Table 3 Follow-up status of study smoke alarms, by follow-up visit and alarm type

Characteristic	Follow-up 1			Follow-up 2			Follow-up 2*
Characteristic	Ionization (n = 332) (%)	Photoelectric (n = 340) (%)	p-value†	Ionization (n = 311) (%)	Photoelectric (n = 314) (%)	p-value†	Ionization (n = 248) (%)	Photoelectric (n = 296) (%)	p-value†
Participant-reported status since last visit
Fire	2.4	2.4	0.97	1.6	2.6	0.41	0.8	2.4	0.19
Study alarm sounded	77.9	38.7	<0.001	76.8	42.0	<0.001	74.2	41.8	<0.001
No. of times study alarm sounded‡			<0.001			<0.001			<0.001
1–3	44.1	83.5		56.3	88.2		61.4	89.9
4–10	26.6	11.0		27.7	11.0		21.7	9.2
⩾11	29.4	5.5		16.0	0.8		16.9	0.8
Other alarm(s) sounded§	42.5	45.4	0.49	33.0	37.6	0.26	30.2	36.7	0.14
No. of times other alarm(s) sounded‡			0.89			0.28			0.07
1–3	73.0	72.5		65.5	73.0		60.9	73.1
4–10	13.9	15.8		27.6	18.0		32.8	17.2
⩾11	13.0	11.7		6.9	9.0		6.3	9.7

Observed status of study alarm
Functional	79.6	94.7	<0.001	76.8	91.4	<0.001	85.8	93.2	0.004
Non-functional due to
Malfunctioning alarm	0.3	0.0	0.49	0.3	0.3	1.00	0.4	0.3	1.00
Disconnected battery	5.8	0.6	<0.001	5.5	1.3	0.004	4.1	0.7	0.008
Absent battery	12.8	1.2	<0.001	14.5	2.2	<0.001	7.7	2.4	0.004
Dead battery	0.9	1.5	0.73	1.3	1.3	1.00	1.6	1.4	1.00
Incorrect battery placement	0.0	0.0	–	0.0	0.3	1.00	0.0	0.3	1.00
Unknown	0.3	0.3	1.00	0.3	0.0	0.50	0.0	0.0	–
Study alarm removed	0.3	1.8	0.12	1.3	3.2	0.11	0.4	1.7	0.23

*Restricted to households with functional study alarms at follow-up 1.
†Based on Pearson’s χ² test if expected ⩾5 in all cells; based on Fisher’s exact test if expected <5 in any cell.
‡Restricted to households whose smoke alarm(s) had alarmed since the last visit.
§Restricted to households with reported non-study alarms at baseline.