I.3.3 Professional Commentary

Statements published by professional researchers show strong support for open policies on data access. Such statements include the following:

• A "guest opinion" by Chief of Surgery at the National Cancer Institute recommended that "[i]nstitutions that fund research and journals that publish the results of research should insist on the open sharing of information and reagents since lack of cooperation is at odds with their long-term goals."²⁶ The author observed that historically the failure to share research results, whether motivated by intellectual property or individual privacy concerns, has impeded medical progress and resulted in needless deaths.

• An article by three authors in the "Sounding Board" section of the New England Journal of Medicine in 1986, expressed the following position on the need for underlying data access in connection with regulatory decision making ²⁷ :

  We recommend that when a government-initiated study leads to changes in health policy or regulations, data supporting such decisions should be shared with all interested parties. Access to these data should not depend on whether records are in the hands of the government or a private research contractor, or whether a products-liability lawsuit has been filed. Although peer review by journals provides a check on the appropriateness and quality of research methods and analyses, that process applies only to materials selected by the author for presentation to the journal. It is not a substitute for having access to all the data. Therefore, when the federal government initiates research that is intended to provide the basis of a regulatory decision - e.g., research on Reye's syndrome, toxic shock syndrome, Agent Orange, and artificial sweeteners - the granting of access to data should be a condition of government support, and procedures for the release of data should be made part of the study. These procedures should grant any interested party (e.g., litigants, industry representatives, or public-interest groups) access to data so that they may review them at their own expense and with their own expert medical consultants. To protect the quality and integrity of the research, data should be shared only after the report of the results has been accepted for publication. However, data should be shared before publication if the government makes or proposes to make a regulatory or health policy decision based on research results before their publication.

• The "Bellagio Principles" evolved from a conference of government officials and scholars at which the United States, the United Kingdom, West Germany, Sweden and Canada were represented.²⁸ The Principles, which have been published in at least five journals, endorse the provision of government data to individual researchers and research institutions for legitimate research purposes.

  Specific Principles pertaining to data access include:

  (a) "National statistical offices should provide researcher both inside and outside government with the broadest practicable access to information within the bounds of accepted norms of privacy and legal requirements to preserve confidentiality."

  (b) "In making data available to researchers, national statistical offices should provide some means to ensure that decisions on selective access are subject to independent review and appeals."

  (c) "There are legitimate research purposes requiring the use of individual data for which public use samples are inadequate."

  (d) "There are legitimate research uses which require the utilization of identifiable data within the framework of concern for confidentiality."

  (e) "Some research and statistical activities require the linking of individual data for research and statistical purposes. The methods that have been developed to permit record linkage without violating law or social custom regarding privacy should be used whenever possible."

• At a conference convened in London in 1995, nineteen scientists developed a set of recommended principles for the use of epidemiologic data in regulatory risk assessments. Their recommendations, and related discussion, included the following ²⁹ :

  R-1. A commitment to collaboration should be made by epidemiologists and risk assessors that includes (a) sharing of raw data where feasible, (b) exchange of protocols and survey instruments, (c) inclusion of epidemiologists in dose-response modeling exercises, and (d) care and fairness by risk assessors in the critique of original epidemiological studies.

  * * *

  The conferees regarded the data-sharing aspect of the recommendation as an important element of accurately interpreting study findings and deriving the most value from the research. However, they recognized that sharing data raises a number of complex issues. There are legal and ethical issues concerning confidentiality involved in revealing analytical data pertaining to individual study subjects; on the other hand, it might be possible to avoid such issues in large part by releasing only certain types of coding data, or by asking the original investigator to re-analyze or perform additional analysis on the original data set. A further problem is that many research organizations have records retention and destruction policies that require destruction of much study data after just a few years. The conferees strongly recommended that Federal Focus or some other entity explore the entire issue of data sharing in one or more future conferences. In general, the conferees felt that epidemiologists have to become more willing to have the full details of their studies opened up to critiquing. Toxicologists had to change in this way. At one time, they did not have to make their slides available; now they do. Making it possible for a peer reviewer to tell an epidemiologist "I think you measured this exposure incorrectly" is no different from making it possible to tell a toxicologist "I think you misinterpreted this slide."

  * * *

  1. Risk assessors' criticisms and major questions about the methods, analyses, data, or interpretation of a published report should be directed, whenever possible, to the epidemiologist(s) responsible for the paper, and they should be given an opportunity to respond.

  Often study reports will not describe all of the analyses that were done; therefore, before criticizing a study for lack of analysis on a particular point, the reviewers and risk assessors should ask the investigators whether it was done and, if so, how. Although it is not reflected in the principle, some participants commented that the principle should work both ways: The investigators should be encouraged to cooperate in answering any questions regarding analysis or data not reflected in the published report. This point is dealt with to a limited extent in the next principle.

  2. Risk assessors may want access to the study's data set for other analyses to be used in the risk assessment. This would be done with the consent and cooperation of the study epidemiologist(s).

  If access to underlying study data is available, epidemiologists other than the original investigator can sometimes review and analyze them and discern strengths or weaknesses that were not apparent in the study report, or perhaps find new uses for the data. For example, data on individual exposures rather than mean exposures within a subgroup of a cohort may be extracted and put to good use in a dose-response assessment. Sharing of raw data is also addressed in recommendation R-1 for improving future studies and the discussion summary under that recommendation.

For more information on this topic, please visit the Federal Focus Web site:
https://www.fedfocus.org

²⁶ Steven Rosenberg, M.D., Ph.D., "Guest Opinion: Secrecy in Medical Research," Federation of American Societies for Experimental Biology Newsletter (June 1996).

²⁷ Yolles BJ, Connors JC, Grufferman S, "Obtaining Access to Data from Government-Sponsored Medical Research" NEJM 315(26):1669-72 (Dec. 25, 1986).

²⁸ The Bellagio Principles, reprinted in Sharing Research Data, supra, at 224-25.

²⁹ The London Conference took place during three days in October 1995. The report, Principles for Evaluating Epidemiologic Data in Regulatory Risk Assessment, was published by Federal Focus in August 1996. At pp. 61-63, 66-67, Epilogue. (Emphasis added.)