Table 5.

Results of random-effects GLS regression to explore relationships between demographic and clinical variables and SF-6D health state utility values

SF-6D value	Coefficient	Standard error	p > z	Lower confidence interval	Upper confidence interval	Beta coefficientsa
TFC score	0.0091764	0.0011486	0.0000000	0.0069239	0.0114290	0.2328026
Behavioural score	− 0.0047258	0.0002288	0.0000000	− 0.0051744	− 0.0042771	− 0.4053759
Cognition score	0.0000891	0.0000596	0.1350000	− 0.0000278	0.0002060	0.0430957
MOT score	− 0.0002756	0.0002094	0.1880000	− 0.0006862	0.0001350	− 0.0384748
Male gender	0.0188754	0.0052631	0.0000000	0.0085535	0.0291973	0.0686814
Age	− 0.0002946	0.0002174	0.1760000	− 0.0007210	0.0001318	− 0.0270106
Years since diagnosis	0.0014575	0.0008332	0.0800000	− 0.0001766	0.0030917	0.0375215
Constant	0.6527599	0.0233794	0.0000000	0.6069084	0.6986114

n = 1939; F [7,1931] = 120.05; p < 0.001; Adjusted R-squared 0.3007

^aBeta coefficients were obtained by standardising all variables to a mean of 0 and a standard deviation of 1, and then including them in the regression analysis