Relationship between morning hematocrit and percentage of patient-days with bleeding outcomes. (A) Unadjusted percentages of patient-days (95% CIs) with grade ≥2A bleeding. (B) ORs (95% CIs) for grade ≥2A bleeding compared with the reference category of hematocrit >29%, taking into account within-person correlation. The 2 df test for any association between morning hematocrit category and grade ≥2A bleeding had P = .002. (C) Unadjusted percentages of patient-days (95% CIs) with grade ≥3 bleeding. (D) ORs (95% CIs) for grade ≥3 bleeding compared with the reference category of hematocrit >29%, taking into account within-person correlation. The 2 df test for any association between morning hematocrit category and grade ≥3 bleeding had P < .001.