Factors affecting clinical outcomes after massive rotator cuff repair using a data mining technique

A data mining technique using decision tree analysis is a method of discovery of new knowledge and patterns from big data. This study aimed to clarify by decision tree analysis the interaction among preoperative and postoperative prognostic factors that affect patient satisfaction after massive rotator cuff repairs (RCR). The study evaluated 1367 shoulders (right 956, left 411) with massive rotator cuff tears. Exclusion criteria were fracture, dislocation, re-tear, nerve paralysis, and insufficient follow-up. These patients were divided into two groups: satisfaction group and dissatisfaction group. We compared these groups with eight preoperative factors and two surgical factors using a decision tree analysis. Our results revealed three prognostic factors of patient satisfaction after massive RCR: tendon covering rate, preoperative ROM of flexion, and MMT. In clinical practice, decision tree analysis can be useful for understanding the interactions among predictors of patient satisfaction after massive RCR.