Table 6.

Regression model — mean absolute error^a

	Estimate	95% CI	P-value
Intercept	0.62	0.37 to 1.04	0.051
Length of time in service
Years in service (each additional year)	1.08	1.04 to 1.12	<0.001
Frequency of use
>10 times per month	0.94	0.66 to 1.38	0.76
6–10 times per month	0.82	0.61 to 1.12	0.19
1–5 times per month	1	reference
6–11 times per year	0.90	0.68 to 1.20	0.47
1–5 times per year	0.94	0.66 to 1.36	0.74
Less frequently	0.89	0.61 to 1.32	0.53
Not known	0.80	0.38 to 2.07	0.61
Cost of device
Cost (every £1 increase)	1.00	0.99 to 1.01	0.75
Tested (reference: no)
Yes	1.21	0.80 to 1.91	0.38
Not known	1.10	0.87 to 1.41	0.44
Make (reference: Boots)
Lloyds	1.31	0.90 to 1.90	0.13
Omron	1.01	0.75 to 1.37	0.94
Other	1.23	0.80 to 1.92	0.34
Validation status (reference: not validated/not applicable)
Validated	0.77	0.59 to 1.02	0.044

^aThe proportion in MAE change is the unit of predictor unless otherwise stated. MAE = mean absolute error.