Evaluating prediction models in reproductive medicine

S. F P J Coppus, F. Van Der Veen, B. C. Opmeer, B. W J Mol, P. M M Bossuyt

Research output: Contribution to journalComment/debate

37 Citations (Scopus)

Abstract

Prediction models are used in reproductive medicine to calculate the probability of pregnancy without treatment, as well as the probability of pregnancy after ovulation induction, intrauterine insemination or in vitro fertilization. The performance of such prediction models is often evaluated with a receiver operating characteristic (ROC) curve. The area under the ROC curve, also known as c-statistic, is then used as a measure of model performance. The value of this c-statistic is low for most prediction models in reproductive medicine. Here, we demonstrate that low values of the c-statistic are to be expected in these prediction models, but we also show that this does not imply that these models are of limited use in clinical practice. The calibration of the model (the correspondence between model-based probabilities and observed pregnancy rates) as well as the availability of a clinically useful distribution of probabilities and the ability to correctly identify the appropriate form of management are more meaningful concepts for model evaluation.

LanguageEnglish
Pages1774-1778
Number of pages5
JournalHuman Reproduction
Volume24
Issue number8
DOIs
Publication statusPublished - 1 Jan 2009

Keywords

  • Fertility
  • IUI
  • IVF
  • Prediction model
  • Spontaneous pregnancy

ASJC Scopus subject areas

  • Reproductive Medicine
  • Obstetrics and Gynaecology

Cite this

Coppus, S. F. P. J., Van Der Veen, F., Opmeer, B. C., Mol, B. W. J., & Bossuyt, P. M. M. (2009). Evaluating prediction models in reproductive medicine. Human Reproduction, 24(8), 1774-1778. https://doi.org/10.1093/humrep/dep109
Coppus, S. F P J ; Van Der Veen, F. ; Opmeer, B. C. ; Mol, B. W J ; Bossuyt, P. M M. / Evaluating prediction models in reproductive medicine. In: Human Reproduction. 2009 ; Vol. 24, No. 8. pp. 1774-1778.
@article{eacd6009cd15494eadf65c97ef6214c3,
title = "Evaluating prediction models in reproductive medicine",
abstract = "Prediction models are used in reproductive medicine to calculate the probability of pregnancy without treatment, as well as the probability of pregnancy after ovulation induction, intrauterine insemination or in vitro fertilization. The performance of such prediction models is often evaluated with a receiver operating characteristic (ROC) curve. The area under the ROC curve, also known as c-statistic, is then used as a measure of model performance. The value of this c-statistic is low for most prediction models in reproductive medicine. Here, we demonstrate that low values of the c-statistic are to be expected in these prediction models, but we also show that this does not imply that these models are of limited use in clinical practice. The calibration of the model (the correspondence between model-based probabilities and observed pregnancy rates) as well as the availability of a clinically useful distribution of probabilities and the ability to correctly identify the appropriate form of management are more meaningful concepts for model evaluation.",
keywords = "Fertility, IUI, IVF, Prediction model, Spontaneous pregnancy",
author = "Coppus, {S. F P J} and {Van Der Veen}, F. and Opmeer, {B. C.} and Mol, {B. W J} and Bossuyt, {P. M M}",
year = "2009",
month = "1",
day = "1",
doi = "10.1093/humrep/dep109",
language = "English",
volume = "24",
pages = "1774--1778",
journal = "Human Reproduction",
issn = "0268-1161",
publisher = "Oxford University Press",
number = "8",

}

Coppus, SFPJ, Van Der Veen, F, Opmeer, BC, Mol, BWJ & Bossuyt, PMM 2009, 'Evaluating prediction models in reproductive medicine', Human Reproduction, vol. 24, no. 8, pp. 1774-1778. https://doi.org/10.1093/humrep/dep109

Evaluating prediction models in reproductive medicine. / Coppus, S. F P J; Van Der Veen, F.; Opmeer, B. C.; Mol, B. W J; Bossuyt, P. M M.

In: Human Reproduction, Vol. 24, No. 8, 01.01.2009, p. 1774-1778.

Research output: Contribution to journalComment/debate

TY - JOUR

T1 - Evaluating prediction models in reproductive medicine

AU - Coppus, S. F P J

AU - Van Der Veen, F.

AU - Opmeer, B. C.

AU - Mol, B. W J

AU - Bossuyt, P. M M

PY - 2009/1/1

Y1 - 2009/1/1

N2 - Prediction models are used in reproductive medicine to calculate the probability of pregnancy without treatment, as well as the probability of pregnancy after ovulation induction, intrauterine insemination or in vitro fertilization. The performance of such prediction models is often evaluated with a receiver operating characteristic (ROC) curve. The area under the ROC curve, also known as c-statistic, is then used as a measure of model performance. The value of this c-statistic is low for most prediction models in reproductive medicine. Here, we demonstrate that low values of the c-statistic are to be expected in these prediction models, but we also show that this does not imply that these models are of limited use in clinical practice. The calibration of the model (the correspondence between model-based probabilities and observed pregnancy rates) as well as the availability of a clinically useful distribution of probabilities and the ability to correctly identify the appropriate form of management are more meaningful concepts for model evaluation.

AB - Prediction models are used in reproductive medicine to calculate the probability of pregnancy without treatment, as well as the probability of pregnancy after ovulation induction, intrauterine insemination or in vitro fertilization. The performance of such prediction models is often evaluated with a receiver operating characteristic (ROC) curve. The area under the ROC curve, also known as c-statistic, is then used as a measure of model performance. The value of this c-statistic is low for most prediction models in reproductive medicine. Here, we demonstrate that low values of the c-statistic are to be expected in these prediction models, but we also show that this does not imply that these models are of limited use in clinical practice. The calibration of the model (the correspondence between model-based probabilities and observed pregnancy rates) as well as the availability of a clinically useful distribution of probabilities and the ability to correctly identify the appropriate form of management are more meaningful concepts for model evaluation.

KW - Fertility

KW - IUI

KW - IVF

KW - Prediction model

KW - Spontaneous pregnancy

UR - http://www.scopus.com/inward/record.url?scp=67651087306&partnerID=8YFLogxK

U2 - 10.1093/humrep/dep109

DO - 10.1093/humrep/dep109

M3 - Comment/debate

VL - 24

SP - 1774

EP - 1778

JO - Human Reproduction

T2 - Human Reproduction

JF - Human Reproduction

SN - 0268-1161

IS - 8

ER -

Coppus SFPJ, Van Der Veen F, Opmeer BC, Mol BWJ, Bossuyt PMM. Evaluating prediction models in reproductive medicine. Human Reproduction. 2009 Jan 1;24(8):1774-1778. https://doi.org/10.1093/humrep/dep109