Smoothing in occupational cohort studies

An illustration based on penalised splines

E. A. Eisen, Ilir Agalliu, S. W. Thurston, B. A. Coull, H. Checkoway

Research output: Contribution to journalArticle

53 Citations (Scopus)

Abstract

Aims: To illustrate the contribution of smoothing methods to modelling exposure-response data, Cox models with penalised splines were used to reanalyse lung cancer risk in a cohort of workers exposed to silica in California's diatomaceous earth industry. To encourage application of this approach, computer code is provided. Methods: Relying on graphic plots of hazard ratios as smooth functions of exposure, the sensitivity of the curve to amount of smoothing, length of the exposure lag, and the influence of the highest exposures was evaluated. Trimming and data transformations were used to down-weight influential observations. Results: The estimated hazard ratio increased steeply with cumulative silica exposure before flattening and then declining over the sparser regions of exposure. The curve was sensitive to changes in degrees of freedom, but insensitive to the number or location of knots. As the length of lag increased, so did the maximum hazard ratio, but the shape was similar. Deleting the two highest exposed subjects eliminated the top half of the range and allowed the hazard ratio to continue to rise. The shape of the splines suggested a parametric model with log hazard as a linear function of log transformed exposure would fit well. Conclusions: This flexible statistical approach reduces the dependence on a priori assumptions, while pointing to a suitable parametric model if one exists. In the absence of an appropriate parametric form, however, splines can provide exposure-response information useful for aetiological research and public health intervention.

Original languageEnglish (US)
Pages (from-to)854-860
Number of pages7
JournalOccupational and Environmental Medicine
Volume61
Issue number10
DOIs
StatePublished - Oct 2004
Externally publishedYes

Fingerprint

smoothing
Silicon Dioxide
Cohort Studies
Diatomaceous Earth
Proportional Hazards Models
hazard
Lung Neoplasms
Industry
Public Health
Weights and Measures
Research
silica
diatomite
exposure
public health
industry
modeling

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health
  • Environmental Science(all)

Cite this

Smoothing in occupational cohort studies : An illustration based on penalised splines. / Eisen, E. A.; Agalliu, Ilir; Thurston, S. W.; Coull, B. A.; Checkoway, H.

In: Occupational and Environmental Medicine, Vol. 61, No. 10, 10.2004, p. 854-860.

Research output: Contribution to journalArticle

Eisen, E. A. ; Agalliu, Ilir ; Thurston, S. W. ; Coull, B. A. ; Checkoway, H. / Smoothing in occupational cohort studies : An illustration based on penalised splines. In: Occupational and Environmental Medicine. 2004 ; Vol. 61, No. 10. pp. 854-860.
@article{b97c9fa3e0034ca7abc25f4f4f6198f6,
title = "Smoothing in occupational cohort studies: An illustration based on penalised splines",
abstract = "Aims: To illustrate the contribution of smoothing methods to modelling exposure-response data, Cox models with penalised splines were used to reanalyse lung cancer risk in a cohort of workers exposed to silica in California's diatomaceous earth industry. To encourage application of this approach, computer code is provided. Methods: Relying on graphic plots of hazard ratios as smooth functions of exposure, the sensitivity of the curve to amount of smoothing, length of the exposure lag, and the influence of the highest exposures was evaluated. Trimming and data transformations were used to down-weight influential observations. Results: The estimated hazard ratio increased steeply with cumulative silica exposure before flattening and then declining over the sparser regions of exposure. The curve was sensitive to changes in degrees of freedom, but insensitive to the number or location of knots. As the length of lag increased, so did the maximum hazard ratio, but the shape was similar. Deleting the two highest exposed subjects eliminated the top half of the range and allowed the hazard ratio to continue to rise. The shape of the splines suggested a parametric model with log hazard as a linear function of log transformed exposure would fit well. Conclusions: This flexible statistical approach reduces the dependence on a priori assumptions, while pointing to a suitable parametric model if one exists. In the absence of an appropriate parametric form, however, splines can provide exposure-response information useful for aetiological research and public health intervention.",
author = "Eisen, {E. A.} and Ilir Agalliu and Thurston, {S. W.} and Coull, {B. A.} and H. Checkoway",
year = "2004",
month = "10",
doi = "10.1136/oem.2004.013136",
language = "English (US)",
volume = "61",
pages = "854--860",
journal = "Occupational and Environmental Medicine",
issn = "1351-0711",
publisher = "BMJ Publishing Group",
number = "10",

}

TY - JOUR

T1 - Smoothing in occupational cohort studies

T2 - An illustration based on penalised splines

AU - Eisen, E. A.

AU - Agalliu, Ilir

AU - Thurston, S. W.

AU - Coull, B. A.

AU - Checkoway, H.

PY - 2004/10

Y1 - 2004/10

N2 - Aims: To illustrate the contribution of smoothing methods to modelling exposure-response data, Cox models with penalised splines were used to reanalyse lung cancer risk in a cohort of workers exposed to silica in California's diatomaceous earth industry. To encourage application of this approach, computer code is provided. Methods: Relying on graphic plots of hazard ratios as smooth functions of exposure, the sensitivity of the curve to amount of smoothing, length of the exposure lag, and the influence of the highest exposures was evaluated. Trimming and data transformations were used to down-weight influential observations. Results: The estimated hazard ratio increased steeply with cumulative silica exposure before flattening and then declining over the sparser regions of exposure. The curve was sensitive to changes in degrees of freedom, but insensitive to the number or location of knots. As the length of lag increased, so did the maximum hazard ratio, but the shape was similar. Deleting the two highest exposed subjects eliminated the top half of the range and allowed the hazard ratio to continue to rise. The shape of the splines suggested a parametric model with log hazard as a linear function of log transformed exposure would fit well. Conclusions: This flexible statistical approach reduces the dependence on a priori assumptions, while pointing to a suitable parametric model if one exists. In the absence of an appropriate parametric form, however, splines can provide exposure-response information useful for aetiological research and public health intervention.

AB - Aims: To illustrate the contribution of smoothing methods to modelling exposure-response data, Cox models with penalised splines were used to reanalyse lung cancer risk in a cohort of workers exposed to silica in California's diatomaceous earth industry. To encourage application of this approach, computer code is provided. Methods: Relying on graphic plots of hazard ratios as smooth functions of exposure, the sensitivity of the curve to amount of smoothing, length of the exposure lag, and the influence of the highest exposures was evaluated. Trimming and data transformations were used to down-weight influential observations. Results: The estimated hazard ratio increased steeply with cumulative silica exposure before flattening and then declining over the sparser regions of exposure. The curve was sensitive to changes in degrees of freedom, but insensitive to the number or location of knots. As the length of lag increased, so did the maximum hazard ratio, but the shape was similar. Deleting the two highest exposed subjects eliminated the top half of the range and allowed the hazard ratio to continue to rise. The shape of the splines suggested a parametric model with log hazard as a linear function of log transformed exposure would fit well. Conclusions: This flexible statistical approach reduces the dependence on a priori assumptions, while pointing to a suitable parametric model if one exists. In the absence of an appropriate parametric form, however, splines can provide exposure-response information useful for aetiological research and public health intervention.

UR - http://www.scopus.com/inward/record.url?scp=5044226832&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=5044226832&partnerID=8YFLogxK

U2 - 10.1136/oem.2004.013136

DO - 10.1136/oem.2004.013136

M3 - Article

VL - 61

SP - 854

EP - 860

JO - Occupational and Environmental Medicine

JF - Occupational and Environmental Medicine

SN - 1351-0711

IS - 10

ER -