The impact of different imputation methods on estimates and model performance: an example using a risk prediction model for premature mortality.
Imputation methods
Missing data
Perforamance measures
Population health
Prediction model
Prediction models
Journal
Population health metrics
ISSN: 1478-7954
Titre abrégé: Popul Health Metr
Pays: England
ID NLM: 101178411
Informations de publication
Date de publication:
17 Jun 2024
17 Jun 2024
Historique:
received:
19
01
2024
accepted:
06
06
2024
medline:
18
6
2024
pubmed:
18
6
2024
entrez:
17
6
2024
Statut:
epublish
Résumé
To compare how different imputation methods affect the estimates and performance of a prediction model for premature mortality. Sex-specific Weibull accelerated failure time survival models were run on four separate datasets using complete case, mode, single and multiple imputation to impute missing values. Six performance measures were compared to access predictive accuracy (Nagelkerke R The highest proportion of missingness for a single variable was 10.86% for the female model and 8.24% for the male model. Comparing the performance measures for complete case, mode, single and multiple imputation: the Nagelkerke R In the scenarios examined in this study, mode imputation performed well when using a population health survey compared to single and multiple imputation when predictive performance measures is the main model goal. To generate unbiased hazard ratios, multiple imputation methods were superior. This study shows the need to consider the best imputation approach for a predictive model development given the conditions of missing data and the goals of the analysis.
Identifiants
pubmed: 38886744
doi: 10.1186/s12963-024-00331-3
pii: 10.1186/s12963-024-00331-3
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
13Informations de copyright
© 2024. The Author(s).
Références
Kang H. The prevention and handling of the missing data. Korean J Anesthesiol. 2013;64(5):402–6. https://doi.org/10.4097/kjae.2013.64.5.402 .
doi: 10.4097/kjae.2013.64.5.402
pubmed: 23741561
pmcid: 3668100
Newman DA. Missing Data: five practical guidelines. Organizational Res Methods. 2014;17(4):372–411. https://doi.org/10.1177/1094428114548590 .
doi: 10.1177/1094428114548590
Moons KGM, Royston P, Vergouwe Y, Grobbee DE, Altman DG. Prognosis and prognostic research: what, why, and how? BMJ. 2009;338:b375. https://doi.org/10.1136/bmj.b375 .
doi: 10.1136/bmj.b375
pubmed: 19237405
Manuel DG, Rosella LC, Hennessy D, et al. Predictive risk algorithms in a population setting: an overview. J Epidemiol Community Health. 2012;66:859–65.
doi: 10.1136/jech-2012-200971
pubmed: 22859516
Nijman SWJ, Leeuwenberg AM, Beekers I, Verkouter I, Jacobs J, Bots ML, et al. Missing data is poorly handled and reported in prediction model studies using machine learning: a literature review. J Clin Epidemiol. 2022;142:218–229.
Collins GS, Reitsma JB, Altman DG, Moons KG. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD). Ann Intern Med. 2015;162(10):735–6.
doi: 10.7326/L15-5093-2
pubmed: 25984857
Collins GS, et al. A systematic review finds prediction models for chronic kidney disease were poorly reported and often developed using inappropriate methods. J Clin Epidemiol. 2013;66(3):268–77.
doi: 10.1016/j.jclinepi.2012.06.020
pubmed: 23116690
Tsvetanova A, et al. Missing data was handled inconsistently in UK prediction models: a review of method used. J Clin Epidemiol. 2021;140:149–58.
doi: 10.1016/j.jclinepi.2021.09.008
pubmed: 34520847
Karahalios A, et al. A review of the reporting and handling of missing data in cohort studies with repeated assessment of exposure measures. BMC Med Res Methodol. 2012;12(1):1–10.
doi: 10.1186/1471-2288-12-96
O’Neill M, Hurst M, Palagan L, Diemert L, Kornas K et al. Development and validation of a Population based risk algorithm for premature mortality: The Premature Mortality Population Risk Tool (PreMPoRT).
Rosella LC, O’Neill M, Fisher S, Hurst M, Diemert L, Kornas K, et al. A study protocol for a predictive algorithm to assess population-based premature mortality risk: premature Mortality Population Risk Tool (PreMPoRT). Diagn Progn Res. 2020;4(1):18.
doi: 10.1186/s41512-020-00086-z
pubmed: 33292834
pmcid: 7640636
van Buuren S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat Methods Med Res. 2007;16(3):219–42.
doi: 10.1177/0962280206074463
pubmed: 17621469
Beland Y. Canadian community health survey–methodological overview. Health Rep. 2002;13(3):9–14.
pubmed: 12743956
Statistics Canada. Canadian Vital Death Statistics Database (CVSD) linked to Discharge Abstract Database (DAD) and National Ambulatory Care Reporting System (NACRS). [ https://www.statcan.gc.ca/en/microdata/data-centres/data/cvsd-nacrs ].
Canadian Institute for Health Information. Health Indicators e-Publication. [ https://www.cihi.ca/en/health-indicators-e-publication ].
Statistics Canada. Canadian Community Health Survey (CCHS) Household weights documentation. [ https://www23.statcan.gc.ca/imdb-bmdi/pub/document/3226_D57_T9_V1-eng.htm ].
Thomas S, Wannell B. Combining cycles of the Canadian Community Health Survey. Health Rep. 2009;20(1):53–8.
pubmed: 19388369
Harrell. Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis. Springer; 2001.
Daniel RM, Kenward MG, Cousens SN, De Stavola BL. Using causal diagrams to guide analysis in missing data problems. Stat Methods Med Res. 2012;21(3):243–56. https://doi.org/10.1177/0962280210394469 .
doi: 10.1177/0962280210394469
pubmed: 21389091