Table 6

Major inconsistencies between conference abstract and final publication (n = 76) for all studies, as well as for randomized-controlled trials (RCT) and observational studies

RCT

Observational studies

Experimental study design

Total Publications


Percentage %

95% CI

Percentage %

95% CI

N

95% CI

n


Major inconsistency

8 (80%)

44-97

41 (54%)

52-77

0

49 (65%)

53-75

49


Study hypothesis different

0 (0%)

0-30

7 (11%)

5-22

3

10 (13%)

6-23

69


Study design different

0 (0%)

0-30

8 (13%)

6-24

3

11 (15%)

7-24

68


Primary outcome measure different

1 (10%)

0-46

10 (16%)

8-27

3

14(18%)

10-29

65


Sample size different

7 (70%)

35-93

25 (40%)

28-53

0

32 (42%)

31-54

32


Increased

4 (40%)

12-74

15 (24%)

14-36

0

19 (25%)

23-45

19


Decreased

3 (30%)

7-65

12 (19%)

10-31

0

15 (20%)

11-30

15


Not stated*

0 (%)

0-30

4 (6%)

2-16

2

6 (8%)

3-16

6


Sample size same

3 (30%)

7-65

32 (51%)

38-64

1

36 (48%)

36-59

36


Statistic Not stated

9 (90%)

56-100

59 (94%)

85-98

3

71 (93%)

85-98

71


Same

1 (10%)

0-46

2 (3%)

0-11

0

3 (4%)

1-11

3


Result different

9 (90%)

56-100

43 (68%)

55-79

3

55 (72%)

60-82

24


Data added

6 (60%)

26-88

19 (30%)

19-43

0

25 (33%)

23-45

25


Data deleted

4 (40%)

12-74

19 (30%)

19-43

0

23 (30%)

20-42

23


Complete changed

0

0-30

3 (5%)

1-13

0

3 (4%)

1-11

3


SD/CI/, Injury rate, Incidence

changed

3 (30%)

7-65

15 (24%)

14-36

0

18 (24%)

15-35

18


*in both conference and full-text abstract

Yoon and Knobloch BMC Medical Research Methodology 2012 12:47   doi:10.1186/1471-2288-12-47

Open Data