariate Behav Res2012; 47: 590–634.
56.
GilthorpeMDahlyDTuY-Ket al.Challenges in modelling the random structure correctly in growth mixture models and the impact this has on model mixtures. J Dev Orig Health Dis2014; 5: 197–205.
57.
MuthénB. The potential of growth mixture modelling. Infant Child Dev2006; 15: 623–625.
58.
RamNGrimmKJ. Methods and measures: growth mixture modeling: a method for identifying differences in longitudinal change among unobserved groups. Int J Behav Dev2009; 33: 565–576.
59.
PicklesACroudaceT. Latent mixture models for multivariate and longitudinal outcomes. Stat Methods Med Res2009; 19: 271–289.
60.
EroshevaEAMatsuedaRLTelescaD. Breaking bad: two decades of life-course data analysis in criminology, developmental psychology, and beyond. Ann Rev Stat Appl2014; 1: 301–332.
61.
McIntoshCN. Pitfalls in subgroup analysis based on growth mixture models: a commentary on van Leeuwen et al. (2012). Qual Life Res2013; 22: 2625–2629.
62.
AndruffHCarraroNThompsonAet al.Latent class growth modelling: a tutorial. Tutor Quant Methods Psychol2009; 5: 11–24.
63.
NylundKLAsparouhovTMuthénBO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Model Multidisc J2007; 14: 535–569.
64.
TofighiDEndersCKIdentifying the correct number of classes in growth mixture models. In: HancockGRSamuelsenKM (eds). Advances in latent variable mixture models, Greenwich, CT: Information Age Publishing, Inc, 2008, pp. 317–341.
65.
FriesemaIVeenstraMYZwieteringPJet al.Measurement of lifetime alcohol intake: utility of a self-administered questionnaire. Am J Epidemiol2004; 159: 809–817.
66.
FriesemaIHZwieteringPJVeenstraMYet al.The effect of alcohol intake on cardiovascular disease and mortality disappeared after taking lifetime drinking and covariates into account. Alcohol Clin Exp Res2008; 32: 645–651.
67.
VeenstraMFriesemaIZwieteringPet al.Lower prevalence of heart disease but higher mortality risk during follow-up was found among nonrespondents to a cohort study. J Clin Epidemiol2006; 59: 412–420.
68.
VeenstraMYLemmensPHFriesemaIHet al.Coping style mediates impact of stress on alcohol use: a prospective population‐based study. Addiction2007; 102: 1890–1898.
69.
IalongoN. Steps substantive researchers can take to build a scientifically strong case for the existence of trajectory groups. Dev Psychopathol2010; 22: 273–275.
70.
TwiskJHoekstraT. Classifying developmental trajectories over time should be done with great caution: a comparison between methods. J Clin Epidemiol2012; 65: 1078–1087.
ShahNHHipwellAESteppSDet al.Measures of discrimination for latent group-based trajectory models. J Appl Stat2015; 42: 1–11.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
