Search

Purpose: Severe toxicity is reported in about 30% of gastrointestinal cancer patients receiving 5-Fluorouracil (5-FU)-based chemotherapy. To date, limited tools exist to identify at risk patients in this setting. The objective of this study was to address this need by designing a predictive model using a Bayesian network, a probabilistic graphical model offering robust, explainable predictions. Methods: We utilized a dataset of 267 gastrointestinal cancer patients, conducting preprocessing, and splitting it into TRAIN and TEST sets (80%:20% ratio). The RandomForest algorithm assessed variable importance based on MeanDecreaseGini coefficient. The bnlearn R library helped design a Bayesian network model using a 10-fold cross-validation on the TRAIN set and the aic-cg method for network structure optimization. The model’s performance was gauged based on accuracy, sensitivity, and specificity, using cross-validation on the TRAIN set and independent validation on the TEST set. Results: The model demonstrated satisfactory performance with an average accuracy of 0.85 (±0.05) and 0.80 on TRAIN and TEST datasets, respectively. The sensitivity and specificity were 0.82 (±0.14) and 0.87 (±0.07) for the TRAIN dataset, and 0.71 and 0.83 for the TEST dataset, respectively. A user-friendly tool was developed for clinical implementation. Conclusions: Despite several limitations, our Bayesian network model demonstrated a high level of accuracy in predicting the risk of developing severe haematological toxicity in gastrointestinal cancer patients receiving 5-FU-based chemotherapy. Future research should aim at model validation in larger cohorts of patients and different clinical settings.

Continuous and discrete unobserved heterogeneity have been widely used in modeling discrete choice models. In this dissertation I investigate how these modeling strategies can be used to capture and model spatial heterogeneity or locally varying coefficients for different latent structures. In the first chapter, I outline the main advantages and disadvantages of both continuous and discrete spatial modeling strategies. Then I conduct a simulation experiment in order to understand the ability of both approaches to retrieve the true representation of the spatially varying process under small sample size situations. The results show that the data requirement to achieve lower bias in the continuous case is substantial compared with the discrete case. I have also found that, as the number of individuals per spatial unit increases, both models are able to identify the regional-specific estimates. However, the discrete case is able to retrieve the true spatial heterogeneity surface with lower bias and better coverage. In the second chapter, I show the Rchoice package for R that allows estimating models with individual heterogeneity for both cross-sectional and panel data. In particular, the package allows binary, ordinal and count response, as well as continuous and discrete covariates. This chapter is a general description of Rchoice and all functionalities are illustrated using real databases. The last chapter shows how continuous and discrete spatial heterogeneity models can be applied in order to analyze whether monetary subjective well-being eval- uations vary across space using a cross-sectional dataset from Chile. The results show that focusing just on the average estimates of compensating variations veils useful local variation. Moreover, the discrete approach shows some weak superiority over the continuous case in terms of model fit and interpretation.