Age-period-cohort models: A comparative study of available methodologies

Chris Robertson, Sara Gandini, Peter Boyle

Research output: Contribution to journalArticle

83 Citations (Scopus)

Abstract

This article compares the estimates produced by a number of solutions to the identifiability problem in age-period-cohort models using a series of disease rates with known structure. The results suggest that only those methods that are based on the estimable functions such as curvatures can be recommended for use in all circumstances. The other common approaches that give parameter estimates that are easier to interpret all have induced bias in the estimates. In particular methods based on the minimization of a penalty function to achieve identifiability are only of use if there is no change in the rates with time. Any drift in the rates tends to be expressed as a cohort-based trend. The methods based on individual records introduce a bias if there is a strong age effect in the direction of a decreasing cohort trend and a compensating increase based on period effects. The nonparametric testing method has little power to detect trends in the rates in small tables but ascribes a strong drift in the rates to both period and cohort trends. With careful interpretation, all methods estimate nonlinear components correctly. Copyright (C) 1999 Elsevier Science Inc.

Original languageEnglish
Pages (from-to)569-583
Number of pages15
JournalJournal of Clinical Epidemiology
Volume52
Issue number6
DOIs
Publication statusPublished - Jun 1999

Fingerprint

Direction compound

Keywords

  • Age-period-cohort models
  • Chronic disease rates
  • Poisson regression
  • Simulation

ASJC Scopus subject areas

  • Medicine(all)
  • Public Health, Environmental and Occupational Health
  • Epidemiology

Cite this

Age-period-cohort models : A comparative study of available methodologies. / Robertson, Chris; Gandini, Sara; Boyle, Peter.

In: Journal of Clinical Epidemiology, Vol. 52, No. 6, 06.1999, p. 569-583.

Research output: Contribution to journalArticle

@article{7578307094424f91be5a4f453926d9b7,
title = "Age-period-cohort models: A comparative study of available methodologies",
abstract = "This article compares the estimates produced by a number of solutions to the identifiability problem in age-period-cohort models using a series of disease rates with known structure. The results suggest that only those methods that are based on the estimable functions such as curvatures can be recommended for use in all circumstances. The other common approaches that give parameter estimates that are easier to interpret all have induced bias in the estimates. In particular methods based on the minimization of a penalty function to achieve identifiability are only of use if there is no change in the rates with time. Any drift in the rates tends to be expressed as a cohort-based trend. The methods based on individual records introduce a bias if there is a strong age effect in the direction of a decreasing cohort trend and a compensating increase based on period effects. The nonparametric testing method has little power to detect trends in the rates in small tables but ascribes a strong drift in the rates to both period and cohort trends. With careful interpretation, all methods estimate nonlinear components correctly. Copyright (C) 1999 Elsevier Science Inc.",
keywords = "Age-period-cohort models, Chronic disease rates, Poisson regression, Simulation",
author = "Chris Robertson and Sara Gandini and Peter Boyle",
year = "1999",
month = "6",
doi = "10.1016/S0895-4356(99)00033-5",
language = "English",
volume = "52",
pages = "569--583",
journal = "Journal of Clinical Epidemiology",
issn = "0895-4356",
publisher = "Elsevier USA",
number = "6",

}

TY - JOUR

T1 - Age-period-cohort models

T2 - A comparative study of available methodologies

AU - Robertson, Chris

AU - Gandini, Sara

AU - Boyle, Peter

PY - 1999/6

Y1 - 1999/6

N2 - This article compares the estimates produced by a number of solutions to the identifiability problem in age-period-cohort models using a series of disease rates with known structure. The results suggest that only those methods that are based on the estimable functions such as curvatures can be recommended for use in all circumstances. The other common approaches that give parameter estimates that are easier to interpret all have induced bias in the estimates. In particular methods based on the minimization of a penalty function to achieve identifiability are only of use if there is no change in the rates with time. Any drift in the rates tends to be expressed as a cohort-based trend. The methods based on individual records introduce a bias if there is a strong age effect in the direction of a decreasing cohort trend and a compensating increase based on period effects. The nonparametric testing method has little power to detect trends in the rates in small tables but ascribes a strong drift in the rates to both period and cohort trends. With careful interpretation, all methods estimate nonlinear components correctly. Copyright (C) 1999 Elsevier Science Inc.

AB - This article compares the estimates produced by a number of solutions to the identifiability problem in age-period-cohort models using a series of disease rates with known structure. The results suggest that only those methods that are based on the estimable functions such as curvatures can be recommended for use in all circumstances. The other common approaches that give parameter estimates that are easier to interpret all have induced bias in the estimates. In particular methods based on the minimization of a penalty function to achieve identifiability are only of use if there is no change in the rates with time. Any drift in the rates tends to be expressed as a cohort-based trend. The methods based on individual records introduce a bias if there is a strong age effect in the direction of a decreasing cohort trend and a compensating increase based on period effects. The nonparametric testing method has little power to detect trends in the rates in small tables but ascribes a strong drift in the rates to both period and cohort trends. With careful interpretation, all methods estimate nonlinear components correctly. Copyright (C) 1999 Elsevier Science Inc.

KW - Age-period-cohort models

KW - Chronic disease rates

KW - Poisson regression

KW - Simulation

UR - http://www.scopus.com/inward/record.url?scp=0033153187&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033153187&partnerID=8YFLogxK

U2 - 10.1016/S0895-4356(99)00033-5

DO - 10.1016/S0895-4356(99)00033-5

M3 - Article

C2 - 10408997

AN - SCOPUS:0033153187

VL - 52

SP - 569

EP - 583

JO - Journal of Clinical Epidemiology

JF - Journal of Clinical Epidemiology

SN - 0895-4356

IS - 6

ER -