Table 9 Variables selected by LASSO regression for predicting postoperative.

From: Retrospective analysis of open reduction and locking plate fixation in three and four part proximal humeral fractures with efficacy and complications

Variable

Coefficient (Ī²)

Direction

Retained

Age

1.624

+

Yes

Diabetes mellitus

0.098

+

Yes

Osteoporosis

0.266

+

Yes

Formal rehabilitation

āˆ’ 0.563

āˆ’

Yes

Self-rehabilitation

0.163

+

Yes

Fracture healing time

1.384

+

Yes

Rotator cuff repair

āˆ’ 0.809

āˆ’

Yes

Calcar screw

āˆ’ 0.956

āˆ’

Yes

Medial cortical bone defect

0.310

+

Yes

Reduction quality

āˆ’ 0.475

āˆ’

Yes

  1. Direction: ā€œ+ā€ indicates a positive association with postoperative complications, ā€œāˆ’ā€ indicates a negative association.
  2. Variables excluded (coefficients shrunk to zero): Smoking, gender, postoperative drainage, age group, time from injury to surgery, intraoperative blood loss, operation time, salvage rehabilitation.
  3. Coefficients were standardized, and variables were selected via 10-fold cross-validation (optimal Ī» at Log(Ī») = āˆ’ 4).
Back to article page