Skip to main content
. 2017 Feb 9;7(2):e012374. doi: 10.1136/bmjopen-2016-012374

Table 2.

Linear regression of health-related quality of life on community asset participation

Dependent variable: EQ5D Health Utility Index
Model specification (1) (2) (3)
Community asset participation 0.0809*** 0.0626*** 0.0387***
(0.064 to 0.098) (0.048 to 0.077) (0.025 to 0.052)
Female participants −0.0381*** −0.0111 −0.00884
(−0.056 to −0.021) (−0.027 to 0.005) (−0.022 to 0.005)
Aged 65–69 years Reference category
Aged 70–74 years 0.00277 −0.00151 −0.00793
(−0.020 to 0.025) (−0.020 to 0.017) (−0.025 to 0.009)
Aged 75–79 years −0.0180 −0.00695 −0.00853
(−0.042 to 0.006) (−0.027 to 0.013) (−0.026 to 0.009)
Aged 80–84 years −0.0405** −0.0109 −0.0331**
(−0.069 to −0.012) (−0.035 to 0.013) (−0.056 to −0.011)
Aged 85–98 years −0.120*** −0.0680*** −0.0706***
(−0.153 to −0.088) (−0.098 to −0.038) (−0.098 to −0.043)
Socioeconomic characteristics† Yes Yes Yes
Health conditions‡ No Yes Limiting
Goodness of fit (adjusted R2) 0.075 0.361 0.461

Sample size is N=3686.

95% CIs in brackets.

**p<0.01; ***p<0.001.

†Educational qualifications and living arrangements. See data section for full details.

‡See Data section for details on the health conditions. Column 2 includes indicators to show whether or not an individual has any of the 23 conditions. Column 3 includes an indicator to show whether or not the conditions are limiting—defined as a response of 4 or 5 on the Bayliss score.

EQ5D, EuroQol-5D.