Table 2.

Performance and optimal variable combinations produced by multivariate regression analysis to predict risk of grade 2+ CIPN.

					Variable importance
		Estimate	p	OR	Main Effect	Total Effect
Model 1 variables
Taxane type [paclitaxel]		0.98	<0.0001
p16 pre-chemotherapy		−0.67	0.02
Arthritis [yes]		0.62	0.03
Age		−0.46	0.16
Age*Osteoporosis [yes]		−0.49	0.14
Osteoporosis [yes]		4.83	0.16
Age*Arthritis [yes]		−0.03	0.23
Model 1 performance
AICc	150
BIC	172
Chi-square	34.9
p value	0.0001
Model 2 variables
Taxane type [paclitaxel]		0.98	<0.0001	7.2 (2.9–17.5)	NA*	NA*
p16Age Gap		−0.047	0.01	0.95 (0.92–0.99)	0.406	0.519
p16 pre-chemotherapy		1.22	0.04	3.4 (1.1–10.8)	0.363	0.477
Model 2 performance
AICc	161
BIC	173
Chi-square	29.3
p value	0.0001

Variables tested in Model 1- taxane type, race, BMI ≥ 30, diabetes, peripheral circulatory issues, osteoporosis, arthritis, high blood pressure, liver or kidney disease, age, interaction of each comorbidity and age, and p16 prior to chemotherapy.

Variables tested in Model 2- taxane type, race, BMI ≥ 30, diabetes, peripheral circulatory issues, osteoporosis, arthritis, high blood pressure, liver or kidney disease, age, interaction of each comorbidity and age, p16 prior to chemotherapy, and p16Age Gap.

*Variable importance for categorical variables is not calculated.