1. 'Risk of bias' assessment criteria for observational studies.

	Internal validity	External validity
Study group	Selection bias (representative: yes/no): if it consisted of more than 90% of the original cohort of children with ALL treated with glucocorticoids; or if it was a random sample with respect to treatment.	Reporting bias (well defined: yes/no): if treatment protocol was mentioned; and if (cumulative) dose of glucocorticoid treatment was mentioned; and if type of glucocorticoid treatment was mentioned; and if duration of glucocorticoid treatment was mentioned; and if method of cessation of glucocorticoid treatment was mentioned.
Follow‐up	Attrition bias (adequate: yes/no): if outcome was assessed at end date of the study for 60% to 90% of study group; or if outcome was assessed for more than 90% of the study group but with an unknown end date.	Reporting bias (well defined: yes/no): if length of follow‐up was mentioned; and if frequency of measuring outcomes was mentioned.
Outcome	Detection bias (blinding: yes/no): if outcome assessor was blinded to glucocorticoid treatment.	Reporting bias (well defined: yes/no): if methods of detection were described; and if outcome definition was objective and precise.
Risk estimation	Confounding (adjustment for other factors: yes/no): if important prognostic factors (i.e. age, sex, cotreatment) or follow‐up was taken adequately into account.	Analysis (well defined: yes/no): if risk ratio, odds ratio, attributable risk, linear or logistic regression model, mean difference, or Chi2 statistic was calculated.

ALL: acute lymphoblastic leukaemia.