Many academic researchers who use preexisting data to conduct research describe a common experience: waiting for university officials to finalize and sign contracts necessary to transfer the data. These data use agreements (DUAs) detail the terms under which data will be disclosed, transferred, stored, and used, specifying rights and obligations for both the data supplier and the recipient (1). Faculty members often struggle to understand why DUAs for transfers of seemingly low-risk data take so long to conclude. To understand reasons for delays and explore what might be done to streamline the process, we interviewed a sample of university officials responsible for negotiating DUAs. This first empirical investigation of the DUA process found that procedural inefficiencies, incomplete information, data suppliers' lack of incentives and familiarity with academic practices, and faculty unresponsiveness may be more important contributors to delays than the lawyers being at loggerheads over their respective positions. Although researchers may view DUAs as another symptom that research has become “overlawyered,” our study suggests that they may underappreciate the importance of these contracts and the complexity of negotiating them.
DUAs typically require institutional review, negotiation, and signoff because of the obligations and potential risks they impose on the university (including risks of civil liability, regulatory noncompliance, reputational damage, and compromised research principles). The review, usually conducted by attorneys or contract administrators within the university, may be intensive and prolonged. Agreements concerning data transfers from private companies may be particularly difficult because companies may not share the university's norms concerning open science, freedom of publication, intellectual property (IP) ownership, and human-subjects protection.
Delays in executing DUAs may frustrate grant deadlines, trainees' needs, and researchers' ability to respond to current events with timely research. Combined with other regulatory requirements, these delays contribute to a climate in which those responsible for research oversight are perceived as obstructing the progress of science.
Complaints, Barriers, Solutions
Interviews were completed with 59 individuals representing 48 of the top-50 U.S. universities in terms of total research expenditures in 2015 [all 50 were approached; (2)] (see supplementary materials for further details on methods, data, and results). Interviews focused on contracts that governed transfers of data to university researchers from another U.S. entity and were not pure nondisclosure agreements, agreements for transfer of tangible materials, or clinical trial agreements. Two-thirds of the interviewees were based at public institutions. Nearly all respondents dealt with all types of data suppliers, or “counterparties” (academic, industry, government, nonprofit), but 9% did not negotiate industry contracts.
Complaints heard from faculty
Contract negotiators from 83% of institutions said that they had heard complaints from faculty about the DUA execution process. The gist of these complaints was that getting DUAs signed interposed an excessive and unnecessary delay in getting research started. Some faculty reportedly viewed DUAs as unnecessary, questioning why “good faith and the honor system” cannot suffice. Negotiators connected this view to a lack of understanding of why certain contractual provisions are problematic and how “the entire university community can be affected if something is not done properly,” as well as faculty members' misplaced confidence that the university can accept such terms because “nothing's going to go wrong.” Negotiators reported that some faculty complained that the process involves “too much paperwork” for a “simple agreement” and wonder, “Why isn't there just one template for everything?” Some faculty reportedly lamented the “bureaucratic hot potato” involved in obtaining signoff from multiple university offices, which may include the institutional review board (IRB), privacy, information technology/security (IT), clinical trials, compliance, technology transfer, and sponsored research. Negotiators perceived that faculty had unrealistic expectations that reflected a lack of appreciation of the negotiators' workload (see table S2 in supplementary materials).
Process factors contributing to delay
Among nine process factors that may contribute to delays, contract negotiators identified their workload as the most common and serious problem (see fig. S1). Many negotiators commented that other types of contracts took priority over DUAs, and several observed that the volume of DUAs was increasing as organizations were becoming “more careful” about privacy. The next most common problem was delays on the part of faculty in responding to questions or requests, which was also characterized as difficult to overcome. Delays due to the need to comply with university policies (e.g., information security) were characterized as a serious or moderate problem when they did occur by about half the respondents, but reportedly occurred less frequently than delays due to faculty dilatory responsiveness. The problem typically arose from the time needed to consult the IT or privacy office or modify IT architecture. How responsibility for DUAs was allocated across university offices was another process factor with relatively low frequency but high seriousness. Faculty sometimes submitted requests to the wrong office, and it was often unclear which offices needed to be involved because faculty provided insufficient information. Lack of incentives on the part of the data supplier to reach agreement also did not frequently contribute to delays but was a serious problem when it did. This problem generally arose in negotiations with industry because companies did not view sharing their data with researchers as a “revenue-generating stream” or share the university's research mission.
Troublesome contractual provisions
Among seven specific DUA provisions, information privacy and security provisions (including, for example, prohibitions on reidentifying individuals and disclosing data to third parties and requirements for how data must be stored) stood out as both a frequently contested issue in negotiations and a serious problem in getting DUAs executed (see the figure). Negotiators described having to consult the privacy office to determine the classification of the data covered under the Health Insurance Portability and Accountability Act (HIPAA); consulting IT about whether the researcher's IT security satisfied the data supplier's terms; and where it did not, having to build a custom IT solution or renegotiate those terms. They perceived the security measures demanded by data suppliers to be excessive in many cases.
The seriousness question was not asked of respondents who answered that the provision was “Never” a problem. Denominator for proportions is the number of nonmissing responses.
GRAPHIC: N. DESAI/SCIENCE
Indemnification provisions, which require the university to pay liability costs incurred by the data supplier arising from the researcher's use of the data, also frequently arose as a negotiating issue but were somewhat easier to resolve. Public universities were constrained by state law in what they could accept. Some private universities acceded to the data supplier's demands; others engaged in protracted negotiations to obtain the narrowest possible language.
Negotiations over the definition of confidential information were fairly common but relatively easy to resolve; efforts focused on ensuring that it excluded research results. Negotiations over ownership of faculty members' work product were less common but relatively difficult to resolve. These disputes arose mostly with industry data suppliers and involved assertions of ownership rights over inventions and research results that were unacceptable because they conflicted with university policy. Similar issues arose with provisions restricting the researcher's right to publish.
Recommendations for streamlining
Negotiators offered several suggestions for improving the DUA execution process (see the box). Those who emphasized increasing staffing in the office with primary responsibility for DUA review reasoned that efficiency improvements alone will be inadequate, given the rapid growth in DUA volume. Among those who emphasized the value of a universal DUA template, a few mentioned that they use and like the template being pilot tested by the Federal Demonstration Partnership (FDP) with the goal of improving quality and consistency in the DUA process (3, 4). FDP is a consortium of 90 research institutions and 10 federal agencies that collaborates to identify ways to improve the efficiency of research; initial efforts have focused on federal contracts.
Options for Action by Universities
The contributors to delay that we identified fall into three broad categories: bureaucratic problems, incentives gaps, and lawyering issues. Although faculty complaints about delays often appear to conflate overlawyering with bureaucratic problems, the latter presents several distinct reform options.
First, given the increasing use of DUAs, staffing in the offices responsible for DUA negotiation has lagged behind the growth in demand for these services. Second, interoffice coordination problems contribute to inefficiency. Negotiators had little appetite for centralizing review in one office but perceived opportunities to coordinate better with others around the university, such as IT and the IRB; they recommended creating clear workflows from office to office and making faculty aware of them. Third, there was a strong felt need to strengthen communication with faculty at the initial intake phase. Well-designed, online submission portals could substantially reduce the amount of back-and-forth required to obtain missing information.
Even if universities can successfully mitigate these bureaucratic issues, another issue will continue to cause delays: the gap in incentives between the researchers who seek data and those from whom data are sought. Although academic researchers have strong reasons to pursue their work expeditiously, data suppliers often have weak incentives to share their data. University negotiators felt they were at the mercy of counterparties for whom completing the DUA, much less accepting compromises, was not a priority. This incentive gap sometimes slowed the negotiation process and forced universities to yield on substantive provisions.
The third set of issues relates to clashes between DUA negotiators over what is and is not acceptable in the contract. Negotiators reported that the most common and serious of these substantive issues related to provisions concerning information privacy and security, indemnification, and the definition of confidential information; provisions concerning publication rights and ownership of academic researchers' work product were less commonly in dispute but serious problems when they were. These are no mere matters of “legalese”; each implicates potentially important risks to the university and faculty member.
Some of these points of tension appear amenable to efficiency gains. For example, negotiations over privacy provisions may be streamlined by use of a robust intake system that elicits key information from faculty and negotiator “playbooks” that list essential, preferred, and acceptable contract wording and rationales. These measures can quickly determine whether particular actions (for example, designating the data as a Limited Data Set under HIPAA) are appropriate. In our sample, 85% of negotiators used playbooks or other written tools; 78% found them useful and 22% somewhat useful.
Indemnification is another actionable area. At least where low-risk data are involved, university contract negotiators may be spending more time on these provisions than is warranted. If good privacy and security protections are in place, the risk of a data breach is low, and haggling over who pays in the unlikely event of a breach that causes harm should not obstruct timely data transfers for research. Yet, negotiators at 13 of 48 universities had walked away from a negotiation because of indemnification issues.
When it comes to provisions safeguarding publication rights and ownership of faculty members' work product, on the other hand, universities must remain resolute. These provisions implicate core values of the university and of open science. A potential strategy for minimizing haggling over non-negotiable issues is for universities as a group to more clearly signal their unified position. Existing university policies setting forth institutional commitments to academic freedom and policies concerning IP are helpful in communicating norms, but even more helpful would be a universal DUA template.
Onerous data security requests by suppliers pose a thornier challenge. A universal template could help in this area by educating counterparties about customary standards for transferring and storing data of different levels of sensitivity. Efforts by universities to centralize and standardize data protection infrastructure and procedures could also reduce DUA-by-DUA negotiation of data security arrangements.
For reasons outlined above, the ongoing FDP template project is a promising development. The first template, currently in the pilot phase, does not cover all types of data or relationships but does represent careful thinking about de-identified data (4). It consists of a basic agreement and several attachments that can be added for different types of data and situations. The initiative demonstrates the feasibility of crafting DUA language that is acceptable to a broad range of universities. Negotiators in our sample who were participating in the pilot felt that the template should be adopted by other institutions and likened it to the Uniform Biological Material Transfer Agreement (UBMTA), a successful initiative by the U.S. National Institutes of Health (NIH) to disseminate a template MTA (5–7). Other groups also have developed template agreements for data sharing that provide useful models (8, 9).
Agreeing on language concerning higher-risk data likely will be more difficult than crafting the initial FDP template, as will engaging private companies and nonfederal units of government. Moreover, the FDP is a content-focused intervention, and our findings show that clashes over DUA provisions are nested within an array of environmental and process factors, many of them internal to universities.
Recommendations
Contract Negotiators' Suggestions for Streamlining the DUA Execution Process
Improve the intake process (n = 16)
-
Ensure that faculty provide better information on the nature of the data and research.
-
Create an electronic submission system similar to those used for IRB and grant applications.
-
Designate one office as the point of entry into the review process; route to others as needed.
Improve coordination across reviewing offices within the university (n = 12)
-
Have IT office review information security provisions before DUAs are routed for general legal review.
-
Create a “campus map” of all offices that must review a DUA and lay out “an end-to-end process.”
Increase staffing in the office primarily responsible for DUA reviews (n = 11)
-
Calibrate staffing to increasing volume of DUAs.
-
Compare staffing and volume across university offices that review different types of contracts.
-
Hire specialists in data privacy issues or more reviewers whose sole focus is DUAs.
Pursue a universal DUA template (n = 5)
Educate faculty about the complexity and importance of DUAs (n = 5).
-
Elevate visibility of the issue, comparable to training on HIPAA and human subjects protection.
-
Educate as to why DUA negotiation is not always straightforward.
Our study has limitations. Our sample consisted of research-intensive U.S. universities. Universities in other countries may face different challenges owing to different privacy laws, and universities with less research may have lower demand for DUAs but also fewer resources. Our sample size was too small to permit subgroup analyses. Some interview responses may be subject to socially desirable response bias. Finally, in most cases we interviewed only one negotiator per university. Different information might have been obtained by interviewing additional negotiators or more senior officials.
DUAs protect researchers' publication and IP rights and minimize universities' risks of civil liability, fines, and reputational harm from data breaches and violations of state and federal law. There is value in crafting them carefully. Balanced against this imperative, however, is the need to ensure that they do not unduly delay research. Long waits for data are already a fact of life, and the problem will only grow with the inexorable increase in the volume and complexity of data exchanges. New data-sharing requirements for federal grantees will likely speed this increase as researchers are increasingly asked to transmit data to others (4, 10–12). To preserve key interests of universities while avoiding delays that test researchers' patience and impede scientific progress, reforms will need to address the full range of contributing factors.
Supplementary Materials
Acknowledgments: This work was supported by a grant from the Alfred P. Sloan Foundation (grant no. G-2016-7232). The sponsor had no role in the design or execution of the study or the interpretation of data. Data were managed using REDCap electronic data capture tools hosted at Stanford University School of Medicine and supported by Stanford Clinical and Translational Science award no. UL1 TR001085 from NIH–National Center for Research Resources.
"waiting" - Google News
January 10, 2020 at 01:46AM
https://ift.tt/2QGjs6x
Waiting for data: Barriers to executing data use agreements - Science Magazine
"waiting" - Google News
https://ift.tt/2oFixIf
Shoes Man Tutorial
Pos News Update
Meme Update
Korean Entertainment News
Japan News Update
No comments:
Post a Comment