Table 2

Multiple linear regression models of EQ-5D utility values and candidate predictors

Individual predictorsMultiple regressionFinal modelR2R2 change
PredictorCoefficientProbabilityCoefficientProbabilityCoefficientProbability
βpβpβp
Pain−0.582<0.001***−0.361<0.001***−0.393<0.001***0.3980.340
Depression−0.526<0.001***−0.215<0.001***−0.243<0.001***0.4820.084
Anxiety−0.448<0.001***−0.1420.003**−0.1500.001**0.4940.012
Fatigue−0.518<0.001***−0.1390.009**−0.0770.012*0.5010.007
Dryness−0.242<0.001***+0.0270.503+0.0670.5710.5020.001
BMI−0.0840.020*0.5100.008
  • EuroQoL-5 dimension (EQ-5D) utility values are rescaled between 0 and 1 prior to the logit transformation (EQ-5D utility value − EQ-5D utility valuemin)/(EQ-5D utility valuemax − EQ-5D utility valuemin).

  • Note that all regression coefficients are reported after adjustment for age and sex differences; age and sex are included in the model.

  • Note that age and sex alone give an R2 value of 0.058 and, once the first term Pain is added to the model, this increases to 0.398 (total R2=0.510).

  • With pain and depression scores the model accounted for 48% of the variability in EQ-5D utility values.

  • BMI, body mass index.

  • ***p<0.001; **p<0.01; *p<0.05.