Fig. 4
From: Quantifying causal effects from observed data using quasi-intervention
Relationship between baseline NLR and OS under different type, cancer stages (a, g), cancer type (b, f), age (c, h), sex (d, i), and KPS score (e). Among them, a–d is a confounding factor analysis, and e–i is an intermediate value analysis. Due to the lack of data, we used SPSS to fill in the missing data, but for the categories with fewer data (the second stage of cancer), we adopted the omission and merge method