Modeling Risk Factors for Intraindividual Variability: A Mixed-Effects Beta-Binomial Model Applied to Cognitive Function in Older People in the English Longitudinal Study of Ageing

Richard M A Parker; Kate Tilling; Graciela Muniz Terrera; Jessica K. Barrett

doi:10.1093/aje/kwad169

Modeling Risk Factors for Intraindividual Variability: A Mixed-Effects Beta-Binomial Model Applied to Cognitive Function in Older People in the English Longitudinal Study of Ageing

Richard M A Parker^*, Kate Tilling, Graciela Muniz Terrera, Jessica K. Barrett

^*Corresponding author for this work

Research output: Contribution to journal › Article (Academic Journal) › peer-review

Abstract

Cognitive functioning in older age profoundly impacts quality of life and health. Whilst most research in cognition in older age has focussed on mean levels, intraindividual variability (IIV) around this may have risk factors and outcomes independent of the mean. Investigating risk factors associated with IIV has typically involved deriving a summary statistic for each person from residual
error around a fitted mean. However, this ignores uncertainty in the estimates, prohibits exploring associations with time-varying factors, and is biased by floor/ceiling effects. To address this, we propose a mixed-effects location scale beta-binomial model to estimate average probability and IIV in a word recall test in the English Longitudinal Study of Aging. After adjusting for mean
performance, an analysis of 9,873 individuals across 7 (mean: 3.4) waves (2002-2015), found IIV greater: at older ages; with lower education; in females; with more difficulties with activities of daily living; in later cohorts; and when interviewers recorded issues potentially affecting the tests. Our study introduces a novel method to identify groups with greater IIV in bounded discrete outcomes. Our findings have implications for daily functioning and care, with further work needed to identify the impact for future health outcomes.

Original language	English
Article number	kwad169
Pages (from-to)	159-169
Number of pages	11
Journal	American Journal of Epidemiology
Volume	193
Issue number	1
Early online date	11 Aug 2023
DOIs	https://doi.org/10.1093/aje/kwad169
Publication status	Published - 1 Jan 2024

Bibliographical note

Funding Information:
This work was supported by Medical Research Council (MRC) grant MR/N027485/1 awarded to K.T. to fund the work of R.M.A.P. The MRC and the University of Bristol support the MRC Integrative Epidemiology Unit (grant MC_UU_00011/3). J.K.B. was supported by MRC Unit Programme grant MC_UU_00002/5. The English Longitudinal Study of Ageing (ELSA) was developed by a team of researchers based at University College London, the National Centre for Social Research, the Institute for Fiscal Studies, the University of Manchester, and the University of East Anglia (all in the United Kingdom). The data for ELSA were collected by the National Centre for Social Research. Funding for ELSA is currently provided by the National Institute on Aging, US National Institutes of Health (grant R01AG017644), and a consortium of UK government departments coordinated by the National Institute for Health Research. Funding for ELSA has also been provided by the Economic and Social Research Council. The ELSA data set is available to UK Data Service registered users subject to an End User License Agreement (http://www.ukdataservice.ac.uk; DOI: 10.5255/UKDA-SN-5050-24). A preprint of this article is available on medRxiv at https://doi.org/10.1101/2022.06.01.22275869 (posted June 2, 2022). The developers and funders of ELSA do not bear any responsibility for the analyses or interpretations presented here. Conflict of interest: none declared.

Publisher Copyright:
© The Author(s) 2023. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health.

Keywords

heteroscedasticity
mixed effects model
beta binomial
intraindividual variability
Cognitive test
older adults
Bayesian hierarchical model

Access to Document

10.1093/aje/kwad169Licence: CC BY

Handle.net

Persistent link

(IEU) Methods for modelling within-individual variation
Tilling, K. M. (Principal Investigator)
30/06/17 → 30/12/19
Project: Research

HPC (High Performance Computing) and HTC (High Throughput Computing) Facilities
Alam, S. R. (Manager), Eccleston, P. E. (Other), Williams, D. A. G. (Manager) & Atack, S. H. (Other)
Facility/equipment: Facility
Research Data Storage Facility (RDSF)
Alam, S. R. (Manager), Williams, D. A. G. (Manager) & Eccleston, P. E. (Manager)
IT Services
Facility/equipment: Facility

Cite this

@article{dd0830dfa866458aa77a0da9e1ed0bcc,

title = "Modeling Risk Factors for Intraindividual Variability: A Mixed-Effects Beta-Binomial Model Applied to Cognitive Function in Older People in the English Longitudinal Study of Ageing",

abstract = "Cognitive functioning in older age profoundly impacts quality of life and health. Whilst most research in cognition in older age has focussed on mean levels, intraindividual variability (IIV) around this may have risk factors and outcomes independent of the mean. Investigating risk factors associated with IIV has typically involved deriving a summary statistic for each person from residualerror around a fitted mean. However, this ignores uncertainty in the estimates, prohibits exploring associations with time-varying factors, and is biased by floor/ceiling effects. To address this, we propose a mixed-effects location scale beta-binomial model to estimate average probability and IIV in a word recall test in the English Longitudinal Study of Aging. After adjusting for meanperformance, an analysis of 9,873 individuals across 7 (mean: 3.4) waves (2002-2015), found IIV greater: at older ages; with lower education; in females; with more difficulties with activities of daily living; in later cohorts; and when interviewers recorded issues potentially affecting the tests. Our study introduces a novel method to identify groups with greater IIV in bounded discrete outcomes. Our findings have implications for daily functioning and care, with further work needed to identify the impact for future health outcomes.",

keywords = "heteroscedasticity, mixed effects model, beta binomial, intraindividual variability, Cognitive test, older adults, Bayesian hierarchical model",

author = "Parker, {Richard M A} and Kate Tilling and {Muniz Terrera}, Graciela and Barrett, {Jessica K.}",

note = "Funding Information: This work was supported by Medical Research Council (MRC) grant MR/N027485/1 awarded to K.T. to fund the work of R.M.A.P. The MRC and the University of Bristol support the MRC Integrative Epidemiology Unit (grant MC_UU_00011/3). J.K.B. was supported by MRC Unit Programme grant MC_UU_00002/5. The English Longitudinal Study of Ageing (ELSA) was developed by a team of researchers based at University College London, the National Centre for Social Research, the Institute for Fiscal Studies, the University of Manchester, and the University of East Anglia (all in the United Kingdom). The data for ELSA were collected by the National Centre for Social Research. Funding for ELSA is currently provided by the National Institute on Aging, US National Institutes of Health (grant R01AG017644), and a consortium of UK government departments coordinated by the National Institute for Health Research. Funding for ELSA has also been provided by the Economic and Social Research Council. The ELSA data set is available to UK Data Service registered users subject to an End User License Agreement (http://www.ukdataservice.ac.uk; DOI: 10.5255/UKDA-SN-5050-24). A preprint of this article is available on medRxiv at https://doi.org/10.1101/2022.06.01.22275869 (posted June 2, 2022). The developers and funders of ELSA do not bear any responsibility for the analyses or interpretations presented here. Conflict of interest: none declared. Publisher Copyright: {\textcopyright} The Author(s) 2023. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health.",

year = "2024",

month = jan,

day = "1",

doi = "10.1093/aje/kwad169",

language = "English",

volume = "193",

pages = "159--169",

journal = "American Journal of Epidemiology",

issn = "0002-9262",

publisher = "Oxford University Press",

number = "1",

}

Modeling Risk Factors for Intraindividual Variability: A Mixed-Effects Beta-Binomial Model Applied to Cognitive Function in Older People in the English Longitudinal Study of Ageing. / Parker, Richard M A ; Tilling, Kate; Muniz Terrera, Graciela et al.
In: American Journal of Epidemiology, Vol. 193, No. 1, kwad169, 01.01.2024, p. 159-169.

Research output: Contribution to journal › Article (Academic Journal) › peer-review

TY - JOUR

T1 - Modeling Risk Factors for Intraindividual Variability

T2 - A Mixed-Effects Beta-Binomial Model Applied to Cognitive Function in Older People in the English Longitudinal Study of Ageing

AU - Parker, Richard M A

AU - Tilling, Kate

AU - Muniz Terrera, Graciela

AU - Barrett, Jessica K.

N1 - Funding Information: This work was supported by Medical Research Council (MRC) grant MR/N027485/1 awarded to K.T. to fund the work of R.M.A.P. The MRC and the University of Bristol support the MRC Integrative Epidemiology Unit (grant MC_UU_00011/3). J.K.B. was supported by MRC Unit Programme grant MC_UU_00002/5. The English Longitudinal Study of Ageing (ELSA) was developed by a team of researchers based at University College London, the National Centre for Social Research, the Institute for Fiscal Studies, the University of Manchester, and the University of East Anglia (all in the United Kingdom). The data for ELSA were collected by the National Centre for Social Research. Funding for ELSA is currently provided by the National Institute on Aging, US National Institutes of Health (grant R01AG017644), and a consortium of UK government departments coordinated by the National Institute for Health Research. Funding for ELSA has also been provided by the Economic and Social Research Council. The ELSA data set is available to UK Data Service registered users subject to an End User License Agreement (http://www.ukdataservice.ac.uk; DOI: 10.5255/UKDA-SN-5050-24). A preprint of this article is available on medRxiv at https://doi.org/10.1101/2022.06.01.22275869 (posted June 2, 2022). The developers and funders of ELSA do not bear any responsibility for the analyses or interpretations presented here. Conflict of interest: none declared. Publisher Copyright: © The Author(s) 2023. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health.

PY - 2024/1/1

Y1 - 2024/1/1

N2 - Cognitive functioning in older age profoundly impacts quality of life and health. Whilst most research in cognition in older age has focussed on mean levels, intraindividual variability (IIV) around this may have risk factors and outcomes independent of the mean. Investigating risk factors associated with IIV has typically involved deriving a summary statistic for each person from residualerror around a fitted mean. However, this ignores uncertainty in the estimates, prohibits exploring associations with time-varying factors, and is biased by floor/ceiling effects. To address this, we propose a mixed-effects location scale beta-binomial model to estimate average probability and IIV in a word recall test in the English Longitudinal Study of Aging. After adjusting for meanperformance, an analysis of 9,873 individuals across 7 (mean: 3.4) waves (2002-2015), found IIV greater: at older ages; with lower education; in females; with more difficulties with activities of daily living; in later cohorts; and when interviewers recorded issues potentially affecting the tests. Our study introduces a novel method to identify groups with greater IIV in bounded discrete outcomes. Our findings have implications for daily functioning and care, with further work needed to identify the impact for future health outcomes.

AB - Cognitive functioning in older age profoundly impacts quality of life and health. Whilst most research in cognition in older age has focussed on mean levels, intraindividual variability (IIV) around this may have risk factors and outcomes independent of the mean. Investigating risk factors associated with IIV has typically involved deriving a summary statistic for each person from residualerror around a fitted mean. However, this ignores uncertainty in the estimates, prohibits exploring associations with time-varying factors, and is biased by floor/ceiling effects. To address this, we propose a mixed-effects location scale beta-binomial model to estimate average probability and IIV in a word recall test in the English Longitudinal Study of Aging. After adjusting for meanperformance, an analysis of 9,873 individuals across 7 (mean: 3.4) waves (2002-2015), found IIV greater: at older ages; with lower education; in females; with more difficulties with activities of daily living; in later cohorts; and when interviewers recorded issues potentially affecting the tests. Our study introduces a novel method to identify groups with greater IIV in bounded discrete outcomes. Our findings have implications for daily functioning and care, with further work needed to identify the impact for future health outcomes.

KW - heteroscedasticity

KW - mixed effects model

KW - beta binomial

KW - intraindividual variability

KW - Cognitive test

KW - older adults

KW - Bayesian hierarchical model

U2 - 10.1093/aje/kwad169

DO - 10.1093/aje/kwad169

M3 - Article (Academic Journal)

C2 - 37579319

SN - 0002-9262

VL - 193

SP - 159

EP - 169

JO - American Journal of Epidemiology

JF - American Journal of Epidemiology

IS - 1

M1 - kwad169

ER -

Modeling Risk Factors for Intraindividual Variability: A Mixed-Effects Beta-Binomial Model Applied to Cognitive Function in Older People in the English Longitudinal Study of Ageing

Abstract

Bibliographical note

Keywords

Access to Document

Handle.net

Fingerprint

Projects

(IEU) Methods for modelling within-individual variation

Equipment

HPC (High Performance Computing) and HTC (High Throughput Computing) Facilities

Research Data Storage Facility (RDSF)

Cite this