Table 5.

Prediction model development for Neurocognitive deficits

	Univariable analysis odds ratios (95% CI)	Multivariable model based on statistical significance odds ratios (95% CI)
Gender female (ref: male)	1.089 (0.540 to 2.196), p = 0.813
Age (years)	1.036 (1.008 to 1.064), p = 0.011	1.024 (0.987 to 1.063)
Tumor location, skull base (ref: convexity)	1.072 (0.598 to 1.923), p = 0.816
Tumor size before last intervention (mm)	1.019 (1.000 to 1.039), p = .048	1.022 (0.998 to 1.047)
First resection yes (ref: no)	0.729 (0.299 to 1.777), p = 0.487
First resection complications yes (ref: no)	1.500 (0.805 to 2.794), p = 0.201
Second resection yes (ref: no)	4.574 (1.191 to 17.572), p = 0.027	2.662 (0.488 to 14.528)
Simpson grade first resection IV/V (ref: I–III)	1.121 (0.540 to 2.325), p = 0.760
WHO Grade II (ref: I)	2.148 (0.651 to 7.092), p = 0.210
Radiotherapy yes (ref: no)	2.011 (0.956 to 4.230), p = 0.066	2.819 (0.925 to 8.585)
Hand dominance, right (ref: left)	0.659 (0.289 to 1.505), p = 0.323
Charlson Comorbidity Index	1.135 (0.877 to 1.468), p = 0.336
Educational level (1: primary/secondary, 2: tertiary vocational, 3: academic)	0.412 (0.265 to 0.641), p = 0.000	0.359 (0.206 to 0.628)
Years since diagnosis	1.103 (1.011 to 1.203), p = 0.027	1.130 (0.982 to 1.301)

Neurocognitive deficit is defined as a z-score < 1.5 in at least one neurocognitive domain. p-values are only showed for the univariable analysis, as they were used for development of the multivariable models that was based on statistical significance