A flexible approach to modelling stillbirths using the foetuses at risk approach.
Cox model
Foetuses-at-risk
piecewise additive mixed models
survival analysis
Journal
Paediatric and perinatal epidemiology
ISSN: 1365-3016
Titre abrégé: Paediatr Perinat Epidemiol
Pays: England
ID NLM: 8709766
Informations de publication
Date de publication:
08 2023
08 2023
Historique:
revised:
31
05
2023
received:
16
02
2023
accepted:
04
06
2023
medline:
15
8
2023
pubmed:
24
6
2023
entrez:
24
6
2023
Statut:
ppublish
Résumé
Survival analysis methods are increasingly used to model the gestational age-specific risk of perinatal phenomena such as stillbirth. To compare two types of survival analysis models, and highlight differences by estimating the relationships between pre-pregnancy BMI and gestational age-specific rates of stillbirth. The study was based on singleton live births and stillbirths in the United States in 2016-2017, with data obtained from the natality and fetal death files of the National Center for Health Statistics. We compared Cox regression versus piecewise exponential additive mixed models (PAMMs) for modelling the relationship between BMI and stillbirth across gestational age. In a second analysis, we illustrated the performance of both models for assessing the relationship between the trimester-specific number of cigarettes smoked, a time-dependent covariate, and stillbirth. The study population included 7,567,316 births, of which 42,739 were stillbirths (5.6 per 1000 total births). Stillbirth rates increased with increasing pre-pregnancy BMI and increasing gestational age. In analyses with BMI as a categorical variable, the Cox model and PAMM models yielded similar results. Analyses of BMI as a continuous variable also showed similar results when BMI associations were assumed to be linear, and the changes in gestational age-specific rates were modelled parametrically. However, results differed slightly when PAMMs, modelled with data-driven approaches, were used to estimate changes in BMI effects across gestational age; PAMMs provided a more nuanced modelling of time-varying effects. PAMM models showed an approximately linear increase in the effect of smoking on stillbirth with increasing gestational age. For survival analyses using the foetuses-at-risk approach, PAMMs provide a valuable alternative to the traditional Cox model, with increased modelling flexibility when proportional hazards assumptions are violated.
Sections du résumé
BACKGROUND
Survival analysis methods are increasingly used to model the gestational age-specific risk of perinatal phenomena such as stillbirth.
OBJECTIVES
To compare two types of survival analysis models, and highlight differences by estimating the relationships between pre-pregnancy BMI and gestational age-specific rates of stillbirth.
METHODS
The study was based on singleton live births and stillbirths in the United States in 2016-2017, with data obtained from the natality and fetal death files of the National Center for Health Statistics. We compared Cox regression versus piecewise exponential additive mixed models (PAMMs) for modelling the relationship between BMI and stillbirth across gestational age. In a second analysis, we illustrated the performance of both models for assessing the relationship between the trimester-specific number of cigarettes smoked, a time-dependent covariate, and stillbirth.
RESULTS
The study population included 7,567,316 births, of which 42,739 were stillbirths (5.6 per 1000 total births). Stillbirth rates increased with increasing pre-pregnancy BMI and increasing gestational age. In analyses with BMI as a categorical variable, the Cox model and PAMM models yielded similar results. Analyses of BMI as a continuous variable also showed similar results when BMI associations were assumed to be linear, and the changes in gestational age-specific rates were modelled parametrically. However, results differed slightly when PAMMs, modelled with data-driven approaches, were used to estimate changes in BMI effects across gestational age; PAMMs provided a more nuanced modelling of time-varying effects. PAMM models showed an approximately linear increase in the effect of smoking on stillbirth with increasing gestational age.
CONCLUSIONS
For survival analyses using the foetuses-at-risk approach, PAMMs provide a valuable alternative to the traditional Cox model, with increased modelling flexibility when proportional hazards assumptions are violated.
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
547-554Informations de copyright
© 2023 The Authors. Paediatric and Perinatal Epidemiology published by John Wiley & Sons Ltd.
Références
Yudkin PL, Wood L, Redman CW. Risk of unexplained stillbirth at different gestational ages. Lancet Lond Engl. 1987;1(8543):1192-1194.
Kramer MS, Liu S, Luo Z, et al. Analysis of perinatal mortality and its components: time for a change? Am J Epidemiol. 2002;156(6):493-497.
Platt RW, Joseph KS, Ananth CV, Grondines J, Abrahamowicz M, Kramer MS. A proportional hazards model with time-dependent covariates and time-varying effects for analysis of fetal and infant death. Am J Epidemiol. 2004;160(3):199-206.
Joseph KS. Theory of obstetrics: “the fetuses-at-risk approach” as a causal paradigm. J Obstet Gynaecol Can. 2004;26(11):953-956.
Platt RW. The fetuses-at-risk approach: an evolving paradigm. In: Buck Louis GM, Platt RW, eds. Reproductive and Perinatal Epidemiology. Oxford University Press; 2011.
Joseph KS, Kramer MS. The fetuses-at-risk approach: survival analysis from a fetal perspective. Acta Obstet Gynecol Scand. 2018;97(4):454-465.
Janevic T, Zeitlin J, Auger N, et al. Association of race/ethnicity with very preterm neonatal morbidities. JAMA Pediatr. 2018;172(11):1061-1069.
Yao R, Ananth CV, Park BY, Pereira L, Plante LA. Obesity and the risk of stillbirth: a population-based cohort study. Am J Obstet Gynecol. 2014;210(5):457.e1-457.e9.
Cox DR. Regression models and life-tables. J R Stat Soc Ser B Methodol. 1972;34(2):187-220.
Wood SN. Generalized Additive Models: An Introduction with R. 2nd ed. CRC Press; 2017:341.
Bender A, Groll A, Scheipl F. A generalized additive model approach to time-to-event analysis. Stat Model. 2018;18(3-4):299-321.
Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. Springer; 2000.
Wood SN. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J R Stat Soc Ser B Stat Methodol. 2011;73(1):3-36.
Wood SN, Li Z, Shaddick G, Augustin NH. Generalized additive models for Gigadata: modeling the U.K. black smoke network daily data. J Am Stat Assoc. 2017;112(519):1199-1210.
Therneau TM. (Until 2009) TL (original S >R port and R maintainer) Elizabeth A, Cynthia C. Survival: Survival Analysis [Internet]. 2022 https://CRAN.R-project.org/package=survival
Bender A, Scheipl F. Pammtools: piece-wise exponential additive mixed modeling tools. arXiv. 2018. http://arxiv.org/abs/1806.01042
Wei LJ. The accelerated failure time model: a useful alternative to the Cox regression model in survival analysis. Stat Med. 1992;11(14-15):1871-1879.
Nardi A, Schemper M. Comparing Cox and parametric models in clinical studies. Stat Med. 2003;22(23):3597-3610.
Chu SY, Kim SY, Lau J, et al. Maternal obesity and risk of stillbirth: a metaanalysis. Am J Obstet Gynecol. 2007;197(3):223-228.
Aune D, Saugstad OD, Henriksen T, Tonstad S. Maternal body mass index and the risk of fetal death, stillbirth, and infant death: a systematic review and meta-analysis. Jama. 2014;311(15):1536-1546.
Bodnar LM, Ness RB, Markovic N, Roberts JM. The risk of preeclampsia rises with increasing prepregnancy body mass index. Ann Epidemiol. 2005;15(7):475-482.
Lisonkova S, Joseph KS. Incidence of preeclampsia: risk factors and outcomes associated with early- versus late-onset disease. Am J Obstet Gynecol. 2013;209(6):544.e1-544.e12.
Najafi F, Hasani J, Izadi N, et al. The effect of prepregnancy body mass index on the risk of gestational diabetes mellitus: a systematic review and dose-response meta-analysis. Obes Rev. 2019;20(3):472-486.
Marufu TC, Ahankari A, Coleman T, Lewis S. Maternal smoking and the risk of still birth: systematic review and meta-analysis. BMC Public Health. 2015;15:239.
Ramjith J, Bender A, Roes KCB, Jonker MA. Recurrent events analysis with piece-wise exponential additive mixed models. Statistical Modelling. 2022. doi:10.1177/1471082X221117612