Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity

Objective To explore factors that potentially impact external validation performance while developing and validating a prognostic model for hospital admissions (HAs) in complex older general practice patients.Study design and setting Using individual participant data from four cluster-randomised tri...

Full description

Saved in:
Bibliographic Details
Main Authors: Rafael Perera, Marjan Van Den Akker, Kym I E Snell, Joerg J Meerpohl, Ferdinand M Gerlach, Daniela Küllenberg de Gaudry, Ana Isabel González-González, Jeanet Blom, Christiane Muth, Walter Emil Haefeli, Karin M A Swart, Petra Elders, Henrik Rudolf, Ulrich Thiem, Andreas Daniel Meid, Truc Sophia Dinh, Donna Bosch-Lenders, Hans J Trampisch, Benno Flaig, Ghainsom Kom
Format: Article
Language:English
Published: BMJ Publishing Group 2021-08-01
Series:BMJ Open
Online Access:https://bmjopen.bmj.com/content/11/8/e045572.full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850138119252213760
author Rafael Perera
Marjan Van Den Akker
Kym I E Snell
Joerg J Meerpohl
Ferdinand M Gerlach
Daniela Küllenberg de Gaudry
Ana Isabel González-González
Jeanet Blom
Christiane Muth
Walter Emil Haefeli
Karin M A Swart
Petra Elders
Henrik Rudolf
Ulrich Thiem
Andreas Daniel Meid
Truc Sophia Dinh
Donna Bosch-Lenders
Hans J Trampisch
Benno Flaig
Ghainsom Kom
author_facet Rafael Perera
Marjan Van Den Akker
Kym I E Snell
Joerg J Meerpohl
Ferdinand M Gerlach
Daniela Küllenberg de Gaudry
Ana Isabel González-González
Jeanet Blom
Christiane Muth
Walter Emil Haefeli
Karin M A Swart
Petra Elders
Henrik Rudolf
Ulrich Thiem
Andreas Daniel Meid
Truc Sophia Dinh
Donna Bosch-Lenders
Hans J Trampisch
Benno Flaig
Ghainsom Kom
author_sort Rafael Perera
collection DOAJ
description Objective To explore factors that potentially impact external validation performance while developing and validating a prognostic model for hospital admissions (HAs) in complex older general practice patients.Study design and setting Using individual participant data from four cluster-randomised trials conducted in the Netherlands and Germany, we used logistic regression to develop a prognostic model to predict all-cause HAs within a 6-month follow-up period. A stratified intercept was used to account for heterogeneity in baseline risk between the studies. The model was validated both internally and by using internal-external cross-validation (IECV).Results Prior HAs, physical components of the health-related quality of life comorbidity index, and medication-related variables were used in the final model. While achieving moderate discriminatory performance, internal bootstrap validation revealed a pronounced risk of overfitting. The results of the IECV, in which calibration was highly variable even after accounting for between-study heterogeneity, agreed with this finding. Heterogeneity was equally reflected in differing baseline risk, predictor effects and absolute risk predictions.Conclusions Predictor effect heterogeneity and differing baseline risk can explain the limited external performance of HA prediction models. With such drivers known, model adjustments in external validation settings (eg, intercept recalibration, complete updating) can be applied more purposefully.Trial registration number PROSPERO id: CRD42018088129.
format Article
id doaj-art-0014ac625b614f349b8b0d9ca196fab9
institution OA Journals
issn 2044-6055
language English
publishDate 2021-08-01
publisher BMJ Publishing Group
record_format Article
series BMJ Open
spelling doaj-art-0014ac625b614f349b8b0d9ca196fab92025-08-20T02:30:39ZengBMJ Publishing GroupBMJ Open2044-60552021-08-0111810.1136/bmjopen-2020-045572Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validityRafael Perera0Marjan Van Den Akker1Kym I E Snell2Joerg J Meerpohl3Ferdinand M Gerlach4Daniela Küllenberg de Gaudry5Ana Isabel González-González6Jeanet Blom7Christiane Muth8Walter Emil Haefeli9Karin M A Swart10Petra Elders11Henrik Rudolf12Ulrich Thiem13Andreas Daniel Meid14Truc Sophia Dinh15Donna Bosch-Lenders16Hans J Trampisch17Benno Flaig18Ghainsom Kom19Department of Primary Care Health Sciences, University of Oxford, Oxford, UK21 Institute of General Practice, University of Frankfurt, Frankfurt am Main, GermanyInstitute of Applied Health Research, University of Birmingham, Birmingham, UKInstitute for Evidence in Medicine, Medical Center – Faculty of Medicine, University of Freiburg, Freiburg im Breisgau, GermanyInstitute of General Practice, Goethe University Frankfurt, Frankfurt am Main, Germanyresearcher4Network for Research on Chronicity, Primary Care, and Health Promotion (RICAPPS), Tenerife, Spain5 Department of Public Health and Primary Care, Leiden University Medical Center, Leiden, Zuid-Holland, Netherlands20 Department of General Practice and Family Medicine, University Hospital OWL of Bielefeld University Campus Hospital Lippe, Detmold, Germany1 Department of Clinical Pharmacology and Pharmacoepidemiology, University of Heidelberg, Heidelberg, Baden-Württemberg, GermanyPHARMO Institute for Drug Outcomes Research, Utrecht, NetherlandsHealth Behaviours & Chronic Diseases, Amsterdam Public Health Research Institute, Amsterdam, The NetherlandsMedical Informatics, Biometry and Epidemiology, Ruhr-University Bochum, Bochum, GermanyChair of Geriatrics and Gerontology, University Clinic Eppendorf, Hamburg, GermanyDepartment of Clinical Pharmacology & Pharmacoepidemiology, Heidelberg University, Heidelberg, Baden-Württemberg, GermanyInstitute of General Practice, Goethe University Frankfurt, Frankfurt am Main, Hessen, GermanySchool of CAPHRI, Department of Family Medicine, Maastricht University, Maastricht, The NetherlandsDepartment of Medical Informatics, Biometry and Epidemiology, Ruhr University, Bochum, GermanyInstitute of General Practice, Goethe University Frankfurt, Frankfurt am Main, GermanyTechniker Krankenkasse (TK), Hamburg, GermanyObjective To explore factors that potentially impact external validation performance while developing and validating a prognostic model for hospital admissions (HAs) in complex older general practice patients.Study design and setting Using individual participant data from four cluster-randomised trials conducted in the Netherlands and Germany, we used logistic regression to develop a prognostic model to predict all-cause HAs within a 6-month follow-up period. A stratified intercept was used to account for heterogeneity in baseline risk between the studies. The model was validated both internally and by using internal-external cross-validation (IECV).Results Prior HAs, physical components of the health-related quality of life comorbidity index, and medication-related variables were used in the final model. While achieving moderate discriminatory performance, internal bootstrap validation revealed a pronounced risk of overfitting. The results of the IECV, in which calibration was highly variable even after accounting for between-study heterogeneity, agreed with this finding. Heterogeneity was equally reflected in differing baseline risk, predictor effects and absolute risk predictions.Conclusions Predictor effect heterogeneity and differing baseline risk can explain the limited external performance of HA prediction models. With such drivers known, model adjustments in external validation settings (eg, intercept recalibration, complete updating) can be applied more purposefully.Trial registration number PROSPERO id: CRD42018088129.https://bmjopen.bmj.com/content/11/8/e045572.full
spellingShingle Rafael Perera
Marjan Van Den Akker
Kym I E Snell
Joerg J Meerpohl
Ferdinand M Gerlach
Daniela Küllenberg de Gaudry
Ana Isabel González-González
Jeanet Blom
Christiane Muth
Walter Emil Haefeli
Karin M A Swart
Petra Elders
Henrik Rudolf
Ulrich Thiem
Andreas Daniel Meid
Truc Sophia Dinh
Donna Bosch-Lenders
Hans J Trampisch
Benno Flaig
Ghainsom Kom
Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity
BMJ Open
title Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity
title_full Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity
title_fullStr Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity
title_full_unstemmed Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity
title_short Predicting hospital admissions from individual patient data (IPD): an applied example to explore key elements driving external validity
title_sort predicting hospital admissions from individual patient data ipd an applied example to explore key elements driving external validity
url https://bmjopen.bmj.com/content/11/8/e045572.full
work_keys_str_mv AT rafaelperera predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT marjanvandenakker predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT kymiesnell predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT joergjmeerpohl predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT ferdinandmgerlach predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT danielakullenbergdegaudry predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT anaisabelgonzalezgonzalez predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT jeanetblom predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT christianemuth predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT walteremilhaefeli predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT karinmaswart predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT petraelders predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT henrikrudolf predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT ulrichthiem predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT andreasdanielmeid predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT trucsophiadinh predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT donnaboschlenders predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT hansjtrampisch predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT bennoflaig predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity
AT ghainsomkom predictinghospitaladmissionsfromindividualpatientdataipdanappliedexampletoexplorekeyelementsdrivingexternalvalidity