Statistical Modelling of Comorbidity Effect on Second Cancer

Abstract

Second cancer is a new cancer that occurs in someone who had history of cancer. The incidence of cancer is on the increase in the global scene and especially in Sub-Saharan Africa. This phenomenon constitutes huge health problems especially with comorbidity effects with other health conditions which have made the diagnosis and treatment of cancer patients a complex issue. Hence, this study determined the comorbidity effect on second cancer based on a retrospective study of 474 patients attending University of Abuja Teaching Hospital (UATH). These patients were first treated of cancer, and after one year developed another cancer or cancer free. The results revealed that the incidence of second cancer was approximately 39.5% and 41.4% of the patients with one or more other disease(s) had second cancer. Adjusted and unadjusted odds ratio from logistic regression showed that patients with history of smoking were 3.58 times more likely to develop second cancer when no adjustment was made to the model while the risk increased after adjustment. Furthermore, cormobid patients are 1.56 times more likely to develop second cancer than cancer patients without other diseases. Based on the area under the receiver characteristics curve, logistic regression model effectively distinguished between the two groups of cancer patients. Comorbidity and smoking were identified as significant factors on the incidence of second cancer among cancer patients attending UATH. Therefore, emphasis should be given to formulating policies on controlling tobacco smoking and to create health awareness on the effect of clinical factors on second cancer within the study area.
 Keywords: Comorbidity, Cancer, Statistical modelling, Sensitivity, Specificity, Odd Ratio