Skip to main content

Table 5 Intention-to-treat quality-adjusted life-years and costs at 12 weeks: TAU vs TAU + hearing aid

From: Effectiveness and cost effectiveness of digital hearing aids in patients with tinnitus and hearing loss: a randomised feasibility trial (THE HUSH Trial)

Variable

TAU (N = 41)

TAU + hearing aid (N = 42)

Mean difference

EQ-5D (n)

(27)

(33)

(60)

 Unadjusted

0·196 (0·033)

0·178 (0·053)

 − 0·017 (− 0·041 to 0·004)

 Baseline adjusted

0·185 (0·179 to 0·191)

0·187 (0·181 to 0·192)

0·002 (− 0·006 to 0·009)

 Covariate adjusted

0·186 (0·18 to 0·192)

0·186 (0·181 to 0·192)

0·001 (− 0·007 to 0·009)

EQ-VAS (n)

(27)

(33)

(60)

 Unadjusted

0·199 (0·039)

0·186 (0·046)

 − 0·013 (− 0·034 to 0·009)

 Baseline adjusted

0·189 (0·181 to 0·197)

0·194 (0·187 to 0·202)

0·006 (− 0·006 to 0·016)

 Covariate adjusted

0·19 (0·181 to 0·198)

0·194 (0·186 to 0·201)

0·004 (− 0·008 to 0·016)

HUI3 (n)

(26)

(31)

(57)

 Unadjusted

0·141 (0·085)

0·132 (0·076)

 − 0·009 (− 0·05 to 0·035)

 Baseline adjusted

0·131 (0·119 to 0·142)

0·14 (0·13 to 0·151)

0·01 (− 0·007 to 0·029)

 Covariate adjusted

0·132 (0·12 to 0·143)

0·14 (0·129 to 0·15)

0·008 (− 0·01 to 0·022)

NHS Costs (n)

(18)

(20)

(38)

 Unadjusted

104·63 (46·09 to 163·17)

251·38 (145·09 to 163·17)

146·75 (25·70 to 272·07)

 Multicentre adjusted

108·64 (28·84 to 188·45)

247·77 (155·61 to 339·92)

139·13 (12·58 to 254·85)

  1. Data are mean (SD), or mean difference (95% CI). EQ-5D, EuroQoL-5D. EQ-VAS, EuroQol Visual Analogue Score. HUI3, Health Utilities Index Mark 3. We report univariate distributions as deriving the joint distributions of complete case data presenting differential observational set sizes would inherently exclude a significant number of observations from the larger (QALY) set. Covariate adjustment included continuous variables for baseline utility and age, and binary variables for sex (male as base) and multicentre effects (site one as base—site six was excluded due to collinearity and so included as base). EQ-VAS estimates were divided by one hundred to aid comparison with the other instruments. Estimates of costs considered all reported service use as NHS applicable, as such reporting a lower n to avoid double counting the items which we could not guarantee were not private service use