A randomised controlled feasibility trial to evaluate local heat preconditioning on wound healing after reconstructive breast surgery: the preHEAT trial

Table 5 Performance matrix for necrosis outcomes

Criteria	Outcome measure
Criteria	Necrosis yes/no	Depth (SKIN score)	Total necrosis area
Subjectivity of measurement (κ [95% CI] for the primary and secondary clinical assessor*)	0.84 [0.62 to 1.00]	0.79 [0.58 to 0.99]	0.57 [0.21 to 0.94]
Total sample size required to demonstrate statistically significant difference between treatment and control group (based on observed data)^†	1096	1556	2866
Proportion of patients with observed response**	95.0%	89.4%	88.7%

*For total necrosis area, we include κ where area is assumed to be 0 mm² when no necrosis present is recorded. ^†For necrosis yes/no, the sample size was determined for a two-sample proportion test of 26% (heated) versus 35% (control). For total area, the sample size was computed using non-parametric methods for non-normally distributed continuous data. For necrosis, depth sample size calculation for ordered categorical data was performed using the observed proportions in each category (71, 18, 11 and 0% versus 78, 14, 6 and 2%). All sample size calculations use a 5% level of significance and 90% power. **For total necrosis area, area is assumed to be 0 mm² where no necrosis present is recorded

ISSN: 2055-5784