Investigating surgical risk in coronary artery disease: Utilizing ordinal regression for predictive modeling

In this study, an ordinal regression model with a Negative Log-log link function was used to predict the likelihood of having multiple blocked arteries based on various clinical and lifestyle factors. This analysis is essential for identifying individuals at higher risk of severe cardiovascular conditions, enabling earlier and more effective interventions. The model showed a good fit with the observed data, as evidenced by a Pearson Chi-Square value of 244.927 (p = 0.224) and a Deviance value of 167.846 (p = 0.999). Additionally, the model yielded high Pseudo R-square values: Cox and Snell = 0.710 and Nagelkerke = 0.723, indicating a strong explanatory power for the outcome variable. Significant predictors identified include blood urea, low-density lipoprotein (LDL) cholesterol levels, hemoglobin (HGB), red cell distribution width (RDW), and smoking status (smoker: yes). Notably, smoking increases the odds of having multiple blocked arteries by a factor of 3.27 (OR = 3.2704, p = 0.028), while higher RDW is also associated with increased risk (OR = 1.5945, p = 0.0048). In contrast, higher blood urea levels are linked to a reduced risk (OR = 0.9594, p = 0.0045). These findings highlight the value of statistical modeling in identifying cardiovascular risk factors, helping healthcare providers make more informed decisions regarding patient treatment and preventive care.