Abstract
This study examines the extent to which trials registered in ClinicalTrials.gov by individual sponsors for the same drug and condition were available publicly in ClinicalTrials.gov and/or PubMed.
Missing or incomplete reporting of clinical trial results and its scientific and ethical consequences are well documented.1,2 One concrete example of this problem occurs when a sponsor conducts several studies of a particular drug for a particular condition but only some (or none) of the studies make their way into the public domain, leaving a distorted body of public evidence.3 Policies that require reporting to ClinicalTrials.gov are designed to ameliorate that problem. ClinicalTrials.gov currently has more than 255 000 registered studies, with summary results information for more than 28 000 studies, of which an estimated half lack corresponding publications.2,4,5 To evaluate the influence of this resource on the evidence base, we examined the extent to which sets of trials registered in ClinicalTrials.gov by individual sponsors for the same drug and condition had results available publicly in ClinicalTrials.gov and/or PubMed.
Methods
We identified ClinicalTrials.gov-registered records for industry-sponsored, phase 2 to 4 trials of drugs or biologics, with at least 1 US study location, completed or terminated from January 1, 2007, through December 31, 2009 (allowing at least 7 years from trial completion for results reporting). We manually grouped trials into a sponsor-drug-condition trial set when the listed sponsor, drug, and condition appeared to be identical; we limited the analysis to a convenience sample of the first 96 sponsor-drug-condition trial sets identified. On July 25, 2017, we assessed the availability of results in PubMed-cited publications and/or results posted in ClinicalTrials.gov based on methods used in prior research.4 We also evaluated whether the drug was approved by the US Food and Drug Administration (FDA) for any use.
Results
Our sample consisted of 329 trials studying 86 drugs and representing 96 unique sponsor-drug-condition trial sets (eg, Amgen-sponsored trials of alendronate for osteoporosis). The median number of trials per set was 3 (range, 2-11). The median proportion of trials in each set with results publicly available (in PubMed and/or ClinicalTrials.gov) was 100% (interquartile range, 62.5%-100%). Most trials used parallel assignment, double-blinding, randomization, and multiple sites and had more than 100 participants enrolled (Table 1). As of July 25, 2017, results were publicly available for 255 of 329 trials (77.5%): 79 (24.0%) from ClinicalTrials.gov only, 63 (19.1%) from PubMed only, and 113 (34.3%) from both sources. Overall, 58 (60.4%) of 96 sponsor-drug-condition trial sets had results available for all trials, and 13 (13.5%) had no results available from either source (Table 2). There were 15 (15.6%) trial sets with results from ClinicalTrials.gov only (10 sets with results posted for all trials and 5 sets with results posted for some trials) (Table 2).
Table 1. Study Design Characteristics of the 329 ClinicalTrials.gov-Registered Trials in the Sponsor-Drug-Condition Trial Sets by Results Availability Category.
| Study Design Characteristic | No. (%) of Trials by Availability of Results as of April 27, 2017 | ||||
|---|---|---|---|---|---|
| Results Disclosure | Any vs No Results Disclosure on ClinicalTrials.gov and PubMed | ||||
| ClinicalTrials.gov Only (n = 109) | PubMed Only (n = 42) | Both ClinicalTrials.gov and PubMed (n = 81) | Results Disclosure Total (n = 232) | Neither ClinicalTrials.gov nor PubMed (n = 97) | |
| Interventional model | |||||
| Parallel assignment (n = 242) | 79 (32.6) | 33 (13.6) | 68 (28.1) | 180 (74.4) | 62 (25.6) |
| Single-group assignment (n = 72) | 26 (36.1) | 7 (9.7) | 8 (11.1) | 41 (56.9) | 31 (43.1) |
| Crossover assignment (n = 9) | 2 (22.2) | 1 (11.1) | 4 (44.4) | 7 (77.8) | 2 (22.2) |
| Factorial assignment (n = 4) | 2 (50.0) | 0 | 1 (25.0) | 3 (75.0) | 1 (25.0) |
| Missing data (n = 2) | 0 | 1 (50.0) | 0 | 0 (50.0) | 1 (50.0) |
| Masking | |||||
| Double-blind (n = 209) | 62 (30.5) | 32 (15.3) | 57 (27.3) | 151 (72.2) | 58 (27.8) |
| Single-blind (n = 11) | 6 (54.5) | 0 | 2 (18.2) | 8 (72.6) | 3 (27.3) |
| Open label (n = 99) | 37 (37.4) | 9 (9.1) | 18 (18.2) | 64 (64.6) | 35 (35.4) |
| Missing data (n = 10) | 4 (40.0) | 1 (10.0) | 4 (40.0) | 9 (90.0) | 1 (10.0) |
| Allocation | |||||
| Randomized (n = 246) | 75 (30.5) | 34 (13.8) | 70 (28.5) | 179 (72.8) | 67 (27.2) |
| Nonrandomized (n = 53) | 17 (32.1) | 4 (7.5) | 10 (18.9) | 31 (58.5) | 22 (41.5) |
| Missing data (n = 30) | 17 (56.7) | 4 (13.3) | 1 (3.3) | 22 (73.3) | 8 (26.7) |
| No. of sites | |||||
| Multiple (n = 218) | 70 (32.1) | 29 (13.3) | 62 (28.4) | 161 (73.9) | 57 (26.1) |
| Single (n = 58) | 19 (32.8) | 6 (10.3) | 3 (5.2) | 28 (48.3) | 30 (51.7) |
| Missing data (n = 53) | 20 (37.7) | 7 (13.2) | 16 (30.2) | 43 (81.1) | 10 (18.9) |
| No. of participants enrolled | |||||
| 1-100 (n = 102) | 37 (36.2) | 13 (12.7) | 11 (10.8) | 61 (59.8) | 41 (40.2) |
| 101-500 (n = 150) | 52 (34.7) | 21 (14.0) | 32 (21.3) | 105 (70.0) | 45 (30.0) |
| >500 (n = 75) | 20 (26.7) | 8 (10.7) | 38 (50.7) | 66 (88.0) | 9 (12.0) |
| Missing data (n = 2) | 0 | 0 | 0 | 0 | 2 (100) |
Table 2. Results Availability by Sponsor-Drug-Condition Trial Set.
| Results Availabilitya | No. (%) of Sponsor-Drug-Condition Trial Sets (N = 96) |
|---|---|
| Complete | 58 (60) |
| ClinicalTrials.gov only | 10 (10) |
| PubMed only | 9 (9) |
| Both ClinicalTrials.gov and PubMed | 12 (13) |
| Other combinationsb | 27 (28) |
| Partial | 25 (26) |
| ClinicalTrials.gov only | 5 (5) |
| PubMed only | 9 (9) |
| Both ClinicalTrials.gov and PubMed | 2 (2) |
| Other combinationsb | 9 (9) |
| None | 13 (14) |
For complete, results were available for all trials in sponsor-drug-condition trial set; partial, results were available for some trials in sponsor-drug-condition trial set; none, no results were available for any trial in sponsor-drug-condition trial set.
Other combinations means that individual trials within each set had different reporting categories for results (a mix of ClinicalTrials.gov only, PubMed only, or both).
Of 214 trials studying 55 FDA-approved drugs, 184 (86.0%) had results available, including 64 (29.9%) in ClinicalTrials.gov only. Of 115 trials studying 31 unapproved drugs, 70 (60.9%) had results publicly available, including 15 (13.0%) with results in ClinicalTrials.gov only.
Discussion
ClinicalTrials.gov was a unique source of results for nearly one-quarter of sampled drug trials and more than one-tenth of sampled sponsor-drug-condition trial sets. However, results remain unavailable in ClinicalTrials.gov or PubMed 7 or more years after study completion for nearly one-quarter of sampled drug trials and more than one-tenth of sampled sponsor-drug-condition trial sets. This study is limited by the convenience sampling approach, omission of unregistered trials, and lack of assessment of reporting content. Availability of results in ClinicalTrials.gov, particularly before or in lieu of any publication, adds critical information to the evidence base. However, this advance in access to trial results only accrues if ClinicalTrials.gov is used; for example, systematic reviewers who rely entirely on literature searches are at risk of missing relevant evidence.6 Despite advances in making trial results publicly available, it is critical to continue identifying and ameliorating gaps in results reporting.
References
- 1.Anderson ML, Chiswell K, Peterson ED, Tasneem A, Topping J, Califf RM. Compliance with results reporting at ClinicalTrials.gov. N Engl J Med. 2015;372(11):1031-1039. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results database: update and key issues. N Engl J Med. 2011;364(9):852-860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med. 2008;358(3):252-260. [DOI] [PubMed] [Google Scholar]
- 4.Ross JS, Tse T, Zarin DA, Xu H, Zhou L, Krumholz HM. Publication of NIH funded trials registered in ClinicalTrials.gov: cross sectional analysis. BMJ. 2012;344:d7292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Riveros C, Dechartres A, Perrodeau E, Haneef R, Boutron I, Ravaud P. Timing and completeness of trial results posted at ClinicalTrials.gov and published in journals. PLoS Med. 2013;10(12):e1001566. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Baudard M, Yavchitz A, Ravaud P, Perrodeau E, Boutron I. Impact of searching clinical trial registries in systematic reviews of pharmaceutical treatments: methodological systematic review and reanalysis of meta-analyses. BMJ. 2017;356:j448. [DOI] [PMC free article] [PubMed] [Google Scholar]