How to develop a more accurate risk prediction model when there are few events

Abstract

When the number of events is low relative to the number of predictors, standard regression could produce overfitted risk models that make inaccurate predictions. Use of penalised regression may improve the accuracy of risk prediction

Prognostic Value of Right Ventricular Dysfunction in Heart Failure With Reduced Ejection Fraction

“…Penalized regression is a flexible shrinkage approach that is effective when the number of events per variable is low (<10). 22 Because lasso coefficients are good for prediction but are not straightforward, multivariate Cox proportional hazard model is also presented. Two multivariable models were constructed based on inclusion of RVFWS either as a continuous variable or as binary categorical variable.…”

Validation of the American College of Surgeons Risk Calculator for preoperative risk stratification

“…Similarly, the events rate for surgical site infection, urinary tract infection and venous thromboembolism were very low (n <10), contributing to the low power of the study for these outcomes. When the number of actual events is smaller than the number of predictors in a risk model, overfitting can ensue, resulting in the underestimation of the event among those labelled as low risk and overestimation among those labelled as high risk 7. Hence, definitive conclusions regarding the predictive ability of the ACS NSQIP Surgical Risk Calculator for renal failure, return to OR, surgical site infection, urinary that infection and venous thromboembolism could not be made and could only be addressed by increasing the sample size and corresponding events rate of these outcomes.…”

Assessing risk in pulmonary arterial hypertension: what we know, what we don't

“…Statistical models are often used to predict the probability that an individual with a given set of risk factors will experience a health outcome, usually termed an "event" [6]. In PAH, it is widely agreed that a multivariate risk prediction model that is validated serially represents the best predictor of morbidity and mortality, especially compared to measures assessed individually or at a single time point.…”

“…If widely adopted, appropriate risk prediction provides an opportunity to learn about various risk phenotypes in PAH, enhance consistency of treatment approaches across practitioners and assist in the timely referral for lung transplantation. Lastly, risk model-derived equations can enhance clinical study design both by selecting the appropriate study cohort and serving as a study end-point.Statistical models are often used to predict the probability that an individual with a given set of risk factors will experience a health outcome, usually termed an "event" [6]. In PAH, it is widely agreed that a…”

“…Statistical models are often used to predict the probability that an individual with a given set of risk factors will experience a health outcome, usually termed an "event" [6]. In PAH, it is widely agreed that a…”