Table 1.

Reasons given for choice of trial arm

		Phase I (N = 510)			Phase 2 (N = 589)
Correct arm	Reason given	Correct	Incorrect	Unsure	Correct	Incorrect	Unsure
	1. Factors relating to initial presentation	10%	14%	1%	6%	24%	0%
	a. Asymptomatic, low PSA test result, localised disease, early presentation	10	2	0	6	3	0
	b. Symptomatic, high PSA test result, locally advanced/metastatic disease, late presentation	0	12	1	0	21	0
	2. Study-specific factors	8%	30%	4%	3%	22%	2%
Intervention	a. No formal prostate cancer diagnosis	0	3	0	0	2	0
arm	b. PSA test not performed	0	21	0	0	11	0
	c. Timing of PSA test, investigations or treatments	4	5	2	1	2	1
	d. Presence of co-morbidities or other cancers	0	0	1	0	4	1
	e. Investigations, treatments received	3	0	0	1	2	0
	f. Other study-specific factors	1	1	1	1	1	0
	3. No reason stated	3%	1%	25%	0%	1%	42%
	(Total for each phase = 100%)	21%	46%	33%	9%	46%	45%
	1. Factors relating to initial presentation	14%	1%	1%	23%	3%	3%
	a. Asymptomatic, low PSA test result, localised disease, early presentation	1	1	1	0	3	1
	b. Symptomatic, high PSA test result, locally advanced/metastatic disease, late presentation	13	0	0	23	0	2
	2. Study-specific factors	50%	1%	4%	20%	2%	1%
Control	a. No formal prostate cancer diagnosis	8	0	1	1	0	0
arm	b. PSA test not performed	31	0	1	10	0	0
	c. Timing of PSA test, investigations or treatments	6	1	1	2	1	0
	d. Presence of co-morbidities or other cancers	1	0	0	3	0	0
	e. Investigations, treatments received	2	0	0	1	1	1
	f. Other study-specific factors	2	0	1	3	0	0
	3. No reason stated	4%	0%	25%	3%	0%	44%
	(Total for each phase = 100%)	68%	2%	30%	46%	6%	48%

Table 1 shows the reasons reviewers gave for their choice of trial arm, for correct, incorrect and unsure categories; for both phases. Highlighted in bold are the total percentages. Figures have been rounded so may not add up to 100%.