Multivariable binary logistic regression assessing the effect of demographic variables, smoking status, and medical condition on the occurrence of xerostomia, taste disturbances, and smell disturbances

ParametersAgeFemale vs. maleSmoking vs. no smokingMedical condition vs. no medical conditionLong medication vs. no long medication
XerostomiaAOR1.031.452.471.070.96
95% CI1.01, 1.041.02, 2.051.39, 4.400.56, 2.030.55, 1.65
p-value0.002*0.038*0.002*0.8380.871
Taste disturbancesAOR1.021.452.390.951.19
95% CI1.01, 1.041.05, 2.001.50, 3.820.54, 1.690.72, 1.95
p-value< 0.0001*0.025*< 0.0001*0.8710.501
Smell disturbancesAOR1.021.902.451.131.19
95% CI1.01, 1.041.36, 2.651.46, 4.090.61, 2.110.69, 2.03
p-value0.005*< 0.0001*0.001*0.6910.533

AOR: adjusted odds ratio; CI: confidence interval; * statistically significant difference at p-value < 0.05. Models’ summary: xerostomia (χ2 = 38.63, df = 5, p-value < 0.0001, –2 log likelihood = 914.78, Nagelkerke R Square = 0.067, percentage of correctly classified = 73.6%), taste disturbances (χ2 = 54.13, df = 5, p-value < 0.0001, –2 log likelihood = 1,083.94, Nagelkerke R Square = 0.085, percentage of correctly classified = 61.1%), smell disturbances (χ2 = 55.57, df = 5, p-value < 0.0001, –2 log likelihood = 982.533, Nagelkerke R Square = 0.091, percentage of correctly classified = 67.6%)