A Method to Predict Comorbid Conditions Using Risk Factor Profile of Multiple Fluid Biomarkers.

Abstract

The co-occurrence of diabetes, hypertension, and cardiovascular diseases together can make clinical management and treatment more complex. Early detection of comorbid conditions can help in creating personalized treatment plans. Multiple fluid biomarkers can be used to enhance the diagnostic accuracy of identifying comorbidity. This study aims to distinguish non-comorbid and comorbid conditions using the risk factor profile of multiple fluid biomarkers, such as creatine phosphokinase, platelet count, serum creatinine, and ejection fraction. Area feature is computed by utilizing risk factor profile of the biomarkers, and a random forest classifier is used to distinguish the two conditions. The results indicate that the area of the radar plot is more significant for differentiating comorbid from non-comorbid conditions. RF classifier achieves the highest accuracy of 59.91% to differentiate the two conditions. Thus, multiple fluid biomarkers could be used to accurately detect the comorbid condition and improve the treatment plan individually.