BMJ No 7109 Volume 315
Papers Saturday 13 September 1997
Publication bias: evidence of delayed publication in a cohort study of clinical research projects
Jerome M Stern, R John Simes
| Table 1-Characteristics of eligible studies | |
|---|---|
| Characteristic | No (%) of studies |
| All studies (n=748) | |
| Research design: | |
| Observational study | 165 (22) |
| Clinical trial | 418 (56) |
| Non-trial experiment | 165 (22) |
| Pilot study: | |
| Pilot study | 39 (5) |
| Non-pilot study | 709 (95) |
| Study groups: | |
| Uncontrolled | 392 (52) |
| Controlled (non-concurrent, historical) | 5 (1) |
| Controlled (concurrent) | 351 (47) |
| Clinical trials (n=418) | |
| Clinical trial phase: | |
| 1 | 18 (4) |
| 2 | 178 (43) |
| 3 | 221 (53) |
| 4 | 1 (<1) |
| Pilot study: | |
| Pilot study | 22 (5) |
| Non-pilot study | 396 (95) |
| Study groups: | |
| Uncontrolled | 142 (34) |
| Controlled (historical) | 3 (1) |
| Controlled (concurrent), non-randomised | 76 (18) |
| Controlled, randomised | 197 (47) |
| Placebo controlled study: | |
| Placebo controlled | 81 (19) |
| Not placebo controlled | 337 (81) |
| Blinded treatment (single or double blind): | |
| Blinded | 107 (26) |
| Non-blinded | 311 (74) |
| Table 2-Stage of studies for 520 questionnaires with completed questionnaires | |
|---|---|
| Stage | No of studies |
| Not started: | |
| Abandoned | 97 |
| Yet to be started | 3 |
| Started: | |
| Abandoned | 70 |
| Still in progress | 63 |
| Completed | 287 |
Fig 1-Proportion of quantitative studies not published, according to
type of results.
Table 4-Risk factors for time to publication using univariate
Cox regression analysis
Publication bias: evidence of delayed publication in a cohort
study of clinical research projects
Table 3-Outcome (results) of quantitative and qualitative studies
All analysed studies
Clinical trials
Quantitative studies
Total No
218
130
Results:
Significant
146 (67)
76 (58)
Non-significant trend
20 (9)
15 (12)
Non-significant
52 (24)
39 (30)
Qualitative studies
Total No
103
37
Results:
Striking
27 (26)
6 (16)
Important and definite
59 (57)
18 (49)
Unimportant and negative
17 (17)
13 (35)
Table 4-Risk factors for time to publication using univariate
Cox regression analysis
Characteristic
No not published
No published
Hazard ratio (95% CI)
P value
All studies (n=321)
Outcome:
Quantitative studies:
<0.0001
Null
29
23
1.00
Non-significant trend
16
4
0.39 (0.13 to 1.12)
Significant
47
99
2.32 (1.47 to 3.66)
Qualitative studies:
0.25
Negative and unimportant
8
9
1.00
Important and definite
24
35
1.22 (0.59 to 2.55)
Striking
8
19
1.56 (0.70 to 3.45)
Research design:
0.008
Observational study
42
48
1.00
Clinical trial
71
96
1.03 (0.73 to 1.46)
Non-trial experiment
19
45
1.74 (1.15 to 2.61)
Funding:
0.02
Internal or none
59
58
1.00
External
73
133
1.45 (1.06 to 1.99)
Pilot study:
0.07
Non-pilot study
120
183
1.00
Pilot study
12
6
0.51 (0.23 to 1.15)
Study groups:
0.02
Non-comparative
62
100
1.00
Comparative
70
89
0.72 (0.54 to 0.96)
Scientific importance:
<0.0001
Low
74
51
1.00
High
58
138
2.35 (1.69 to 3.26)
Sample size:
0.44
<100
80
100
1.00
|LX100
52
82
1.12 (0.84 to 1.51)
Clinical trials (n=167)
Outcome:
Quantitative trials:
<0.0001
Null
24
15
1.00
Non-significant trend
12
3
0.46 (0.13 to 1.59)
Significant
21
55
3.13 (1.76 to 5.58)
Qualitative trials:
0.39
Negative and unimportant
4
9
1.00
Important and definite
7
11
0.97 (0.40 to 2.36)
Striking
3
3
0.53 (0.14 to 1.95)
Clinical trial phase:
0.13
1
3
5
1.15 (0.45 to 2.91)
2
22
40
1.00
3
46
51
0.67 (0.44 to 1.02)
4
0
0
Study groups:
0.02
Non-comparative
19
33
1.00
Comparative (non-randomised)
10
17
1.07 (0.60 to 1.93)
Randomised
42
46
0.61 (0.39 to 0.96)
Home | Current contents | Past issues |
Classified ads | Career Focus | Feedback
Collections |
About this site | About the BMJ | BMA | Medline
