Table 3.

Effect of endoscopic procedure for each outcome tested by inverse probability of treatment weighting methods

Outcome	B	Standard error	p value	Beta	95% confidence interval for beta
Bleeding amount	−241.2	68.5	<0.001*	1.00 × 10−13	1.00 × 10−13−1.01 × 10−13
Operative time	−28.8	12.5	0.02*	4.10 × 10−13	1.00 × 10−13−0.013
Stay time in the operation room	−62.8	15.1	<0.001*	1.00 × 10−13	1.00 × 10−13−1.04 × 10−13
GCS score 7 days postop	−0.383	1.80	0.83	0.68	0.02–23.10
mRS at 6 months	−0.076	0.72	0.92	0.93	0.23–3.82

Weights are based on results from a treatment selection model, estimated using logistic regression with endoscopy or craniotomy as the dependent variable and the baseline characteristics of age, sex, history of cardiovascular disease, use of antithrombotic drugs, GCS score on admission, hematoma volume, midline shift, preoperative APTT, PT-INR, and platelet count as independent variables. The weights for each patient were calculated as the inverse of the probability of receiving the treatment. After weighting, we made generalized estimating equations to assess the association between the surgical procedure and each outcome. As to the probability of receiving the treatment, the p value was 0.646 by the Hosmer–Lemeshow test, and the c-statistic was 0.923.

p <0.05 by weighted generalized estimating equations.

APTT: activated partial thromboplastin time, GCS: Glasgow Coma Scale, mRS: modified Rankin Scale, PT-INR: prothrombin time international normalized ratio