🔍 Key Findings

• Cook system had good reliability across all cases (ICC = 0.848), even when complications, sequelae, and failure-to-cure were included.
• aCD system had excellent reliability when excluding sequelae/failure-to-cure cases (ICC = 0.975) but only moderate reliability when including them (ICC = 0.620).
• Majority of complications (60–63%) were graded as Grade 3 in the aCD system, corresponding to surgical or anesthetic intervention.
• Cook system classified most cases (78–81%) as major complications.
• Assessors had difficulty distinguishing sequelae from Grade 1 complications, and failure-to-cure from Grades 1–2, reducing aCD reliability.
• The aCD system offers more resolution (5 grades vs. 3) and less subjectivity through objective definitions.
• Novel terminology (e.g., sequelae, failure to cure) in the aCD system may hinder its uptake without proper training.
• Reclassifying expected events (e.g., swelling, bruising) as sequelae could significantly reduce reported complication rates in other studies.