Table 4. Subgroup analyses: repeatability of the health questionnaire (n = 286).

	n	ra	1. HQb cups/day	diff.c (%)	Pd
Age (years)
20–29	—	—	—	—	—
30–59	242	0.78	6.03	0.01	0.59
≥60	40	0.67	6.35	−3.1	0.60
Sex
Male	153	0.76	5.97	−1.3	0.66
Female	129	0.78	6.21	2.6	0.27
Marital status
Married	232	0.76	6.11	−1.0	0.62
Living alonee	50	0.78	5.92	7.8	0.08
Education (years)
1–9	218	0.74	6.29	0.6	0.77
≥10	63	0.86	5.41	0	1.00
Community density
Rural	56	0.82	6.25	2.9	0.48
Urban	226	0.75	6.04	1.3	0.55
Alcohol consumption (grams/day)
0	111	0.81	6.45	1.2	0.63
<12	131	0.75	5.96	0	1.00
≥12	40	0.70	5.43	−0.5	0.95
Body mass index (kg/m2)
<25	136	0.78	6.15	1.3	0.64
25–29.9	105	0.75	5.66	−0.9	0.80
≥30	41	0.70	6.90	0.7	0.87
Physical activityf
None	67	0.69	5.96	3.2	0.48
Light activity ≥4 hours/week	177	0.76	6.17	0	1.00
Heavy activity ≥3 hours/week	35	0.93	5.74	−3.0	0.34
Heavy activity almost daily	3	—	—	—	—
Smoking
Never	139	0.83	5.91	4.1	0.06
Past	64	0.79	6.00	−2.3	0.53
Current	79	0.69	6.43	−3.1	0.51
Hypertension
No	209	0.80	6.11	0.5	0.82
Yes	73	0.67	6.00	0.5	0.92
Serum cholesterol
<7.23 mmol/lg	141	0.77	5.96	−2.2	0.41
≥7.23 mmol/l	141	0.76	6.20	3.1	0.25

^aIntraclass correlation coefficient, coffee consumption measured as a continuous variable cups/day.

^bThe first health questionnaire, average daily coffee consumption in cups.

^cDifferences between the first and second health questionnaires (%), where negative values indicate that reported coffee consumption was higher on the second health questionnaire.

^dStatistical significance of the difference between the 2 measurements.

^eSingle, divorced or widow/er.

^fLeisure-time physical activity.

^g7.23 mmol/l was the median serum cholesterol level of the subjects.