Literature DB >> 33549087

An empirical analysis of dealing with patients who are lost to follow-up when developing prognostic models using a cohort design.

Jenna M Reps1, Peter Rijnbeek2, Alana Cuthbert3, Patrick B Ryan4, Nicole Pratt5, Martijn Schuemie4.   

Abstract

BACKGROUND: Researchers developing prediction models are faced with numerous design choices that may impact model performance. One key decision is how to include patients who are lost to follow-up. In this paper we perform a large-scale empirical evaluation investigating the impact of this decision. In addition, we aim to provide guidelines for how to deal with loss to follow-up.
METHODS: We generate a partially synthetic dataset with complete follow-up and simulate loss to follow-up based either on random selection or on selection based on comorbidity. In addition to our synthetic data study we investigate 21 real-world data prediction problems. We compare four simple strategies for developing models when using a cohort design that encounters loss to follow-up. Three strategies employ a binary classifier with data that: (1) include all patients (including those lost to follow-up), (2) exclude all patients lost to follow-up or (3) only exclude patients lost to follow-up who do not have the outcome before being lost to follow-up. The fourth strategy uses a survival model with data that include all patients. We empirically evaluate the discrimination and calibration performance.
RESULTS: The partially synthetic data study results show that excluding patients who are lost to follow-up can introduce bias when loss to follow-up is common and does not occur at random. However, when loss to follow-up was completely at random, the choice of addressing it had negligible impact on model discrimination performance. Our empirical real-world data results showed that the four design choices investigated to deal with loss to follow-up resulted in comparable performance when the time-at-risk was 1-year but demonstrated differential bias when we looked into 3-year time-at-risk. Removing patients who are lost to follow-up before experiencing the outcome but keeping patients who are lost to follow-up after the outcome can bias a model and should be avoided.
CONCLUSION: Based on this study we therefore recommend (1) developing models using data that includes patients that are lost to follow-up and (2) evaluate the discrimination and calibration of models twice: on a test set including patients lost to follow-up and a test set excluding patients lost to follow-up.

Entities:  

Keywords:  Best practices; Censoring; Loss to follow-up; Model development; PatientLevelPrediction; Prognostic model

Mesh:

Year:  2021        PMID: 33549087      PMCID: PMC7866757          DOI: 10.1186/s12911-021-01408-x

Source DB:  PubMed          Journal:  BMC Med Inform Decis Mak        ISSN: 1472-6947            Impact factor:   2.796


  19 in total

1.  Lipid modification: cardiovascular risk assessment and the modification of blood lipids for the primary and secondary prevention of cardiovascular disease.

Authors:  J Robson
Journal:  Heart       Date:  2008-08-13       Impact factor: 5.994

2.  Adapting machine learning techniques to censored time-to-event health record data: A general-purpose approach using inverse probability of censoring weighting.

Authors:  David M Vock; Julian Wolfson; Sunayan Bandyopadhyay; Gediminas Adomavicius; Paul E Johnson; Gabriela Vazquez-Benitez; Patrick J O'Connor
Journal:  J Biomed Inform       Date:  2016-03-16       Impact factor: 6.317

3.  Development of pediatric comorbidity prediction model.

Authors:  Derek Tai; Paul Dick; Teresa To; James G Wright
Journal:  Arch Pediatr Adolesc Med       Date:  2006-03

4.  Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD Statement.

Authors:  Gary S Collins; Johannes B Reitsma; Douglas G Altman; Karel G M Moons
Journal:  BMC Med       Date:  2015-01-06       Impact factor: 8.775

5.  When and how should multiple imputation be used for handling missing data in randomised clinical trials - a practical guide with flowcharts.

Authors:  Janus Christian Jakobsen; Christian Gluud; Jørn Wetterslev; Per Winkel
Journal:  BMC Med Res Methodol       Date:  2017-12-06       Impact factor: 4.615

6.  Multiple imputation for handling missing outcome data when estimating the relative risk.

Authors:  Thomas R Sullivan; Katherine J Lee; Philip Ryan; Amy B Salter
Journal:  BMC Med Res Methodol       Date:  2017-09-06       Impact factor: 4.615

Review 7.  Prognosis Research Strategy (PROGRESS) 3: prognostic model research.

Authors:  Ewout W Steyerberg; Karel G M Moons; Danielle A van der Windt; Jill A Hayden; Pablo Perel; Sara Schroter; Richard D Riley; Harry Hemingway; Douglas G Altman
Journal:  PLoS Med       Date:  2013-02-05       Impact factor: 11.069

8.  Risk prediction model for heart failure and cardiomyopathy after adjuvant trastuzumab therapy for breast cancer.

Authors:  Ghideon Ezaz; Jessica B Long; Cary P Gross; Jersey Chen
Journal:  J Am Heart Assoc       Date:  2014-02-28       Impact factor: 5.501

Review 9.  How to Establish Clinical Prediction Models.

Authors:  Yong Ho Lee; Heejung Bang; Dae Jung Kim
Journal:  Endocrinol Metab (Seoul)       Date:  2016-03

10.  Design and implementation of a standardized framework to generate and evaluate patient-level prediction models using observational healthcare data.

Authors:  Jenna M Reps; Martijn J Schuemie; Marc A Suchard; Patrick B Ryan; Peter R Rijnbeek
Journal:  J Am Med Inform Assoc       Date:  2018-08-01       Impact factor: 4.497

View more
  1 in total

1.  Dynamic Digital Twin: Diagnosis, Treatment, Prediction, and Prevention of Disease During the Life Course.

Authors:  Skander Tahar Mulder; Amir-Houshang Omidvari; Régine Steegers-Theunissen; Anja J Rueten-Budde; Pei-Hua Huang; Ki-Hun Kim; Babette Bais; Melek Rousian; Rihan Hai; Can Akgun; Jeanine Roeters van Lennep; Sten Willemsen; Peter R Rijnbeek; David Mj Tax; Marcel Reinders; Eric Boersma; Dimitris Rizopoulos; Valentijn Visch
Journal:  J Med Internet Res       Date:  2022-09-14       Impact factor: 7.076

  1 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.