State of the art in selection of variables and functional forms in multivariable analysis—outstanding issues

Abstract

How to select variables and identify functional forms for continuous variables is a key concern when creating a multivariable model. Ad hoc 'traditional' approaches to variable selection have been in use for at least 50 years. Similarly, methods for determining functional forms for continuous variables were first suggested many years ago. More recently, many alternative approaches to address these two challenges have been proposed, but knowledge of their properties and meaningful comparisons between them are scarce. To define a state-of-the-art and to provide evidence-supported guidance to researchers who have only a basic level of statistical knowledge many outstanding issues in multivariable modelling remain. Our main aims are to identify and illustrate such gaps in the literature and present them at a moderate technical level to the wide community of practitioners, researchers and students of statistics.We briefly discuss general issues in building descriptive regression models, strategies for variable selection, different ways of choosing functional forms for continuous variables, and methods for combining the selection of variables and functions. We discuss two examples, taken from the medical literature, to illustrate problems in the practice of modelling.Our overview revealed that there is not yet enough evidence on which to base recommendations for the selection of variables and functional forms in multivariable analysis. Such evidence may come from comparisons between alternative methods. In particular, we highlight seven important topics that require further investigation and make suggestions for the direction of further research.

Pre-Operative Evaluation of DNA Methylation Profile in Oral Squamous Cell Carcinoma Can Predict Tumor Aggressive Potential

“…This process results in sparse fits similar to lasso-like approaches, with many of the estimated coefficients shrunk toward zero. Boosting algorithms have been shown to be particularly useful to handle models in which the number of candidate predictors exceeds the number of observations (high-dimensional settings) [51,52].…”

Factors affecting attitudes towards caring for terminally ill patients among nursing students in Switzerland: a cross-sectional study

“…Because this last analysis was intended as a descriptive model rather than a predictive model, variables weakly associated with the FATCOD, Form B were not removed. 24 Missing values were handled by multiple imputation: instead of being replaced by a single value, missing values are replaced by several values selected at random from a distribution determined using a model (15 imputations for this study). Statistical analyses were performed using R V.3.6.1 ( www.r-project.org ), and the mice package V.3.5.0 was used for the imputation.…”

Canine helminthiases and associated risk factors in Kigali city, Rwanda

“…A decision on which variable to include in a multivariate analysis can depend on different criteria, or cutoff (determinants screening signi cance) than the signi cance level [19,20,21]. In addition, some variables can adjust the effects of other variables in the model, even if they do not have any low p-value themselves [22]. In the present study, all the potential risk factors were considered for multivariable analysis presented in Table 3 to detect the direction and extent to which variable categories explain the difference in prevalence of canine helminthiases.…”