At-risk set: 1831 patients, 196 events

Missing ASA in the at-risk set: 17 (0.9%)

AKI events among patients with missing ASA: 4

Reproduced the published asa_ge3 column from raw ASA: rule confirmed as missing -> ASA <3.



                     specification           term    n  events   HR  ci_low  ci_high      p

 A as published (missing = ASA <3)    lac_2.1-3.0 1831     196 1.16    0.79     1.69 0.4431

 A as published (missing = ASA <3)    lac_3.8-5.9 1831     196 1.48    1.04     2.10 0.0278

 A as published (missing = ASA <3) lac_8.6 capped 1831     196 2.13    1.22     3.74 0.0080

 A as published (missing = ASA <3)  lactate_score 1831     196 1.25    1.08     1.45 0.0034

B complete case (missing excluded)    lac_2.1-3.0 1814     192 1.19    0.81     1.73 0.3801

B complete case (missing excluded)    lac_3.8-5.9 1814     192 1.50    1.05     2.13 0.0258

B complete case (missing excluded) lac_8.6 capped 1814     192 2.07    1.16     3.68 0.0132

B complete case (missing excluded)  lactate_score 1814     192 1.25    1.07     1.45 0.0043

               C missing indicator    lac_2.1-3.0 1831     196 1.15    0.79     1.68 0.4556

               C missing indicator    lac_3.8-5.9 1831     196 1.47    1.04     2.09 0.0307

               C missing indicator lac_8.6 capped 1831     196 2.13    1.22     3.73 0.0082

               C missing indicator  lactate_score 1831     196 1.25    1.07     1.45 0.0038



Top stratum hazard ratio across specifications: 2.13 (1.22-3.74), 2.07 (1.16-3.68), 2.13 (1.22-3.73)

Per-stratum trend across specifications: 1.25 (1.08-1.45), 1.25 (1.07-1.45), 1.25 (1.07-1.45)



Largest difference in the top-stratum hazard ratio: 0.06

The published handling is reportable if this is small relative to the width of the confidence interval; otherwise the primary model must be refitted under a stated rule.



Outputs: <local path>\analysis\results\manuscript_v3

