Escalation with Overdose Control Using Ordinal Toxicity Grades for Cancer Phase I Clinical Trials

Table 2

Summary statistics for trial safety and efficiency under model misspecification when the true MTD𝛾=0.1. Rows 2–6 give the summary statistics based on all patients from all 1000 trials.

Statistics	Model	𝜌1
		0.2	0.5	0.8
Proportion of DLTs	EWOC-POM	0.337	0.312	0.299
	M1	0.328	0.304	0.289
	M2	0.357	0.342	0.331
Proportion overdosed	EWOC-POM	0.274	0.155	0.200
	M1	0.240	0.203	0.173
	M2	0.404	0.351	0.322
MTD estimate	EWOC-POM	0.093	0.090	0.087
	M1	0.094	0.092	0.090
	M2	0.095	0.090	0.086
Bias	EWOC-POM	−0.007	−0.010	−0.013
	M1	−0.006	−0.007	−0.010
	M2	−0.005	−0.010	−0.013
Root MSE	EWOC-POM	0.022	0.021	0.023
	M1	0.016	0.016	0.017
	M2	0.027	0.029	0.031
Percent of trial with; DLT rate >40%	EWOC-POM	6.6	3.0	2.9
	M1	1.5	1.2	0.8
	M2	14.7	11.6	12.6
Percent of trial with; estimated MTD within 0.05 of γ	EWOC-POM	98.4	97.5	96.4
	M1	99.7	99.5	99.5
	M2	91.1	91.2	88.1
Percent of trial with; estimated MTD within 0.10 of γ	EWOC-POM	100.0	100.0	100.0
	M1	100.0	100.0	100.0
	M2	100.0	100.0	100.0