*2.6. Statistical Analysis*

Due to the nature of this retrospective cohort study without a control group, all data are presented using standard descriptive statistics. Due to the small sample size, we applied the Wilcoxon rank-sum test to compare the clinical function of tuberosity healing, the presence of the ER lag sign and the difference in outcome after fracture pattern types I and IV. Analysis of variance (ANOVA) testing was applied to investigate any clinical differences among the various fracture types. Due to the small and nonrepresentational sample sizes for fracture types II and III (*n* = 3 each), we have used ANOVA testing just for investigation of tendencies. In addition, we also compared the outcomes of the "classic" type I fracture to those with comminution (type IV) to achieve a more reliable statistical analysis.

All statistical analyses were completed using SPSS 27 (IBM, Armonk, NY, USA) and the significance level was set to 0.05.
