EUVIMEDOpdag EUVIMED
Open access · CC BY 4.0

Functional Principal Component Analysis as an Alternative to Mixed‐Effect Models for Describing Sparse Repeated Measures in Presence of Missing Data

Corentin Ségalas iD, Catherine Helmer iD, Robin Genuer iD, Cécile Proust-Lima iD

Statistics in Medicine · 2024

DOI: 10.1002/sim.10214 ↗

At a glance

Journal article in Statistics in Medicine, 2024. Written by Corentin Ségalas and 3 co-authors. Institution: Univ. Bordeaux, INSERM, INRIA, BPH, U1219, Bordeaux, France. Major MEDLINE topics: Principal Component Analysis, Computer Simulation, Models, Statistical. Cited 6 times so far (Crossref); 48 references. The full text is freely available (open access, CC BY 4.0). Funded by: France 2030 Program / RRI PHDS.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Analyzing longitudinal data in health studies is challenging due to sparse and error‐prone measurements, strong within‐individual correlation, missing data and various trajectory shapes. While mixed‐effect models (MM) effectively address these challenges, they remain parametric models and may incur computational costs. In contrast, functional principal component analysis (FPCA) is a non‐parametric approach developed for regular and dense functional data that flexibly describes temporal trajectories at a potentially lower computational cost. This article presents an empirical simulation study evaluating the behavior of FPCA with sparse and error‐prone repeated measures and its robustness under different missing data schemes in comparison with MM. The results show that FPCA is well‐suited in the presence of missing at random data caused by dropout, except in scenarios involving most frequent and systematic dropout. Like MM, FPCA fails under missing not at random mechanism. The FPCA was applied to describe the trajectories of four cognitive functions before clinical dementia and contrast them with those of matched controls in a case‐control study nested in a population‐based aging cohort. The average cognitive declines of future dementia cases showed a sudden divergence from those of their matched controls with a sharp acceleration 5 to 2.5 years prior to diagnosis.

Subjects

MeSH terms (MEDLINE)

Principal Component AnalysisComputer SimulationModels, StatisticalHumansLongitudinal StudiesDementiaCase-Control StudiesData Interpretation, Statistical

Author keywords

functional principal component analysismissing datamixed modelssparse functional data

Authors and institutions

  • Univ. Bordeaux, INSERM, INRIA, BPH, U1219, Bordeaux, France.
  • Univ. Bordeaux, INSERM, BPH, U1219, Bordeaux, France.

ORCID profiles: Corentin Ségalas, Catherine Helmer, Robin Genuer, Cécile Proust-Lima

Funding

  • France 2030 Program / RRI PHDS

Cite

APA

Ségalas, C., Helmer, C., Genuer, R., & Proust-Lima, C. (2024). Functional Principal Component Analysis as an Alternative to Mixed‐Effect Models for Describing Sparse Repeated Measures in Presence of Missing Data. Statistics in Medicine. https://doi.org/10.1002/sim.10214

Vancouver

Ségalas C, Helmer C, Genuer R, Proust-Lima C. Functional Principal Component Analysis as an Alternative to Mixed‐Effect Models for Describing Sparse Repeated Measures in Presence of Missing Data. Statistics in Medicine. 2024. doi:10.1002/sim.10214

BibTeX
@article{Sgalas2024,
  author = {Ségalas, Corentin and Helmer, Catherine and Genuer, Robin and Proust-Lima, Cécile},
  title = {Functional Principal Component Analysis as an Alternative to Mixed‐Effect Models for Describing Sparse Repeated Measures in Presence of Missing Data},
  journal = {Statistics in Medicine},
  year = {2024},
  doi = {10.1002/sim.10214}
}
Start for free →