Recruitment into antibody prevalence studies: a randomized trial of postcards vs. letters as invitations

Table 1 Comparison of invitation response rates and days to response for first and second mailings and their combinations by mailing type (postcard and letter)

		Mailings sent	Responded	No response	Response rate (%)	P 1	OR	CI	Days to response ^a	P 2
		N	N	N	%				Median (Q1-Q3), range
Mailing #1	P	4999	150	4849	3.00	< 0.001	1		5 (3–7), 2–35	0.07
	L	5000	265	4735	5.30		1.81	1.48–2.22	4 (3–7), 2–31
	Total	9999	415	9584	4.15
Mailing #2 ^b (reminder)	P	4942	235	4707	4.76	< 0.001	1		6 (3–15), 1–45 ^c	0.65
	L	5057	441	4616	8.72		1.91	1.63–2.25	5 (3–12), 1–49 ^c
	Total	9999	676	9323	6.76
Combinations of mailings ^b	P0 or P/P	2575	197	2378	7.7	< 0.001	1		5 (3–9), 1–45	0.87
	P0 or P/L	2574	244	2330	9.5		1.26	1.04–1.54	5 (3–9), 1–49
	L0 or L/P	2632	303	2329	11.5		1.57	1.30–1.90	5 (3–7), 1–33
	L0 or L/L	2633	347	2286	13.2		1.83	1.52–2.20	5 (3–8), 1–49

P0 Postcard only, P/P Postcard/postcard, P/L Postcard/letter, L0 Letter only, L/P Letter/postcard, L/L Letter/letter. P1: comparison of response rates, P2: comparison of days to response, OR Odds ratio, CI 95% confidence interval, Q1-Q3 Interquartile range
^a Time to response: for those missing survey completion date (~ 20%), estimated as that of next dated record in database
^b Households who responded to their first mailing are included in each category, as they would have received a second mailing had they not already responded (alternatively, N received after second mailing but not including initial mailing]: P: 85, L: 176)
^c Calculation includes only the responses from the second mailing

ISSN: 1471-2288