A simulation study on estimating biomarker–treatment interaction effects in randomized trials with prognostic variables

Haller, Bernhard; Ulm, Kurt

doi:10.1186/s13063-018-2491-0

Trials

Table 4 Proportions of rejected null hypotheses and numbers of included covariates for scenarios with K=36

From: A simulation study on estimating biomarker–treatment interaction effects in randomized trials with prognostic variables

K	Σ	β _k	Censoring	Main	True	AIC_A	AIC_B	Significance	Full
No interaction
36	Σ ₁	β _eq	Low	0.047	0.065	0.080	0.054	0.056	0.065
36	Σ ₁	β _v	Low	0.059	0.067	0.084	0.066	0.069	0.073
36	Σ ₂	β _eq	Low	0.077	0.053	0.073	0.051	0.053	0.053
36	Σ ₂	β _v	Low	0.075	0.054	0.074	0.052	0.056	0.056
36	Σ ₃	β _eq	Low	0.067	0.060	0.083	0.057	0.064	0.060
36	Σ ₃	β _v	Low	0.063	0.055	0.074	0.053	0.069	0.069
36	Σ ₁	β _eq	High	0.052	0.071	0.086	0.059	0.055	0.071
36	Σ ₁	β _v	High	0.047	0.063	0.080	0.058	0.050	0.066
36	Σ ₂	β _eq	High	0.085	0.080	0.082	0.057	0.080	0.080
36	Σ ₂	β _v	High	0.085	0.064	0.086	0.054	0.070	0.070
36	Σ ₃	β _eq	High	0.057	0.069	0.094	0.056	0.063	0.069
36	Σ ₃	β _v	High	0.075	0.063	0.081	0.057	0.076	0.071
Quantitative interaction
36	Σ ₁	β _eq	Low	0.103	0.150	0.165	0.130	0.138	0.150
36	Σ ₁	β _v	Low	0.095	0.131	0.150	0.120	0.125	0.136
36	Σ ₂	β _eq	Low	0.121	0.120	0.141	0.109	0.120	0.120
36	Σ ₂	β _v	Low	0.128	0.128	0.147	0.121	0.134	0.134
36	Σ ₃	β _eq	Low	0.115	0.121	0.142	0.103	0.122	0.121
36	Σ ₃	β _v	Low	0.119	0.143	0.172	0.125	0.140	0.143
36	Σ ₁	β _eq	High	0.081	0.108	0.133	0.098	0.094	0.108
36	Σ ₁	β _v	High	0.095	0.112	0.132	0.101	0.102	0.118
36	Σ ₂	β _eq	High	0.100	0.085	0.103	0.072	0.085	0.085
36	Σ ₂	β _v	High	0.092	0.109	0.127	0.093	0.118	0.118
36	Σ ₃	β _eq	High	0.091	0.104	0.134	0.099	0.105	0.104
36	Σ ₃	β _v	High	0.097	0.093	0.132	0.093	0.102	0.115
Qualitative interaction
36	Σ ₁	β _eq	Low	0.551	0.652	0.657	0.603	0.558	0.652
36	Σ ₁	β _v	Low	0.517	0.700	0.688	0.658	0.599	0.669
36	Σ ₂	β _eq	Low	0.280	0.570	0.582	0.518	0.570	0.570
36	Σ ₂	β _v	Low	0.266	0.555	0.575	0.517	0.542	0.542
36	Σ ₃	β _eq	Low	0.408	0.609	0.637	0.582	0.581	0.609
36	Σ ₃	β _v	Low	0.451	0.623	0.637	0.592	0.605	0.620
36	Σ ₁	β _eq	High	0.389	0.447	0.456	0.403	0.385	0.447
36	Σ ₁	β _v	High	0.390	0.472	0.486	0.437	0.418	0.453
36	Σ ₂	β _eq	High	0.219	0.368	0.411	0.334	0.368	0.368
36	Σ ₂	β _v	High	0.228	0.385	0.419	0.353	0.389	0.389
36	Σ ₃	β _eq	High	0.282	0.364	0.396	0.328	0.352	0.364
36	Σ ₃	β _v	High	0.303	0.402	0.416	0.350	0.382	0.391
Mean number of prognostic covariates included
		β _eq	Low	0	36	19.6	19.9	24.5	36
		β _v	Low	0	24	19.0	19.4	25.2	36
		β _eq	High	0	36	15.2	15.7	23.1	36
		β _v	High	0	24	16.0	16.4	23.9	36

For the scenarios with no true biomarker–treatment interaction, results for methods/strategies with an observed type I error probability above 7% are in italics. For scenarios with a true biomarker–treatment interaction, the observed power is in bold if the type I error probability did not exceed 7%

Back to article page

ISSN: 1745-6215

Contact us

Submission enquiries: Access here and click Contact Us
General enquiries: info@biomedcentral.com