Table 5.

Linear mixed regressions with the pooled dataset

Variable	HbA1c	Cholesterol	CVD risk${}_{\mathit{log}}$
Intercept	4.716 [4.663; 4.769]	4.849 [4.822; 4.876]	− 0.658 [− 0.683; − 0.633]
Degree (ref. other qualification)	− 0.171 [− 0.203; − 0.138]	− 0.033 [− 0.062; − 0.005]	− 0.127 [− 0.150; − 0.104]
No qualification	0.160 [0.136; 0.185]	0.268 [0.247; 0.289]	0.136 [0.119; 0.152]
Year	0.033 [0.032; 0.035]	− 0.026 [− 0.027; − 0.025]	− 0.015 [− 0.016; − 0.014]
Degree x Year	0.002 [− 0.000; 0.003]	0.001 [− 0.001; 0.003]	0.000 [− 0.001; 0.001]
No qualification x Year	0.000 [− 0.001; 0.002]	− 0.015 [− 0.017; − 0.014]	− 0.004 [− 0.005; − 0.003]
HSE (ref. NHANES)	− 0.911 [− 1.017; − 0.805]	0.432 [0.392; 0.472]	0.139 [0.103; 0.174]
SHS	v0.277 [− 0.406; − 0.148]	0.537 [0.482; 0.592]	0.252 [0.204; 0.300]
n	211251	235794	170684
m	200	201	200

Coefficients $\beta$ and 99% confidence intervals. m and n: number of random-effect strata levels and sample size, respectively. HbA1c glycated haemoglobin, CVD cardiovascular disease. Log: logarithm