Empirical evidence for outcome reporting bias in randomized clinical trials of acupuncture: comparison of registered records and subsequent publications

Table 3 Proportion of discrepancies in the specified primary outcomes between registered trials and published articles on acupuncture and discrepancies favoring statistically significant results

Discrepancy in published articles relative to registered trials	Total	Western countries	Eastern countries
	n = 71 (%)	n = 50 (%)	n = 21 (%)
Articles with different primary outcomes in the trial registration and the published article	32 ^a(45.1)	26 ^b(52.0)	6 ^c(28.6)
Registered primary outcome omitted in published articles	22 (31.0)	18 (36.0)	4 (19.0)
An absent primary outcome in the registry defined in the published article	9 (12.7)	7 (14.0)	2 (9.5)
A published primary outcome registered as a secondary outcome	1 (1.4)	1 (2.0)	0 (0.0)
A registered primary outcome defined as a secondary outcome in the published article	15 (21.1)	13 (26.0)	2 (9.5)
Different timing of assessment of the primary outcome	12 (16.9)	8 (16.0)	4 (19.0)
Discrepancies in primary outcome favoring statistically significant results ^d	32	26	6
Yes	15 (46.9)	13 (50.0)	2 (33.3)
No	6 (18.8)	4 (15.4)	2 (33.3)
Impossible to conclude	11 (34.3)	9 (34.6)	2 (33.3)

^aFourteen articles had two reasons for the difference in primary outcome; five articles had three reasons for the difference in primary outcome; one article had four reasons for the difference in primary outcome.
^bEleven articles had two reasons for the difference in primary outcome; five articles had three reasons for the difference in primary outcome.
^cThree article had two reasons for the difference in primary outcome; one article had four reasons for the difference in primary outcome. Compared with articles from western countries: P = 0.07.
^dA discrepancy in primary outcome was said to favor statistically significant results when a new, statistically significant primary outcome was introduced in the article or when a statistically nonsignificant primary outcome was defined as a nonprimary outcome in the published article.

ISSN: 1745-6215