Table 6

Association between the trial retention rate (% of randomised participants with valid primary outcome data for analysis) and trial characteristics

Characteristic (n=388)		n	Median	IQR	P value
Setting	Hospital	219	91.5	82.2–97.8	0.001*†
	General practice	55	84.0	76.6–91.3
	Mixed	61	87.3	79.7–97.3
	Community	34	84.9	75.4–90.8
	Other	19	84.2	74.9–96.5
Arms	2	290	89.9	81–97.4	<0.001‡
	3	61	84.4	72.4–93.6
	4	24	83.2	79.6–88.2
	>4	13	80.2	73.4–96.4
Control type	Placebo	87	89.8	79.1–97.3	0.614§
Control type	Active	301	87.8	80.3–96.4
Final target recruitment	≤200	53	88.6	79.6–96.4	0.003‡
	201–400	112	86.1	77.1–94.1
	401–600	84	86.8	78.9–95.7
	601–800	42	84.4	80.4–90.9
	>800	97	96.3	85.3–99.1
Total recruitment	≤200	72	87.9	74.5–96.2	0.001‡
	201–400	99	87.3	79.3–94.9
	401–600	82	86.4	80.6–94.1
	601–800	39	86.2	82.2–91.4
	>800	96	95.8	82.4–99
Timing of final follow-up	≤1 month	20	92.2	78.7–99	0.518‡
	1< months ≤6	87	88.5	79.8–96.7
	6< months ≤18	181	88.2	79.5–96.4
	>18 months	88	87.8	80–95.5

*The category ‘other’ was not included in Kruskal-Wallis test.
†P values are reported from a Kruskal-Wallis test.
‡P values are reported from a Jonckheere-Terpstra test.
§P values are reported from a Mann-Whitney U test.