Table 5.

Parameters for different episode 2 non-enrolment scenarios (simulation study 2a). For all scenarios, we set ${\mathrm{\alpha }}^{R_2}=0.05$ and ${\mathrm{\gamma }}^{R_2}=0.10$

Scenario	${\mathrm{\beta }}_{X_{\mathit{PL}}}$	${\mathrm{\beta }}_{X_{\mathit{EL}}}$	${\mathrm{\beta }}_{X_{\mathit{PL}}}^{R_2}$	${\mathrm{\beta }}_{X_{\mathit{EL}}}^{R_2}$	${\mathrm{\delta }}_{\mathit{Xpl}}^{R_2}$	${\mathrm{\delta }}_{\mathit{Xel}}^{R_2}$
Scenario 1 – Non-enrolment depends on previous treatment allocation	0	0	0	0	0	0
Scenario 2 – Non-enrolment depends on previous treatment allocation and previous outcome	10	0	0.25	0	0	0
Scenario 3 – Non-enrolment depends on previous treatment allocation and baseline prognosis at episode 2	0	10	0	0.25	0	0
Scenario 4 – Non-enrolment is differential between treatment groups based on previous outcome	10	0	0	0	0.5	0
Scenario 5 – Non-enrolment is differential between treatment groups based on baseline prognosis at episode 2	0	10	0	0	0	0.5