- Defining Clinical Trial Transparency
- Clinical Trial Spin and Suppression
- Benefits and Costs of Transparency
- Legislative Strategies to Encourage Transparency
- Bottom-Up Strategies for Clinical Trial Transparency
- Conclusions: Towards Greater Transparency
- Clinical Trial Transparency References
Defining Clinical Trial Transparency
There are varying degrees of transparency of clinical trials and the associated data. We can distinguish three major types of trial transparency. At the most basic level is trial registration. Registration of clinical trials ensures that regulators and the public are aware of the existence of a trial. Trial registration is a necessary for any of the other forms of transparency, but only in the last decade have there been serious efforts to encourage trial registration and open public access to registries. Trial registries may contain only the most basic information describing the trial, the condition and intervention under investigation, and the investigators’ and trial sponsors’ identities. A trial registry can also contain information about the trial protocol which should include a description of the trials’ intent, the mechanisms by which patients are selected, the pre-specified outcomes under investigation, and methods of analysis to be employed. By requiring protocol registration before a trial starts, researchers can be prevented from changing the questions that they ask based on the data.
After a trial has been performed, the second level of transparency is reporting of summarized results. An important consideration in the reporting of results is whether they are made available as data-only, as data with narrative summaries. Although results databases help to make information more easily accessible, a risk of allowing narrative summaries is that they might make it possible for a submitter to introduce a biased interpretation of results without being subject to peer review. Therefore, there have been strong arguments from the publishing community to restrict results databases to rigidly specified forms and tabular data, offered without interpretation, which is left “to the reader”. The problem of how to balance the need for accessibility and comprehensible result summaries with the risk of allowing data to be subjected to spin and misinterpretation is an important open question that will likely require novel technical and policy solutions to fully solve.
Finally, the highest degree of clinical trial transparency is to make all primary data available for examination (with appropriate measures in place to ensure confidentiality for research subjects). The advantage of making primary data available is that an outside scientist can assess whether analysis has been done correctly and potentially identify whether there are discrepancies or problems with the way that data was collected. Full access to raw data poses many potential problems as well; external scientists (or worse, lay people) might misinterpret the data. Exposing primary research data could also threaten patient privacy, unless any identifying information was carefully scrubbed. Depending on the type of data recorded, it might even be difficult or impossible to fully anonymize the clinical data — genetic information such as SNP measurements (single-nucleotide polymorphisms) might be sufficient to uniquely identify a patient, even if biographical details are not available.
A related issue that I’m not going to address in this series is the question of disclosure of financial relationships between trial sponsors and the physicians who conduct studies . Under the current system, the vast majority of clinical research is sponsored by companies that hope to have their investigational products eventually approved. These companies compensate physicians directly for the costs of conducting the trials, and may also pay these clinical researchers for consulting or other services. Not all research sponsored by companies is bad; indeed, there is evidence suggesting that industrially sponsored studies are better run and producer higher-quality data than studies funded solely from public sources, possibly because the company studies have better funding and larger incentives to be performed well in order to garner regulatory approvals. However, there have been concerns raised that payments to doctors could be directly or indirectly affecting their scientific independence and clinical judgment. In response to these concerns, the “Physician Payments Sunshine Act of 2008”  was a bill proposed by Rep. P. DeFazio in the 110th Congress. The Sunshine Act would have amended the Social Security Act to “provide for transparency in the relationship between physicians and manufacturers of drugs, devices, or medical supplies for which payment is made under Medicare, Medicaid, or SCHIP, and for other purposes.” The Sunshine Act would have required sponsors to make quarterly reports to the Department of Health and Human Services of all payments to physicians or professional medical organizations. The Sunshine act never became law, but illustrates how one element of the healthcare system, in this case the US Government as the funder of Medicaid and Medicare, can exert its fiscal power to encourage transparency.