Skip to main content

Table 7 Estimated parameters, Goodness-of-fit statistics and dose reduction of GQrank

From: Regression models for analyzing radiological visual grading studies – an empirical comparison

Model Coefficient Goodness-of-fit Dose Reduction
logCTDI id2 id4 AIC Pseudo R2 id2 id4
  Est. P-value Est. P-value Est. P-value
regressa −4.671 <0.001 −0.817 <0.001 −0.463 <0.001 - 0.1134 16.04 % 9.43 %
(−5.158, −4.183) (−.989, −.644) (−0.635, −0.290) (13.36 %, 18.73 %) (6.44 %, 12.42 %)
ologita −9.433 <0.001 −1.691 <0.001 −1.033 <0.001 2427.97 0.1269 16.41 % 10.37 %
(−10.550, −8.317) (−2.039, −1.343) (−1.395, −0.672) (13.93 %, 18.88 %) (7.39 %, 13.36 %)
rologita −4.766 <0.001 −0.709 <0.001 −0.576 <0.001 1337.79 0.1270 13.81 % 11.38 %
(−5.516, −4.015) (−0.940, −0.477) (−0.823, −0.328) (10.23 %, 17.40 %) (7.35 %, 15.41 %)
mixedb −4.671 <0.001 −0.8167 <0.001 −0.4625 <0.001 2619.45 0.0000 16.04 % 9.43 %
(−5.145, −4.196) (−0.985, −0.648) (−0.631, −0.294) (13.42 %, 18.66 %) (6.51 %, 12.34 %)
meologitb −9.380 <0.001 −1.680 <0.001 −1.027 <0.001 2367.96 0.0105 16.40 % 13.37 %
(−9.506, −9.255) (−1.956, −1.405) (−1.331, −0.723) (13.91 %, 18.89 %) (7.45 %, 13.29 %)
  1. 95 % confidence limits of each estimate given in parentheses
  2. aregression model with fixed effects only
  3. bregression model with fixed and random effects