+ Site Statistics
+ Search Articles
+ PDF Full Text Service
How our service works
Request PDF Full Text
+ Follow Us
Follow on Facebook
Follow on Twitter
Follow on LinkedIn
+ Subscribe to Site Feeds
Most Shared
PDF Full Text
+ Translate
+ Recently Requested

Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors

Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors

Statistics in Medicine 15(4): 361-387

Multivariable regression models are powerful tools that are used frequently in studies of clinical outcomes. These models can use a mixture of categorical and continuous variables and can handle partially observed (censored) responses. However, uncritical application of modelling techniques can result in models that poorly fit the dataset at hand, or, even more likely, inaccurately predict outcomes on new subjects. One must know how to measure qualities of a model's fit in order to avoid poorly fitted or overfitted models. Measurement of predictive accuracy can be difficult for survival time data in the presence of censoring. We discuss an easily interpretable index of predictive discrimination as well as methods for assessing calibration of predicted survival probabilities. Both types of predictive accuracy should be unbiasedly validated using bootstrapping or cross-validation, before using predictions in a new data series. We discuss some of the hazards of poorly fitted and overfitted regression models and present one modelling strategy that avoids many of the problems discussed. The methods described are applicable to all regression models, but are particularly needed for binary, ordinal, and time-to-event outcomes. Methods are illustrated with a survival analysis in prostate cancer using Cox regression.

Please choose payment method:

(PDF emailed within 0-6 h: $19.90)

Accession: 009063781

Download citation: RISBibTeXText

PMID: 8668867

DOI: 10.1002/(sici)1097-0258(19960229)15:4<361::aid-sim168>3.0.co;2-4

Related references

Models for combining random and systematic errors. assumptions and consequences for different models. Clinical Chemistry and Laboratory Medicine 39(7): 589-595, 2001

Developing Useful Multivariable Models. Journal of the American College of Surgeons 222(1): 102-103, 2016

Subspace Algorithms for the Identification of Multivariable Dynamic Errors-in-Variables Models. Automatica 33(10): 1857-1869, 1997

On stability issues in deriving multivariable regression models. Biometrical Journal. Biometrische Zeitschrift 57(4): 531-555, 2016

Evaluating the assumptions of population projection models used for conservation. Biological Conservation 237: 145-154, 2019

Effect of errors in assumptions underlying genetic models on components of variance and covariance. Egyptian Journal of Genetics & Cytology 2(2): 370, 1973

Systematic review of multivariable prognostic models for mild traumatic brain injury. Journal of Neurotrauma 32(8): 517-526, 2016

Validation and Recalibration of Two Multivariable Prognostic Models for Survival and Independence in Acute Stroke. Plos One 11(5): E0153527, 2017

Contribution of Patient-reported Symptoms Before Palliative Radiotherapy to Development of Multivariable Prognostic Models. Anticancer Research 38(3): 1705-1709, 2018

Agricultural risk analysis: adequacy of models, data, and issues. American Journal of Agricultural Economics 85(5): 1249-1256, 2003

Discrimination-based sample size calculations for multivariable prognostic models for time-to-event data. Bmc Medical Research Methodology 15: 82, 2016

Tubeless procedure is most important factor in reducing length of hospitalization after percutaneous nephrolithotomy: results of univariable and multivariable models. Urology 77(2): 299-304, 2011

Tubeless Procedure is Most Important Factor in Reducing Length of Hospitalization After Percutaneous Nephrolithotomy: Results of Univariable and Multivariable Models. Yearbook of Urology 2011: 5-6, 2011

Camera traps and mark-resight models: The value of ancillary data for evaluating assumptions. Journal of Wildlife Management 79(7): 1163-1172, 2015

Evaluating the Adequacy of Molecular Clock Models Using Posterior Predictive Simulations. Molecular Biology and Evolution 32(11): 2986-2995, 2016