Using linear and natural cubic splines, SITAR, and latent trajectory models to characterise nonlinear longitudinal growth trajectories in cohort studies
Tóm tắt
Longitudinal data analysis can improve our understanding of the influences on health trajectories across the life-course. There are a variety of statistical models which can be used, and their fitting and interpretation can be complex, particularly where there is a nonlinear trajectory. Our aim was to provide an accessible guide along with applied examples to using four sophisticated modelling procedures for describing nonlinear growth trajectories. This expository paper provides an illustrative guide to summarising nonlinear growth trajectories for repeatedly measured continuous outcomes using (i) linear spline and (ii) natural cubic spline linear mixed-effects (LME) models, (iii) Super Imposition by Translation and Rotation (SITAR) nonlinear mixed effects models, and (iv) latent trajectory models. The underlying model for each approach, their similarities and differences, and their advantages and disadvantages are described. Their application and correct interpretation of their results is illustrated by analysing repeated bone mass measures to characterise bone growth patterns and their sex differences in three cohort studies from the UK, USA, and Canada comprising 8500 individuals and 37,000 measurements from ages 5–40 years. Recommendations for choosing a modelling approach are provided along with a discussion and signposting on further modelling extensions for analysing trajectory exposures and outcomes, and multiple cohorts. Linear and natural cubic spline LME models and SITAR provided similar summary of the mean bone growth trajectory and growth velocity, and the sex differences in growth patterns. Growth velocity (in grams/year) peaked during adolescence, and peaked earlier in females than males e.g., mean age at peak bone mineral content accrual from multicohort SITAR models was 12.2 years in females and 13.9 years in males. Latent trajectory models (with trajectory shapes estimated using a natural cubic spline) identified up to four subgroups of individuals with distinct trajectories throughout adolescence. LME models with linear and natural cubic splines, SITAR, and latent trajectory models are useful for describing nonlinear growth trajectories, and these methods can be adapted for other complex traits. Choice of method depends on the research aims, complexity of the trajectory, and available data. Scripts and synthetic datasets are provided for readers to replicate trajectory modelling and visualisation using the R statistical computing software.
Tài liệu tham khảo
Ben-Shlomo Y, Cooper R, Kuh D. The last two decades of life course epidemiology, and its relevance for research on ageing. Int J Epidemiol. 2016;45(4):973–88.
Grimm KJ, Ram N, Hamagami F. Nonlinear growth curves in developmental research. Child Dev. 2011;82(5):1357–71.
Howe LD, Tilling K, Matijasevich A, Petherick ES, Santos AC, Fairley L, et al. Linear spline multilevel models for summarising childhood growth trajectories: a guide to their application using examples from five birth cohorts. Stat Methods Med Res. 2016;25(5):1854–74.
Lourenço BH, Villamor E, Augusto RA, Cardoso MA. Influence of early life factors on body mass index trajectory during childhood: a population-based longitudinal analysis in the Western Brazilian Amazon. Matern Child Nutr. 2015;11(2):240–52.
Cole TJ, Donaldson MD, Ben-Shlomo Y. SITAR--a useful instrument for growth curve analysis. Int J Epidemiol. 2010;39(6):1558–66.
Herle M, Micali N, Abdulkadir M, Loos R, Bryant-Waugh R, Hübel C, et al. Identifying typical trajectories in longitudinal data: modelling strategies and interpretations. Eur J Epidemiol. 2020;35(3):205–22.
Sauerbrei W, Abrahamowicz M, Altman DG, le Cessie S, Carpenter J, initiative obotS. STRengthening analytical thinking for observational studies: the STRATOS initiative. Stat Med. 2014;33(30):5413–32.
Tu YK, Tilling K, Sterne JA, Gilthorpe MS. A critical evaluation of statistical approaches to examining the role of growth trajectories in the developmental origins of health and disease. Int J Epidemiol. 2013;42(5):1327–39.
Curran PJ, Obeidat K, Losardo D. Twelve frequently asked questions about growth curve modeling. J Cogn Dev. 2010;11(2):121–36.
Macdonald-Wallis C, Lawlor DA, Palmer T, Tilling K. Multivariate multilevel spline models for parallel growth processes: application to weight and mean arterial pressure in pregnancy. Stat Med. 2012;31(26):3147–64.
Twisk JW. Longitudinal data analysis. A comparison between generalized estimating equations and random coefficient analysis. Eur J Epidemiol. 2004;19(8):769–76.
Fitzmaurice G, Davidian M, Verbeke G, Molenberghs G. Longitudinal data analysis. USA: Chapman & Hall/CRC; 2009.
Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38(4):963–74.
Goldstein H, De Stavola B. Statistical modelling of repeated measurement data. Longitud Life Course Stud. 2010;1(2):170–85.
Cole TJ: sitar: Super Imposition by Translation and Rotation growth curve analysis. R package version 1.2.0. 2021. https://cran.r-project.org/web/packages/sitar/index.html.
Perperoglou A, Sauerbrei W, Abrahamowicz M, Schmid M. A review of spline function procedures in R. BMC Med Res Methodol. 2019;19(1):46.
Suk HW, West SG, Fine KL, Grimm KJ. Nonlinear growth curve modeling using penalized spline models: a gentle introduction. Psychol Methods. 2019;24(3):269–90.
Aris IM, Bernard JY, Chen LW, Tint MT, Pang WW, Lim WY, et al. Infant body mass index peak and early childhood cardio-metabolic risk markers in a multi-ethnic Asian birth cohort. Int J Epidemiol. 2017;46(2):513–25.
Fonseca MJ, Moreira C, Santos AC. Adiposity rebound and cardiometabolic health in childhood: results from the generation XXI birth cohort. Int J Epidemiol. 2021;50(4):1260-71.
Harrell F. Regression modeling strategies with applications to linear models, logistic regression, and survival analysis. 1st ed. New York: Springer; 2001.
Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med. 2010;29(9):1037–57.
Mackenzie ML, Donovan CR, McArdle BH. Regression spline mixed models: A forestry example. J Agric Biol Environ Stat. 2005;10(4):394.
James G, Witten D, Hastie T, Tibshirani R. An introduction to statistical learning with applications in R. New York: Springer; 2017.
Naumova EN, Must A, Laird NM. Tutorial in biostatistics: evaluating the impact of ‘critical periods’ in longitudinal studies of growth using piecewise mixed effects models. Int J Epidemiol. 2001;30(6):1332–41.
Beath KJ. Infant growth modelling using a shape invariant model with random effects. Stat Med. 2007;26(12):2547–64.
Cole TJ, Kuh D, Johnson W, Ward KA, Howe LD, Adams JE, et al. Using super-imposition by translation and rotation (SITAR) to relate pubertal growth to bone health in later life: the Medical Research Council (MRC) National Survey of health and development. Int J Epidemiol. 2016.
Berlin KS, Parra GR, Williams NA. An introduction to latent variable mixture modeling (part 2): longitudinal latent class growth analysis and growth mixture models. J Pediatr Psychol. 2013;39(2):188–203.
van de Schoot R, Sijbrandij M, Winter SD, Depaoli S, Vermunt JK. The GRoLTS-checklist: guidelines for reporting on latent trajectory studies. Struct Equ Model Multidiscip J. 2017;24(3):451–67.
Lennon H, Kelly S, Sperrin M, Buchan I, Cross AJ, Leitzmann M, et al. Framework to construct and interpret latent class trajectory modelling. BMJ Open. 2018;8(7):e020683.
Proust-Lima C, Philipps V, Liquet B. Estimation of extended mixed models using latent classes and latent processes: the R package lcmm. J Stat Softw. 2017;1(Issue 2):2017.
Harvey N, Dennison E, Cooper C. Osteoporosis: a lifecourse approach. J Bone Miner Res. 2014;29(9):1917–25.
Fraser A, Macdonald-Wallis C, Tilling K, Boyd A, Golding J, Davey Smith G, et al. Cohort profile: the Avon longitudinal study of parents and children: ALSPAC mothers cohort. Int J Epidemiol. 2013;42(1):97–110.
Boyd A, Golding J, Macleod J, Lawlor DA, Fraser A, Henderson J, et al. Cohort profile: the ‘children of the 90s’-the index offspring of the Avon longitudinal study of parents and children. Int J Epidemiol. 2013;42(1):111–27.
McCormack SE, Cousminer DL, Chesi A, Mitchell JA, Roy SM, Kalkwarf HJ, et al. Association between linear growth and bone accrual in a diverse cohort of children and adolescents. JAMA Pediatr. 2017;171(9):e171769.
Baxter-Jones AD, Faulkner RA, Forwood MR, Mirwald RL, Bailey DA. Bone mineral accrual from 8 to 30 years of age: an estimation of peak bone mass. J Bone Miner Res. 2011;26(8):1729–39.
Nowok B, Raab GM, Dibben C. synthpop: bespoke creation of synthetic data in R. J Stat Softw. 2016;74(11):26.
Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects models using lme4. J Stat Software. 2015;67(1):1–48. https://doi.org/10.18637/jss.v067.i01.
Elhakeem A, Frysz M, Tilling K, Tobias JH, Lawlor DA. Association between age at puberty and bone accrual from 10 to 25 years of age. JAMA Netw Open. 2019;2(8):e198918.
Jackowski SA, Erlandson MC, Mirwald RL, Faulkner RA, Bailey DA, Kontulainen SA, et al. Effect of maturational timing on bone mineral content accrual from childhood to adulthood: evidence from 15 years of longitudinal data. Bone. 2011;48(5):1178–85.
Cousminer DL, Mitchell JA, Chesi A, Roy SM, Kalkwarf HJ, Lappe JM, et al. Genetically determined later puberty impacts lowered bone mineral density in childhood and adulthood. J Bone Miner Res. 2018;33(3):430–6.
Khera AV, Chaffin M, Wade KH, Zahid S, Brancale J, Xia R, et al. Polygenic prediction of weight and obesity trajectories from birth to adulthood. Cell. 2019;177(3):587–596.e589.
Jensen SM, Ritz C, Ejlerskov KT, Mølgaard C, Michaelsen KF. Infant BMI peak, breastfeeding, and body composition at age 3 y. Am J Clin Nutr. 2014;101(2):319–25.
Cousminer DL, Wagley Y, Pippin JA, Elhakeem A, Way GP, Pahl MC, et al. Genome-wide association study implicates novel loci and reveals candidate effector genes for longitudinal pediatric bone accrual. Genome Biol. 2021;22(1):1.
O'Keeffe LM, Simpkin AJ, Tilling K, Anderson EL, Hughes AD, Lawlor DA, et al. Sex-specific trajectories of measures of cardiovascular health during childhood and adolescence: a prospective cohort study. Atherosclerosis. 2018;278:190–6.
Lambert PC, Abrams KR, Jones DR, Halligan AW, Shennan A. Analysis of ambulatory blood pressure monitor data using a hierarchical model incorporating restricted cubic splines and heterogeneous within-subject variances. Stat Med. 2001;20(24):3789–805.
Snijders T. Power and sample size in multilevel modeling. In: Everitt BS, Howell DC, editors. Encyclopedia of Statistics in Behavioral Science. Chicester: Wiley; 2005.
Guo Y, Logan HL, Glueck DH, Muller KE. Selecting a sample size for studies with repeated measures. BMC Med Res Methodol. 2013;13(1):100.
Simpkin AJ, Sayers A, Gilthorpe MS, Heron J, Tilling K. Modelling height in adolescence: a comparison of methods for estimating the age at peak height velocity. Ann Hum Biol. 2017;44(8):715–22.
Tilling K, Macdonald-Wallis C, Lawlor DA, Hughes RA, Howe LD. Modelling childhood growth using fractional polynomials and linear splines. Ann Nutr Metab. 2014;65(2–3):129–38.
Kwong ASF, Manley D, Timpson NJ, Pearson RM, Heron J, Sallis H, et al. Identifying critical points of trajectories of depressive symptoms from childhood to young adulthood. J Youth Adolesc. 2019;48(4):815–27.
Cole TJ. Optimal design for longitudinal studies to estimate pubertal height growth in individuals. Ann Hum Biol. 2018;45(4):314–20.
Wood SN. Generalized additive models an introduction with R. 2nd ed: Chapman & Hall/CRC; USA: 2017.
Pedersen EJ, Miller DL, Simpson GL, Ross N. Hierarchical generalized additive models in ecology: an introduction with mgcv. PeerJ. 2019;7:e6876.
Wood SN. mgcv: Mixed GAM Computation Vehicle with Automatic Smoothness Estimation. 2021. p. 1.8-136 https://cran.r-project.org/web/packages/mgcv/index.html.
Wood SN, Scheipl F. gamm4: generalized additive mixed models using ‘mgcv’and ‘lme4’; 2017. p. 0.2–5. http://cran.nexr.com/web/packages/gamm4/index.html
Kohli N, Harring JR, Zopluoglu C. A finite mixture of nonlinear random coefficient models for continuous repeated measures data. Psychometrika. 2016;81(3):851–80.
Lock EF, Kohli N, Bose M. Detecting multiple random changepoints in Bayesian piecewise growth mixture models. Psychometrika. 2018;83(3):733–50.
Ding M, Chavarro JE, Fitzmaurice GM. Development of a mixture model allowing for smoothing functions of longitudinal trajectories. Stat Methods Med Res. 2021;30(2):549–62.
Buscot M-J, Thomson RJ, Juonala M, Sabin MA, Burgner DP, Lehtimäki T, et al. Distinct child-to-adult body mass index trajectories are associated with different levels of adult cardiometabolic risk. Eur Heart J. 2018;39(24):2263–70.
Kwong ASF, Lopez-Lopez JA, Hammerton G, Manley D, Timpson NJ, Leckie G, et al. Genetic and environmental risk factors associated with trajectories of depression symptoms from adolescence to young adulthood. JAMA Netw Open. 2019;2(6):e196587.
Elhakeem A, Heron J, Tobias JH, Lawlor DA. Physical activity throughout adolescence and peak hip strength in young adults. JAMA Netw Open. 2020;3(8):e2013463.
Hulman A, Witte DR, Vistisen D, Balkau B, Dekker JM, Herder C, et al. Pathophysiological characteristics underlying different glucose response curves: a latent class trajectory analysis from the prospective EGIR-RISC study. Diabetes Care. 2018;41(8):1740–8.
Lévêque E, Lacourt A, Philipps V, Luce D, Guénel P, Stücker I, et al. A new trajectory approach for investigating the association between an environmental or occupational exposure over lifetime and the risk of chronic disease: application to smoking, asbestos, and lung cancer. Plos One. 2020;15(8):e0236736.
Lawlor DA, Tilling K, Davey Smith G. Triangulation in aetiological epidemiology. Int J Epidemiol. 2016;45(6):1866–86.
Madden JM, Li X, Kearney PM, Tilling K, Fitzgerald AP. Exploring diurnal variation using piecewise linear splines: an example using blood pressure. Emerg Themes Epidemiol. 2017;14:1–1.
Brilleman SL, Howe LD, Wolfe R, Tilling K. Bayesian piecewise linear mixed models with a random change point: an application to BMI rebound in childhood. Epidemiology. 2017;28(6):827–33.
Crozier SR, Johnson W, Cole TJ, Macdonald-Wallis C, Muniz-Terrera G, Inskip HM, et al. A discussion of statistical methods to characterise early growth and its impact on bone mineral content later in childhood. Ann Hum Biol. 2019;46(1):17–26.
Sayers A, Heron J, Smith A, Macdonald-Wallis C, Gilthorpe MS, Steele F, et al. Joint modelling compared with two stage methods for analysing longitudinal data and prospective outcomes: a simulation study of childhood growth and BP. Stat Methods Med Res. 2017;26(1):437–52.
Parker RMA, Leckie G, Goldstein H, Howe LD, Heron J, Hughes AD, et al. Joint modeling of individual trajectories, within-individual variability, and a later outcome: systolic blood pressure through childhood and left ventricular mass in early adulthood. Am J Epidemiol. 2021;190(4):652-62.
Smith AD, Hardy R, Heron J, Joinson CJ, Lawlor DA, Macdonald-Wallis C, et al. A structured approach to hypotheses involving continuous exposures over the life course. Int J Epidemiol. 2016;45(4):1271–9.
Lee KJ, Tilling K, Cornish RP, Little RJ, Bell ML, Goetghebeur E, et al. Framework for the treatment and reporting of missing data in observational studies: the TARMOS framework. J Clin Epidemiol. 2021;134:79-88
van Buuren S. Flexible imputation of missing data. 2nd ed. Chapman & Hall/CRC. USA; 2018.
Matteo Quartagno SG, Carpenter J. jomo: a flexible package for two-level joint modelling multiple imputation. R J. 2019;11(2):205–28.
Hughes RA, Heron J, Sterne JAC, Tilling K. Accounting for missing data in statistical analyses: multiple imputation is not always the answer. Int J Epidemiol. 2019;48(4):1294–304.
Twisk J, de Boer M, de Vente W, Heymans M. Multiple imputation of missing values was not necessary before performing a longitudinal mixed-model analysis. J Clin Epidemiol. 2013;66(9):1022–8.
Huque MH, Carlin JB, Simpson JA, Lee KJ. A comparison of multiple imputation methods for missing data in longitudinal studies. BMC Med Res Methodol. 2018;18(1):168.
Huque MH, Moreno-Betancur M, Quartagno M, Simpson JA, Carlin JB, Lee KJ. Multiple imputation methods for handling incomplete longitudinal and clustered data where the target analysis is a linear mixed effects model. Biom J. 2020;62(2):444–66.
VanderWeele TJ. Principles of confounder selection. Eur J Epidemiol. 2019;34(3):211–9.
Groenwold RHH, Palmer TM, Tilling K. To Adjust or Not to Adjust? When a "Confounder" Is Only Measured After Exposure. Epidemiology. 2021;32(2):194-201. https://doi.org/10.1097/EDE.0000000000001312.
Lipsitch M, Tchetgen Tchetgen E, Cohen T. Negative controls: a tool for detecting confounding and bias in observational studies. Epidemiology. 2010;21(3):383–8.
Taylor K, Elhakeem A, Nader JLT, Yang T, Isaevska E, Richiardi L, et al. Effect of maternal prepregnancy/early-pregnancy BMI and pregnancy smoking and alcohol on congenital heart diseases: a parental negative control study. J Am Heart Assoc. 2021;10(11):e020051
Brion MJ, Lawlor DA, Matijasevich A, Horta B, Anselmi L, Araújo CL, et al. What are the causal effects of breastfeeding on IQ, obesity and blood pressure? Evidence from comparing high-income with middle-income cohorts. Int J Epidemiol. 2011;40(3):670–80.
Wills AK, Lawlor DA, Matthews FE, Aihie Sayer A, Bakra E, Ben-Shlomo Y, et al. Life course trajectories of systolic blood pressure using longitudinal data from eight UK cohorts. PLoS Med. 2011;8(6):e1000440.
Jaddoe VWV, Felix JF, Andersen AN, Charles MA, Chatzi L, Corpeleijn E, et al. The LifeCycle project-EU child cohort network: a federated analysis infrastructure and harmonized data of more than 250,000 children and parents. Eur J Epidemiol. 2020;35(7):709–24.
Ronkainen J, Nedelec R, Atehortua A, Balkhiyarova Z, Zhanna A, Dang V, et al. LongITools: dynamic longitudinal exposome trajectories in cardiovascular and metabolic non-communicable diseases. Environ Epidemiol. 2021;6(1):e184. https://doi.org/10.1097/EE9.0000000000000184.
Hughes RA, Tilling K, Lawlor DA. Combining longitudinal data from different cohorts to examine the life-course trajectory. Am J Epidemiol. 2021;190(12):2680-9.
Pinot de Moira A, Haakma S, Strandberg-Larsen K, van Enckevort E, Kooijman M, Cadman T, et al. The EU Child Cohort Network’s core data: establishing a set of findable, accessible, interoperable and re-usable (FAIR) variables. Eur J Epidemiol. 2021;36(5):565–80.
Nader JL, López M, Julvez J, Guxens M, Cadman T, Elhakeem A, et al. Cohort description: measures of early-life behaviour and later psychopathology in the LifeCycle project - EU child cohort network. J Epidemiol. 2021. (Epub ahead of print). https://doi.org/10.2188/jea.JE20210241.
Baxter-Jones AD, Burrows M, Bachrach LK, Lloyd T, Petit M, Macdonald H, et al. International longitudinal pediatric reference standards for bone mineral content. Bone. 2010;46(1):208–16.