Retrospective Study of Prognostic Prediction Based Only on Objective Indicators in End-of-life Patients: A Study Using Biological Prognostic Score Version-2

ABSTRACT

Objective: We retrospectively studied the adaptation and limits of applying Biological Prognostic Score Version-2 composed only of blood test results to prognostic prediction in end-of-life non-cancer patients. Methods: The prognostic score was calculated from the cholinesterase, blood urea nitrogen, and white blood cell counts of hospitalized end-of-life non-cancer patients, divided into three groups with cutoff values, and prediction accuracy analysis, survival analysis, and simple regression analysis were performed. Results: Diagnostic accuracy of 204 non-cancer patients at the same cut-off value and predicted survival time as cancer was 79% accurate at 3 weeks survival and 63% at 9 weeks. Specificity and negative predictive value were highly accurate, sensitivity and positive predictive value were low. In the survival analysis, the discrimination between the 3 groups was significant (p<0.05), but the regression coefficient in the regression analysis was not significant (p=0.43). Conclusion: The prognostic prediction using this score for non-cancer patients has high prediction accuracy in the case of good prognosis. It is suggested that clinical use of this score is also possible if used cautiously.