A new, accurate predictive model for incident hypertension

Henry Völzke; Glenn Fung; Till Ittermann; Shipeng Yu; Sebastian E Baumeister; Marcus Dörr; Wolfgang Lieb; Uwe Völker; Allan Linneberg; Torben Jørgensen; Stephan B Felix; Rainer Rettig; Bharat Rao; Heyo K Kroemer

doi:10.1097/hjh.0b013e328364a16d

A new, accurate predictive model for incident hypertension

Henry Völzke, Glenn Fung, Till Ittermann, Shipeng Yu, Sebastian E Baumeister, Marcus Dörr, Wolfgang Lieb, Uwe Völker, Allan Linneberg, Torben Jørgensen, Stephan B Felix, Rainer Rettig, Bharat Rao, Heyo K Kroemer

15 Citations (Scopus)

Abstract

OBJECTIVE:: Data mining represents an alternative approach to identify new predictors of multifactorial diseases. This work aimed at building an accurate predictive model for incident hypertension using data mining procedures. METHODS:: The primary study population consisted of 1605 normotensive individuals aged 20-79 years with 5-year follow-up from the population-based study, that is the Study of Health in Pomerania (SHIP). The initial set was randomly split into a training and a testing set. We used a probabilistic graphical model applying a Bayesian network to create a predictive model for incident hypertension and compared the predictive performance with the established Framingham risk score for hypertension. Finally, the model was validated in 2887 participants from INTER99, a Danish community-based intervention study. RESULTS:: In the training set of SHIP data, the Bayesian network used a small subset of relevant baseline features including age, mean arterial pressure, rs16998073, serum glucose and urinary albumin concentrations. Furthermore, we detected relevant interactions between age and serum glucose as well as between rs16998073 and urinary albumin concentrations [area under the receiver operating characteristic (AUC 0.76)]. The model was confirmed in the SHIP validation set (AUC 0.78) and externally replicated in INTER99 (AUC 0.77). Compared to the established Framingham risk score for hypertension, the predictive performance of the new model was similar in the SHIP validation set and moderately better in INTER99. CONCLUSION:: Data mining procedures identified a predictive model for incident hypertension, which included innovative and easy-to-measure variables. The findings promise great applicability in screening settings and clinical practice.

Original language	English
Journal	Journal of Hypertension
Volume	31
Issue number	11
Pages (from-to)	2142-2150
Number of pages	9
ISSN	0263-6352
DOIs	https://doi.org/10.1097/hjh.0b013e328364a16d
Publication status	Published - Nov 2013

Access to Document

10.1097/hjh.0b013e328364a16d

Cite this

@article{9d4730b4dbd9423d8bfa4faddb9506c0,

title = "A new, accurate predictive model for incident hypertension",

abstract = "OBJECTIVE:: Data mining represents an alternative approach to identify new predictors of multifactorial diseases. This work aimed at building an accurate predictive model for incident hypertension using data mining procedures. METHODS:: The primary study population consisted of 1605 normotensive individuals aged 20-79 years with 5-year follow-up from the population-based study, that is the Study of Health in Pomerania (SHIP). The initial set was randomly split into a training and a testing set. We used a probabilistic graphical model applying a Bayesian network to create a predictive model for incident hypertension and compared the predictive performance with the established Framingham risk score for hypertension. Finally, the model was validated in 2887 participants from INTER99, a Danish community-based intervention study. RESULTS:: In the training set of SHIP data, the Bayesian network used a small subset of relevant baseline features including age, mean arterial pressure, rs16998073, serum glucose and urinary albumin concentrations. Furthermore, we detected relevant interactions between age and serum glucose as well as between rs16998073 and urinary albumin concentrations [area under the receiver operating characteristic (AUC 0.76)]. The model was confirmed in the SHIP validation set (AUC 0.78) and externally replicated in INTER99 (AUC 0.77). Compared to the established Framingham risk score for hypertension, the predictive performance of the new model was similar in the SHIP validation set and moderately better in INTER99. CONCLUSION:: Data mining procedures identified a predictive model for incident hypertension, which included innovative and easy-to-measure variables. The findings promise great applicability in screening settings and clinical practice.",

author = "Henry V{\"o}lzke and Glenn Fung and Till Ittermann and Shipeng Yu and Baumeister, {Sebastian E} and Marcus D{\"o}rr and Wolfgang Lieb and Uwe V{\"o}lker and Allan Linneberg and Torben J{\o}rgensen and Felix, {Stephan B} and Rainer Rettig and Bharat Rao and Kroemer, {Heyo K}",

year = "2013",

month = nov,

doi = "10.1097/hjh.0b013e328364a16d",

language = "English",

volume = "31",

pages = "2142--2150",

journal = "Journal of Hypertension",

issn = "0263-6352",

publisher = "Lippincott Williams & Wilkins, Ltd.",

number = "11",

}

TY - JOUR

T1 - A new, accurate predictive model for incident hypertension

AU - Völzke, Henry

AU - Fung, Glenn

AU - Ittermann, Till

AU - Yu, Shipeng

AU - Baumeister, Sebastian E

AU - Dörr, Marcus

AU - Lieb, Wolfgang

AU - Völker, Uwe

AU - Linneberg, Allan

AU - Jørgensen, Torben

AU - Felix, Stephan B

AU - Rettig, Rainer

AU - Rao, Bharat

AU - Kroemer, Heyo K

PY - 2013/11

Y1 - 2013/11

N2 - OBJECTIVE:: Data mining represents an alternative approach to identify new predictors of multifactorial diseases. This work aimed at building an accurate predictive model for incident hypertension using data mining procedures. METHODS:: The primary study population consisted of 1605 normotensive individuals aged 20-79 years with 5-year follow-up from the population-based study, that is the Study of Health in Pomerania (SHIP). The initial set was randomly split into a training and a testing set. We used a probabilistic graphical model applying a Bayesian network to create a predictive model for incident hypertension and compared the predictive performance with the established Framingham risk score for hypertension. Finally, the model was validated in 2887 participants from INTER99, a Danish community-based intervention study. RESULTS:: In the training set of SHIP data, the Bayesian network used a small subset of relevant baseline features including age, mean arterial pressure, rs16998073, serum glucose and urinary albumin concentrations. Furthermore, we detected relevant interactions between age and serum glucose as well as between rs16998073 and urinary albumin concentrations [area under the receiver operating characteristic (AUC 0.76)]. The model was confirmed in the SHIP validation set (AUC 0.78) and externally replicated in INTER99 (AUC 0.77). Compared to the established Framingham risk score for hypertension, the predictive performance of the new model was similar in the SHIP validation set and moderately better in INTER99. CONCLUSION:: Data mining procedures identified a predictive model for incident hypertension, which included innovative and easy-to-measure variables. The findings promise great applicability in screening settings and clinical practice.

AB - OBJECTIVE:: Data mining represents an alternative approach to identify new predictors of multifactorial diseases. This work aimed at building an accurate predictive model for incident hypertension using data mining procedures. METHODS:: The primary study population consisted of 1605 normotensive individuals aged 20-79 years with 5-year follow-up from the population-based study, that is the Study of Health in Pomerania (SHIP). The initial set was randomly split into a training and a testing set. We used a probabilistic graphical model applying a Bayesian network to create a predictive model for incident hypertension and compared the predictive performance with the established Framingham risk score for hypertension. Finally, the model was validated in 2887 participants from INTER99, a Danish community-based intervention study. RESULTS:: In the training set of SHIP data, the Bayesian network used a small subset of relevant baseline features including age, mean arterial pressure, rs16998073, serum glucose and urinary albumin concentrations. Furthermore, we detected relevant interactions between age and serum glucose as well as between rs16998073 and urinary albumin concentrations [area under the receiver operating characteristic (AUC 0.76)]. The model was confirmed in the SHIP validation set (AUC 0.78) and externally replicated in INTER99 (AUC 0.77). Compared to the established Framingham risk score for hypertension, the predictive performance of the new model was similar in the SHIP validation set and moderately better in INTER99. CONCLUSION:: Data mining procedures identified a predictive model for incident hypertension, which included innovative and easy-to-measure variables. The findings promise great applicability in screening settings and clinical practice.

U2 - 10.1097/hjh.0b013e328364a16d

DO - 10.1097/hjh.0b013e328364a16d

M3 - Journal article

C2 - 24077244

SN - 0263-6352

VL - 31

SP - 2142

EP - 2150

JO - Journal of Hypertension

JF - Journal of Hypertension

IS - 11

ER -

A new, accurate predictive model for incident hypertension

Abstract

Access to Document

Fingerprint

Cite this