Figure 1 Concerns of authors regarding a potential Journal of the Medical Library Association data sharing policy
Of the 15 authors who responded to our email, 7 expressed no concerns, whereas 8 expressed 1 or more concern about data sharing.
Kevin B. Read, Liz Amos, Lisa M. Federer, Ayaba Logan, T. Scott Plutchak, Katherine G. Akers
doi: http://dx.doi.org/10.5195/jmla.2018.431
Received January 2018: Accepted January 2018
ABSTRACT
Providing access to the data underlying research results in published literature allows others to reproduce those results or analyze the data in new ways. Health sciences librarians and information professionals have long been advocates of data sharing. It is time for us to practice what we preach and share the data associated with our published research. This editorial describes the activity of a working group charged with developing a research data sharing policy for the Journal of the Medical Library Association.
The sharing of research data underlying scholarly literature is essential for research transparency. Health sciences librarians and information professionals strongly believe in the openness of information and have long been advocates of data sharing [1–3]. Members of our community spend considerable time developing and implementing services and policies to encourage researchers to share their data [4]. Among the many reasons for data sharing, access to research data can improve the integrity of the research process by allowing others to reproduce research results or analyze the data in new ways [5]. Some peer-reviewed journals such as PLOS ONE [6] have taken the lead in requiring authors to share their data if they want to publish in these journals, whereas other journals like the New England Journal of Medicine have yielded to their research communities’ concerns around data sharing and released less stringent data sharing guidelines [7]. The latter has justifiably disappointed many open data advocates who wish to see biomedical research improved by greater transparency and access to research data [8, 9].
To date, the Journal of the Medical Library Association (JMLA) has not enacted policies related to sharing research data. However, as discussions around data in health sciences librarianship continue to build and more journals require data sharing as a precondition for publishing, it is time for librarians to practice what we preach and share the data and documentation associated with our published research. As we encourage our user communities to comply with data sharing practices, we must adhere to these practices as well.
In May 2017, the editor-in-chief of the JMLA tasked a working group with developing a data sharing policy for the journal. To develop this data sharing policy, the working group (1) contacted authors of articles recently published in the JMLA to assess their willingness or ability to share their data and to understand their concerns about data sharing, (2) reviewed data sharing policies of other peer-reviewed journals, and (3) is working to align JMLA policy with recommendations of the Research Data Alliance (RDA) Data Policy Standardisation and Implementation Working Group [10]. The group’s goal is to develop a data sharing policy that promotes the rigor and reproducibility of research described by articles published in the JMLA, while also imposing a reasonably minimal burden on authors, peer reviewers, and editorial team members.
To inform the development of a JMLA data sharing policy, the group contacted twenty-one authors of research articles or case studies published in the JMLA in 2016. After explaining our intention to develop a JMLA data sharing policy, we asked the authors if they would send our working group the data that would be needed to reproduce the results described in their articles. If authors could not or chose not to share the data with us, we asked them to provide an explanation. Finally, we asked authors if they had concerns about the implementation of a JMLA data sharing policy. In several cases we had to search, sometimes unsuccessfully, for authors’ new email addresses if we received notice that they had left their former institutions. Nonresponses to our initial email request were followed up by a reminder email.
We heard back from 15 authors, 11 of whom sent us partial or complete datasets, sometimes accompanied by supporting documentation or computer code. Responding authors who did not send us their data gave the following reasons: the protocol approved by their institutional review board (IRB) specifically forbade data sharing (n=1), the statistician who managed the data had left the institution (n=1), data could not be located (n = 1), and article coauthors were not comfortable with sharing the data (n=1). We note that had there been a journal data sharing policy in place at the time these manuscripts were planned, prepared, or submitted, at least some of these barriers to data sharing might not have existed. For instance, authors’ plans to share data could have been approved by the IRB before the study took place, and data could have been submitted to a trustworthy repository and might not have been lost.
Of the authors who responded to our email, approximately half did not express any concerns regarding a potential JMLA data sharing policy (Figure 1), and many (n=8) explicitly voiced support for such a policy. Among those who expressed concerns about a journal data sharing policy, the most common concern was how the policy would address sharing data that contained identifying or personal information. Some of these authors stated that sharing data would require modifications to their IRB protocols, whereas others worried about the lack of standards or procedures in the library and information science field for determining what and how identifying information should be removed. Other authors were concerned about data ownership and intellectual property rights: Would authors still retain rights to the data? How would access be controlled? Would the policy address what other people would be allowed to do with an author’s data? Some authors pointed out that data sharing would be difficult or impossible in cases in which some data were proprietary or included copyrighted materials. Finally, some authors asked how data would need to be formatted or described to be usable or understandable by others, wondered where authors without institutional repositories were to deposit their data, noted the not negligible time needed to prepare a dataset and its documentation for sharing, and wanted clarification on what we meant by “data.”
|
|
||
|
Figure 1 Concerns of authors regarding a potential Journal of the Medical Library Association data sharing policy |
||
Indeed, “data” can conceivably encompass a great number of different types of research products, including spreadsheets, text files, interview recordings and transcripts, survey and assessment instruments, images, videos, and computer codes and scripts, as well as documents describing the data or how it was collected (e.g., study protocols, data dictionaries, codebooks, readme files). These concerns have been duly noted by our working group and are being taken into serious consideration as we develop the policy. Each issue raised by the authors provides valuable information that is helping shape policy details and language.
To familiarize ourselves with the language used in other peer-reviewed journals’ data sharing policies, we reviewed policies addressing data accessibility from journals published by PLOS [6], Gigascience [11], Science [12], journals published by the Nature Publishing Group [13], and Nature Publishing Group’s Scientific Data in particular [14]. The goal of this exercise was to assess different journal policies ranging from those that firmly require data sharing in order to publish (i.e., PLOS ONE) to those that only ask for a data availability statement (i.e., Nature). For each policy, we focused on the journal’s definition of data, acceptable methods for data sharing, permitted reasons for not sharing data, and any required technical specifications (e.g., file formats, documentation). These details have provided valuable insight into how to structure and word the data sharing policy for the JMLA.
To ensure that the JMLA data sharing policy aligns with larger efforts in scholarly publishing, we have kept abreast of the conversations of the RDA Data Policy Standardisation and Implementation Working Group, which includes representatives from several major publishers and publishing organizations and aims to establish common frameworks for journal data policies that can be used by journals across disciplines [10]. As we align our policy with RDA guidelines, we will be asking JMLA authors to adhere to policies consistent with those of other journals to which they may be submitting.
After gathering feedback from JMLA authors, reviewing existing journal data sharing policies, and being informed by RDA Data Policy Standardisation and Implementation Working Group discussions, we created a list of several questions that our working group must answer to develop a JMLA data sharing policy that not only compels authors to share the data underlying their research results, but also addresses common data sharing concerns held by researchers and practitioners in our field. These questions pertain to what our definition of data is, what types of articles to which the policy will apply, where data should be deposited, when data should be made available, what acceptable reasons are for restricting access to data, and what Frequently Asked Questions or other guidance we should provide to help authors comply with the policy. We also plan to converse with and solicit feedback from members of the JMLA Editorial Board, Medical Library Association (MLA) Data Special Interest Group, MLA Research Section, and MLA Board of Directors. Our goal is to officially implement a JMLA data sharing policy in 2018.
We hope that health sciences librarians will be excited by the prospect of sharing the data associated with their published articles in accordance with the new JMLA data sharing policy and leading by example for their user communities. As the corpus of health sciences library and information research data grows under this policy, so will the integrity of our research processes and findings. The more we share data and encourage others to do the same, the better chance we have of increasing collaboration within and outside our domain, making new discoveries, and becoming a leading field in research data sharing practices. Furthermore, by experiencing the process of preparing our own data to be shared, we can learn more about the challenges that our users face when asked to share their data. This newfound knowledge will make us better educators, advocates, and guides for improving the sharing of biomedical data by researchers at our own institutions. To communicate your ideas for or concerns with a JMLA data sharing policy, please contact JMLA Editor-in-Chief Katherine Akers, at jmla@journals.pitt.edu.
We thank the authors of JMLA articles who shared their data or concerns about data sharing with our data sharing policy working group. The contributions to this work by Liz Amos were provided with the support of the National Library of Medicine (NLM), National Institutes of Health.
1 Borgman CL, Wallis JC, Enyedy N. Little science confronts the data deluge: habitat ecology, embedded sensor networks, and digital libraries. Int J Digit Libr. 2007 Oct;7(1):17–30.
2 Cox AM, Pinfield S. Research data management and libraries: current activities and future priorities. J Librariansh Inf Sci. 2014 Dec;46(4):299–316.
3 Wallis JC, Rolando E, Borgman CL. If we share data, will anyone use them? data sharing and reuse in the long tail of science and technology. PLOS ONE. 2013;8(7):e67332.
4 Higman R, Pinfield S. Research data management and openness: the role of data sharing in developing institutional policies and practices. Program. 2015;49(4):364–81.
5 Warren E. Strengthening research through data sharing. N Engl J Med. 2016 Aug 6;375(5):401–3.
6 PLOS ONE. Data availability [Internet]. 3 Mar 2014 [cited 14 Dec 2017]. <http://journals.plos.org/plosone/s/data-availability>.
7 Taichman DB, Sahni P, Pinborg A, Peiperl L, Laine C, James A, Hong ST, Haileamlak A, Gollogly L, Godlee F, Frizelle FA, Florenzano F, Drazen JM, Bauchner H, Baethge C, Backus J. Data sharing statements for clinical trials — a requirement of the International Committee of Medical Journal Editors. N Engl J Med. 2017 Jun 8;376(23):2277–9.
8 Empty rhetoric over data sharing slows science. Nature. 2017 Jun 12;546(7658):327.
9 Marcus A, Oransky I. New science data-sharing rules are two scoops of disappointment [Internet]. STAT News. 6 Jun 2017 [cited 14 Dec 2017]. <https://www.statnews.com/2017/06/06/data-sharing-rules-disappoint>.
10 Research Data Alliance. Data policy standardisation and implementation [Internet]. The Alliance; 2017 [cited 14 Dec 2017]. <https://www.rd-alliance.org/groups/data-policy-standardisation-and-implementation>.
11 Gigascience. Instructions to authors [Internet]. 2016 [cited 14 Dec 2017]. <https://academic.oup.com/gigascience/pages/instructions_to_authors>.
12 Science. Science journals: editorial policies [Internet]. 2015 [cited 14 Dec 2017]. <http://www.sciencemag.org/authors/science-editorial-policies>.
13 Springer Nature. Data availability statements—guidance for authors and editors [Internet]. 2016 [cited 14 Dec 2017]. <http://www.springernature.com/gp/authors/research-data-policy/data-availability-statements/12330880?countryChanged=true>.
14 Scientific Data. Data policies [Internet]. 2017 [cited 14 Dec 2017]. <https://www.nature.com/sdata/policies/data-policies>.
Kevin B. Read, MLIS, MAS, kevin.read@nyumc.org, https://orcid.org/0000-0002-7511-9036, Data Services Librarian and Data Discovery Lead, NYU Health Sciences Library, New York University School of Medicine, 577 First Avenue, New York, NY 10016
Liz Amos, MLIS, liz.amos@nih.gov, Librarian, National Information Center on Health Services Research and Health Care Technology, National Library of Medicine, Bethesda, MD
Lisa M. Federer, MLIS, MA, AHIP, lisa.federer@nih.gov, https://orcid.org/0000-0001-5732-5285, Research Data Informationist, NIH Library, National Institutes of Health, Bethesda, MD
Ayaba Logan, MLIS, MPH, loganay@musc.edu, https://orcid.org/0000-0002-7430-6358, Research and Education Informationist, Libraries, Medical University of South Carolina, Charleston, SC
T. Scott Plutchak, MA, AHIP, FMLA, tscott@uab.edu, https://orcid.org/0000-0003-4712-5233, Director of Digital Data Curation Strategies, Lister Hill Library of the Health Sciences, University of Alabama, Birmingham, AL
Katherine G. Akers, PhD, JMLA@journals.pitt.edu, http://orcid.org/0000-0002-4578-6575, Editor-in-Chief, Journal of the Medical Library Association
Articles in this journal are licensed under a Creative Commons Attribution 4.0 International License.
This journal is published by the University Library System of the University of Pittsburgh as part of its D-Scribe Digital Publishing Program and is cosponsored by the University of Pittsburgh Press.
Journal of the Medical Library Association, VOLUME 106, NUMBER 2, March 2018