Table 1.

Percent of iterations for which each updating methods was recommended by our testing procedure under each simulated scenario, training sample size ($n_t$), and updating sample size ($n_u$) for the complex model

		${\mathit{n}}_{\mathit{t}}$= 1000			${\mathit{n}}_{\mathit{t}}$= 5000			${\mathit{n}}_{\mathit{t}}$= 10000
Scenario	Updating method	${\mathit{n}}_{\mathit{u}}$= 1000	${\mathit{n}}_{\mathit{u}}$= 5000	${\mathit{n}}_{\mathit{u}}$= 10 000	${\mathit{n}}_{\mathit{u}}$= 1000	nu= 5000	nu= 10 000	${\mathit{n}}_{\mathit{u}}$= 1000	${\mathit{n}}_{\mathit{u}}$= 5000	${\mathit{n}}_{\mathit{u}}$= 10 000
No population shift	Not updating	94	7.7	0	100	99.7	87	100	100	100
	Intercept correction	1.2	1.3	0	0	0.2	0	0	0	0
	Linear recalibration	4.8	7.1	0	0	0.1	0	0	0	0
	Flexible recalibration	0	0.1	0	0	0	0	0	0	0
	Model refitting	0	83.8	100	0	0	13	0	0	0
Increased event rate	Not updating	43.6	0.1	0	43.4	0	0	48.1	0	0
	Intercept correction	48	5.2	0	56.6	99.2	81.7	51.9	100	100
	Linear recalibration	8.4	4.4	0	0	0.8	5.2	0	0	0
	Flexible recalibration	0	0	0	0	0	0	0	0	0
	Model refitting	0	90.3	100	0	0	13.1	0	0	0
Less variable case mix	Not updating	79.9	22.5	1	100	88.5	68.4	100	96.3	89.4
	Intercept correction	14.1	14.6	0.5	0	10.3	21.1	0	3.6	7
	Linear recalibration	6	20.7	1.6	0	1.2	5.2	0	0.1	3.6
	Flexible recalibration	0	0.1	0	0	0	0	0	0	0
	Model refitting	0	42.1	96.9	0	0	5.3	0	0	0
More variable case mix	Not updating	59	0	0	95.5	60.1	15.8	96.4	74.7	74.9
	Intercept correction	30.4	0	0	4.5	28.4	13.3	3.6	24.7	18
	Linear recalibration	8.2	0.2	0	0	1.7	7.8	0	0.4	3.5
	Flexible recalibration	0	0	0	0	0	0	0	0	0
	Model refitting	2.4	99.8	100	0	9.8	63.1	0	0.2	3.6
Association changes	Not updating	0	0	0	0	0	0	0	0	0
	Intercept correction	0	0	0	3.6	0	0	2.4	0	0
	Linear recalibration	0	0	0	0	0	0	1.1	0	0
	Flexible recalibration	0	0	0	0	0	0	0	0	0
	Model refitting	100	100	100	96.4	100	100	96.5	100	100