Literature DB >> 29174118

Poor performance of clinical prediction models: the harm of commonly applied methods.

Ewout W Steyerberg1, Hajime Uno2, John P A Ioannidis3, Ben van Calster4.   

Abstract

OBJECTIVE: To evaluate limitations of common statistical modeling approaches in deriving clinical prediction models and explore alternative strategies. STUDY DESIGN AND
SETTING: A previously published model predicted the likelihood of having a mutation in germline DNA mismatch repair genes at the time of diagnosis of colorectal cancer. This model was based on a cohort where 38 mutations were found among 870 participants, with validation in an independent cohort with 35 mutations. The modeling strategy included stepwise selection of predictors from a pool of over 37 candidate predictors and dichotomization of continuous predictors. We simulated this strategy in small subsets of a large contemporary cohort (2,051 mutations among 19,866 participants) and made comparisons to other modeling approaches. All models were evaluated according to bias and discriminative ability (concordance index, c) in independent data.
RESULTS: We found over 50% bias for five of six originally selected predictors, unstable model specification, and poor performance at validation (median c = 0.74). A small validation sample hampered stable assessment of performance. Model prespecification based on external knowledge and using continuous predictors led to better performance (c = 0.836 and c = 0.852 with 38 and 2,051 events respectively).
CONCLUSION: Prediction models perform poorly if based on small numbers of events and developed with common but suboptimal statistical approaches. Alternative modeling strategies to best exploit available predictive information need wider implementation, with collaborative research to increase sample sizes.
Copyright © 2017 Elsevier Inc. All rights reserved.

Entities:  

Keywords:  Events per variable; Prediction model; Regression analysis; Sample size; Simulation; Validation

Mesh:

Year:  2017        PMID: 29174118     DOI: 10.1016/j.jclinepi.2017.11.013

Source DB:  PubMed          Journal:  J Clin Epidemiol        ISSN: 0895-4356            Impact factor:   6.437


  28 in total

1.  Age, Motivation, and Emotion Regulation Skills Predict Treatment Outcome in an Internet-Based Self-Help Intervention for COVID-19 Related Psychological Distress.

Authors:  Noemi Anja Brog; Julia Katharina Hegy; Thomas Berger; Hansjörg Znoj
Journal:  Front Public Health       Date:  2022-06-09

2.  Development of a Prognostic Model for Predicting Multiple Sclerosis After Optic Neuritis: A Secondary Analysis of Data From the Optic Neuritis Treatment Trial.

Authors:  Wenjing Luo; Xinlei Deng; Xiaoyu Xu; Ruitong Song; Meifeng Luo; Heather E Moss; Yi Du
Journal:  J Neuroophthalmol       Date:  2021-10-22       Impact factor: 4.415

3.  Predictive models for musculoskeletal injury risk: why statistical approach makes all the difference.

Authors:  Daniel I Rhon; Deydre S Teyhen; Gary S Collins; Garrett S Bullock
Journal:  BMJ Open Sport Exerc Med       Date:  2022-10-14

4.  Validated Prediction Models for Macular Degeneration Progression and Predictors of Visual Acuity Loss Identify High-Risk Individuals.

Authors:  Johanna M Seddon; Bernard Rosner
Journal:  Am J Ophthalmol       Date:  2018-10-31       Impact factor: 5.258

5.  Neural Responses to Heartbeats Detect Residual Signs of Consciousness during Resting State in Postcomatose Patients.

Authors:  Diego Candia-Rivera; Jitka Annen; Olivia Gosseries; Charlotte Martial; Aurore Thibaut; Steven Laureys; Catherine Tallon-Baudry
Journal:  J Neurosci       Date:  2021-03-23       Impact factor: 6.167

6.  COVID-19 Prediction Models Need Robust and Transparent Development.

Authors:  Gary S Collins
Journal:  Disaster Med Public Health Prep       Date:  2021-04-19       Impact factor: 1.385

7.  The updated Charlson comorbidity index is a useful predictor of mortality in patients with Staphylococcus aureus bacteraemia.

Authors:  H G Ternavasio-de la Vega; F Castaño-Romero; S Ragozzino; R Sánchez González; M P Vaquero-Herrero; M Siller-Ruiz; G Spalter-Glicberg; C Ramírez-Baum; S Rodríguez-Rodríguez; J E García-Sánchez; I García-García; M Marcos
Journal:  Epidemiol Infect       Date:  2018-09-03       Impact factor: 4.434

8.  Cardiovascular Disease Prognostic Models in Latin America and the Caribbean: A Systematic Review.

Authors:  Rodrigo M Carrillo-Larco; Carlos Altez-Fernandez; Niels Pacheco-Barrios; Claudia Bambs; Vilma Irazola; J Jaime Miranda; Goodarz Danaei; Pablo Perel
Journal:  Glob Heart       Date:  2019-03

9.  Development of a clinical prediction model for the onset of functional decline in people aged 65-75 years: pooled analysis of four European cohort studies.

Authors:  Nini H Jonkman; Marco Colpo; Jochen Klenk; Chris Todd; Trynke Hoekstra; Vieri Del Panta; Kilian Rapp; Natasja M van Schoor; Stefania Bandinelli; Martijn W Heymans; Dominique Mauger; Luca Cattelani; Michael D Denkinger; Dietrich Rothenbacher; Jorunn L Helbostad; Beatrix Vereijken; Andrea B Maier; Mirjam Pijnappels
Journal:  BMC Geriatr       Date:  2019-06-27       Impact factor: 3.921

10.  Effect of tranexamic acid by baseline risk of death in acute bleeding patients: a meta-analysis of individual patient-level data from 28 333 patients.

Authors:  Francois-Xavier Ageron; Angele Gayet-Ageron; Katharine Ker; Timothy J Coats; Haleema Shakur-Still; Ian Roberts
Journal:  Br J Anaesth       Date:  2020-03-19       Impact factor: 9.166

View more

北京卡尤迪生物科技股份有限公司 © 2022-2023.