Are non-constant rates and non-proportional treatment effects accounted for in the design and analysis of randomised controlled trials? A review of current practice

Table 1 Reported characteristics of the trials

Reported trial characteristic			N (%)
Sample size calculation approach
	Log rank test		40 (61%)
	Cox model beta coefficient		4 (6%)
	Exponentially distributed survival		4 (6%)
	Simulation		7 (11%)
	Difference in proportions		6 (9%)
	Unclear		5 (6%)
Time-to-event analytical methods^a
	Non-parametric log rank test		58 (88%)
	Cox PH model		64 (97%)
	Parametric regression		7 (11%)
	Landmark analysis		7 (11%)
Proportional hazards (PH) assumption^b
	PH assumption acknowledged		34 (53%)
	PH testing methods documented		31 (48%)
		Analytical test methods	10 (16%)
		Visual assessment methods	6 (9%)
		Visual and analytical methods	7 (11%)
		Unspecified	8 (13%)

ISSN: 1471-2288