Background: Since diagnosis of mood disorder heavily depends on signs and symptoms, emerging researches have been studying biomarkers with the attempt to improve diagnostic accuracy, but none of the findings have been broadly accepted. The purpose of the present study was to construct a preliminary diagnostic model to distinguish major depressive disorder (MDD) and bipolar disorder (BD) using potential commonly tested blood biomarkers.
Methods: Information of 721 inpatients with an ICD-10 diagnosis of MDD or BD were collected from the electronic medical record system. Variables in the nomogram were selected by best subset selection method after a prior univariable screening, and then constructed using logistic regression with inclusion of the psychotropic medication use. The discrimination, calibration and internal validation of the nomogram were evaluated by the receiver operating characteristic curve (ROC), the calibration curve, cross validation and subset validation method.
Results: The nomogram consisted of five variables, including age, eosinophil count, plasma concentrations of prolactin, total cholesterol, and low-density lipoprotein cholesterol. The model could discriminate between MDD and BD with an area under the ROC curve (AUC) of 0.858, with a sensitivity of 0.716 and a specificity of 0.890.
Conclusion: The comprehensive nomogram constructed by the present study can be convenient to distinguish MDD and BD since the incorporating variables were common indicators in clinical practice. It could help avoid misdiagnoses and improve prognosis of the patients.
Keywords: Bipolar disorder; Diagnosis; Major depressive disorder; Nomogram; Validation.
© 2023. BioMed Central Ltd., part of Springer Nature.