Table 3.

Factors associated with clinical trial accrual in randomised and non-randomised studies supporting oncology US Food and Drug Administration approvals, Poisson regression results

Variable	Study cohorts
	Randomised			Non-randomised*
	Rate ratio	95% CI	P value	Rate ratio	95% CI	P value
Blinding
Less than double blind	Reference	Reference	Reference	Not included
Double blind	1.21	0.97 to 1.50	0.07
Endpoint
Response rate	Reference	Reference	Reference	Reference	Reference	Reference
PFS	0.82	0.62 to 1.08	0.15	1.97	0.57 to 9.03	0.32
OS	0.94	0.68 to 1.29	0.68	2.51	0.71 to 19.14	0.25
Crossover
No	Reference	Reference	Reference	Reference	Reference	Reference
Yes	1.18	0.96 to 1.45	0.1	1.69	0.35 to 14.67	0.55
Phase
Phase 1/2	Reference	Reference	Reference	Reference	Reference	Reference
Phase 3	2.13	1.48 to 2.99	<0.01	0.94	0.43 to 2.32	0.89
Line
Third line or beyond	Reference	Reference	Reference	Reference	Reference	Reference
First and second line	1.55	1.15 to 2.04	0.01	1.09	0.62 to 1.84	0.75
Single agent response rate	1.09	0.74 to 1.64	0.64	0.7	0.25 to 1.98	0.54
Sites	1	0.99 to 1.00	0.52	1	0.99 to 1.01	0.11
Manufacturer
Other	Reference	Reference	Reference	Reference	Reference	Reference
Pfizer	0.76	0.53 to 1.07	0.12	0.18	0.05 to 0.75	0.01
Roche	1.02	0.76 to 1.37	0.91	1.13	0.45 to 3.49	0.8
Novartis	0.9	0.59 to 1.43	0.63	1.22	0.57 to 2.90	0.61
Merck	1.47	1.18 to 2.37	0.03	2.03	1.10 to 3.92	0.03
Disease burden†	0.99	0.99 to 1.00	0.13	1	0.99 to 1.00	0.12
Randomisation ratio
Equal	Reference	Reference	Reference	Not included
Skewed	0.87	0.69 to 1.09	0.21

*Certain variables were excluded in the regression model of non-randomised studies if they were deemed collinear.

†Disease burden: number of deaths in 1 year per tumour type using data from Surveillance, Epidemiology and End Results.¹⁶

OS, overall survival; PFS, progression-free survival.