The Wally plot approach to assess the calibration of clinical prediction models

Research output: Contribution to journal › Journal article › Research › peer-review

Standard

The Wally plot approach to assess the calibration of clinical prediction models. / Blanche, Paul; Gerds, Thomas A; Ekstrøm, Claus T.

In: Lifetime Data Analysis, Vol. 25, No. 1, 15.01.2019, p. 150-167.

Research output: Contribution to journal › Journal article › Research › peer-review

Harvard

Blanche, P, Gerds, TA & Ekstrøm, CT 2019, 'The Wally plot approach to assess the calibration of clinical prediction models', Lifetime Data Analysis, vol. 25, no. 1, pp. 150-167. https://doi.org/10.1007/s10985-017-9414-3

APA

Blanche, P., Gerds, T. A., & Ekstrøm, C. T. (2019). The Wally plot approach to assess the calibration of clinical prediction models. Lifetime Data Analysis, 25(1), 150-167. https://doi.org/10.1007/s10985-017-9414-3

Vancouver

Blanche P, Gerds TA, Ekstrøm CT. The Wally plot approach to assess the calibration of clinical prediction models. Lifetime Data Analysis. 2019 Jan 15;25(1):150-167. https://doi.org/10.1007/s10985-017-9414-3

Author

Blanche, Paul ; Gerds, Thomas A ; Ekstrøm, Claus T. / The Wally plot approach to assess the calibration of clinical prediction models. In: Lifetime Data Analysis. 2019 ; Vol. 25, No. 1. pp. 150-167.

Bibtex

@article{023852d0d6274dac989e4bd7393e9fee,
title = "The Wally plot approach to assess the calibration of clinical prediction models",
abstract = "A prediction model is calibrated if, roughly, for any percentage x we can expect that x subjects out of 100 experience the event among all subjects that have a predicted risk of x%. Typically, the calibration assumption is assessed graphically but in practice it is often challenging to judge whether a {"}disappointing{"} calibration plot is the consequence of a departure from the calibration assumption, or alternatively just {"}bad luck{"} due to sampling variability. We propose a graphical approach which enables the visualization of how much a calibration plot agrees with the calibration assumption to address this issue. The approach is mainly based on the idea of generating new plots which mimic the available data under the calibration assumption. The method handles the common non-trivial situations in which the data contain censored observations and occurrences of competing events. This is done by building on ideas from constrained non-parametric maximum likelihood estimation methods. Two examples from large cohort data illustrate our proposal. The 'wally' R package is provided to make the methodology easily usable.",
author = "Paul Blanche and Gerds, {Thomas A} and Ekstr{\o}m, {Claus T}",
year = "2019",
month = jan,
day = "15",
doi = "10.1007/s10985-017-9414-3",
language = "English",
volume = "25",
pages = "150--167",
journal = "Lifetime Data Analysis",
issn = "1380-7870",
publisher = "Springer",
number = "1",

}

RIS

TY - JOUR

T1 - The Wally plot approach to assess the calibration of clinical prediction models

AU - Blanche, Paul

AU - Gerds, Thomas A

AU - Ekstrøm, Claus T

PY - 2019/1/15

Y1 - 2019/1/15

N2 - A prediction model is calibrated if, roughly, for any percentage x we can expect that x subjects out of 100 experience the event among all subjects that have a predicted risk of x%. Typically, the calibration assumption is assessed graphically but in practice it is often challenging to judge whether a "disappointing" calibration plot is the consequence of a departure from the calibration assumption, or alternatively just "bad luck" due to sampling variability. We propose a graphical approach which enables the visualization of how much a calibration plot agrees with the calibration assumption to address this issue. The approach is mainly based on the idea of generating new plots which mimic the available data under the calibration assumption. The method handles the common non-trivial situations in which the data contain censored observations and occurrences of competing events. This is done by building on ideas from constrained non-parametric maximum likelihood estimation methods. Two examples from large cohort data illustrate our proposal. The 'wally' R package is provided to make the methodology easily usable.

AB - A prediction model is calibrated if, roughly, for any percentage x we can expect that x subjects out of 100 experience the event among all subjects that have a predicted risk of x%. Typically, the calibration assumption is assessed graphically but in practice it is often challenging to judge whether a "disappointing" calibration plot is the consequence of a departure from the calibration assumption, or alternatively just "bad luck" due to sampling variability. We propose a graphical approach which enables the visualization of how much a calibration plot agrees with the calibration assumption to address this issue. The approach is mainly based on the idea of generating new plots which mimic the available data under the calibration assumption. The method handles the common non-trivial situations in which the data contain censored observations and occurrences of competing events. This is done by building on ideas from constrained non-parametric maximum likelihood estimation methods. Two examples from large cohort data illustrate our proposal. The 'wally' R package is provided to make the methodology easily usable.

U2 - 10.1007/s10985-017-9414-3

DO - 10.1007/s10985-017-9414-3

M3 - Journal article

C2 - 29214550

VL - 25

SP - 150

EP - 167

JO - Lifetime Data Analysis

JF - Lifetime Data Analysis

SN - 1380-7870

IS - 1

ER -

ID: 198525425