Table 3.

Univariate and multivariable analyses for the risk of dying from acute lymphoblastic leukaemia for patients aged 15–17 years in the Netherlands between 1990 and 2015.

		Univariate analysis				Multivariable analysis, 1st model				Multivariable analysis, 2nd model
	N	HR	95% CI		p value	HRa	95% CI		p value	HRa	95% CI		p value
Period
1990–94	40	Ref.				Ref.				Ref.
1995–99	43	0.9	0.5	1.8	0.85	0.9	0.5	1.7	0.73	1.0	0.5	1.9	0.94
2000–04	41	0.6	0.3	1.3	0.23	0.6	0.3	1.3	0.18	0.7	0.3	1.4	0.30
2005–09	54	0.5	0.2	1.0	0.04	0.4	0.2	0.9	0.03	0.6	0.3	1.4	0.25
2010–15	68	0.5	0.3	1.0	0.06	0.5	0.3	1.0	0.04	0.8	0.4	1.7	0.56
Sex
Male	169	Ref.				Ref.				Ref.
Female	77	1.2	0.7	1.9	0.48	1.2	0.8	2.0	0.41	1.5	0.9	2.4	0.14
Immunophenotype
BCP-ALL	179	Ref.				Ref.				Ref.
T-cell ALL	67	1.5	0.9	2.4	0.12	1.6	1.0	2.6	0.06	1.6	1.0	2.6	0.07
Site of treatment
Outside paediatric oncology centre	83	Ref.								Ref.
Paediatric oncology centre	163	0.3	0.2	0.5	<0.01					0.3	0.2	0.5	<0.01

In the first multivariable model we did not consider site of treatment, and this model shows significantly lower risk of death in recent periods of diagnosis compared to the reference period 1990–1994. In the second multivariable model we added site of treatment which results in disappearance of the discriminative effect of period of diagnosis and a significantly lower risk of death for patients treated in a paediatric oncology centre.

BCP-ALL B-cell precursor acute lymphoblastic leukaemia, HR hazard ratio, CI confidence interval.

^aIn the multivariable analysis, each covariate is simultaneously adjusted for all other covariates, and follow-up time. Hazard ratios represent risk of death within 5 years from diagnosis compared to the reference category.