Table S1.

Assumptions for instrumental variable analysis

	ICP monitoring	Intracranial operation	Primary referral
Assumption 1: instrument must be strongly associated with the intervention itself
Partial F statistic	22.4	7.0	65.9
Assumption 2: substitute variable is not associated with any of the prognostic factors
Spearman’s Rho correlation with Psurvival	−0.09	0.17	0.06
Spearman’s Rho correlation with Pfavorable outcome	−0.18	0.14	0.07

Note: P_survival = the probability of six-month survival; P_{favorable outcome} = the probability of six-month favorable outcome (GOS >3).

Abbreviation: ICP, intracranial pressure.