Abstract
This study summarizes responses to trial registration fields on the ClinicalTrials.gov website asking users if they plan to share individual participant data and what information they would be willing to share.
The US Institute of Medicine and the International Committee of Medical Journal Editors (ICMJE), among others, have called for clinical trialists to declare their plans to share individual participant data (IPD) and related study documents for secondary research prior to study initiation. Individual participant data, which refers to measurements collected from each research participant, is distinct from aggregated data usually reported in journal articles or in the ClinicalTrials.gov results database. The term IPD sharing refers to the practice of making clinical trial data at the individual level available to researchers who were not part of the original study team.
In December 2015, ClinicalTrials.gov added 2 optional registration fields for disclosing trialists’ intentions about IPD sharing: (1) plan to share IPD? and (2) available IPD/information type (Box). In this exploratory study, we characterized responses to both fields to inform the ongoing discussion.
Box. Description of the 2 ClinicalTrials.gov Individual Participant Data (IPD) Sharing-Related Optional Registration Fields.
1. Plan to Share IPD (generally provided at study initiation)
Indicate whether there is a plan to make IPD collected in this study, including data dictionaries, available to other researchers (typically after the end of the study). Options include: Yes: There is a plan to make IPD and related data dictionaries available; No: There is not a plan to make IPD available; Undecided: It is not yet known if there will be a plan to make IPD available.
IPD Sharing Plan Description
If response to the plan to share IPD field is “yes,” briefly describe what specific individual participant data sets are to be shared (for example, all collected IPD, all IPD that underlie results in a publication). If response to the plan to share IPD field is “no” or “undecided,” an explanation may be provided for why IPD will not be shared or why it is not yet decided.
2. Available IPD/Information Type (generally provided after study completion)
The type of data set or supporting information being shared. Options include: individual participant data set, study protocol, statistical analysis plan, informed consent form, clinical study report, analytic code, or other (specify)
Methods
We summarized and examined responses from both ClinicalTrials.gov IPD-related fields for all interventional studies registered from January 2016 to August 2017. We also extracted a nonrandom convenience sample of unique free-text plan descriptions from records indicating plans to share IPD for a qualitative review and to determine if any misunderstanding of the definition of IPD sharing was present.
For trial records from high-volume registrants (registered ≥10 records), we characterized disclosed plans to share IPD. We further characterized responses from the top-10 high-volume registrants within each key funder type. Key funder types were categorized as “NIH” if at least 1 National Institutes of Health (NIH) institute or center was listed as the lead sponsor or collaborator, as “industry” if not classified as NIH and at least 1 company was listed as the lead sponsor or collaborator, and as “other” for all remaining records (eg, funded by foundations).
Results
Of 35 621 trial records analyzed on August 31, 2017, 25 551 (72%) responded to the Plan to Share IPD field; 2782 records (10.9%) indicated yes, 6452 (25.3%) indicated undecided, and 16 317 (63.9%) indicated no. Of records indicating plans to share IPD, we reviewed 154 free-text descriptions of sharing plans and identified at least 5 types of misunderstanding of IPD sharing (eg, suggesting that sharing would be done through publication or posting on trial registers, that data would be provided to the participants, or that data would be shared with a limited number of groups, etc).
Of the 35 621 records analyzed, 20 842 (58.5%) were from 693 high-volume registrants; 14 040 (67%) included a response to the plan to share IPD field, with 5% indicating yes, 16% undecided, and 50% no (Table). Proportions varied by key funder type, from 11% of NIH funders to 0% of industry answering yes. Among the top-10 high-volume registrants by key funder type, the median for percentage of records indicating existence of IPD sharing plans was 7% (range, 0%-18%) for NIH-funded studies, 0% (range, 0%-27%) for industry-funded studies, and 5% (range, 0%-14%) for studies with other funding.
Table. Responses Among High-Volume Registrants to the Plan to Share Individual Participant Data Trial Registration Fielda.
| Summary Statistics | Total | Key Funder Type | ||
|---|---|---|---|---|
| NIH | Industry | Other | ||
| Trial records, No. | 20 842 | 1151 | 3738 | 15 953 |
| ≥1 Response, No. (%) | 14 040 (67) | 515 (45) | 1485 (40) | 12 040 (75) |
| Response, median (range), % | ||||
| Yes | 5 (0-86) | 11 (0-35) | 0 (0-86) | 6 (0-69) |
| Undecided | 16 (0-100) | 12 (0-61) | 7 (0-100) | 18 (0-91) |
| No | 50 (0-100) | 44 (0-72) | 39 (0-100) | 51 (0-100) |
| ≥1 Missing response, No. (%) | 6802 (33) | 636 (55) | 2253 (60) | 3913 (25) |
| No response, median (range), % | 21 (1-100) | 34 (6-100) | 37 (3-100) | 20 (1-100) |
Abbreviation: NIH, National Institutes of Health.
High-volume registrants defined as those that registered 10 or more trial records.
Among 212 trial records (of 35 621) listing entries in the available IPD/information type field, 19 specified that IPD would be available for sharing. Of these 19 records, 7 had responded no to the plan to share IPD registration field, 6 had indicated that IPD sharing was planned, 5 had indicated undecided, and 1 had not responded.
Discussion
Although 36.2% of 25 551 trial records indicated they would share data or were undecided, several descriptions of IPD sharing plans reflected confusion or uncertainty about the term IPD and the meaning of the term sharing. Heterogeneity in responses among top registrants, consistent with findings among industry registrants, may suggest inconsistent institutional-level IPD sharing policies. Limitations include analysis of responses only 20 months after introduction of IPD-related fields on ClinicalTrials.gov and the lack of systematic analysis of sharing plan descriptions.
Based on this early experience, ClinicalTrials.gov added additional subfields with greater structure in late June 2017 to facilitate more complete disclosure of IPD-sharing plans as required by the ICMJE. However, considerable educational, cultural, and scientific changes will be necessary before the sharing of IPD and associated documents is well understood.
Section Editor: Jody W. Zylke, MD, Deputy Editor.
References
- 1.National Academies Press Sharing clinical trial data: maximizing benefits, minimizing risks. https://www.ncbi.nlm.nih.gov/books/NBK269030/. Accessed December 8, 2017. [PubMed]
- 2.Taichman DB, Sahni P, Pinborg A, et al. . Data sharing statements for clinical trials: a requirement of the International Committee of Medical Journal Editors. JAMA. 2017;317(24):2491-2492. [DOI] [PubMed] [Google Scholar]
- 3.Bauchner H, Golub RM, Fontanarosa PB. Data sharing: an ethical and scientific imperative. JAMA. 2016;315(12):1237-1239. [DOI] [PubMed] [Google Scholar]
- 4.Zarin DA, Tse T. Sharing individual participant data (IPD) within the context of the Trial Reporting System (TRS). PLoS Med. 2016;13(1):e1001946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Boutron I, Dechartres A, Baron G, Li J, Ravaud P. Sharing of data from industry-funded registered clinical trials. JAMA. 2016;315(24):2729-2730. [DOI] [PubMed] [Google Scholar]
- 6.US National Library of Medicine ClinicalTrials.gov protocol registration data element definitions for interventional and observational studies. https://prsinfo.clinicaltrials.gov/definitions.html#IPDSharingLabel. Accessed November 3, 2017.