Data Transparency in Pharma
Clinical trials are inarguably the most expensive part of drug development, and the most risky to study subjects. Thus, it is only sensible to promote maximal usage of already available data, through data sharing. However, some stakeholders are hesitant to share data, stating concerns for patient confidentiality and vulnerability of intellectual property in pharmaceutical and biotechnology industries. Still, those who have embraced data sharing recognize that it is a public responsibility, as it promotes research progress.
The pharmaceutical industry does not always readily share this type of information. However, there is growing pressure that sharing the data will help the public. Organizations such as the International Serious Adverse Events Consortium (iSAEC) have taken initiative in this arena. Through collaboration with the FDA, leading pharmaceutical companies, and academic institutions, the iSAEC aims to identify DNA-variants that will give insight to the risk of drug-related serious adverse events. In addition, Johnson & Johnson and Medtronic have partnered with Yale University Open Data Access (YODA) Project in an advocacy for responsible conduct in data sharing and transparency. Since many clinical researches are federally funded and therefore tax financed, one could argue that the public should fully benefit from the research. Data sharing allows outside investigators to test their hypotheses, and publish the results. Moreover, new methods can be developed and tested, and verifiable results and reproducible analysis will strengthen the credibility of already published analyses. Of course, careful consideration on how to obtain consent is needed to enable data use beyond its original purpose.
To statisticians, data sharing will, for example, impact meta-analysis. This is especially true for survival meta-analysis where traditionally, hazard ratios from publications (if available), would be used to determine an effect of covariates of interest, on survival. The availability of data from previous studies will allow for further exploration of the individual patient data (IPD) approach to survival meta-analysis, where raw data from studies is utilized. Also, a study of multiple datasets is a chance to study patterns of missingness. At BSSI, we have experience with IPD analysis and appreciate the convenience of data availability, and the opportunity it provides for research and development.
Though some companies in pharma and biotech industries have embraced data sharing, many have expressed concerns. Analysis through other investigators could highlight flaws in the original studies and analyses. However, it can be argued that data sharing will give researchers an incentive for thorough research, knowing that the data and the results may be questioned later. Still, others express that there is no guarantee of acceptable quality of resulting research findings. But, it can be debated that before data and results are published, they receive scrutiny from peer reviewers. In drug development, entities such as Food and Drug Administration (FDA) are stringent in their examination of research results.
A legitimate concern is the protection of patient information and adherence to Health Insurance Portability and Accountability Act (HIPAA). Furthermore, just because data are available does not mean it is usable. Proper data infrastructure is needed. To be most useful, data must be in the correct format and labeled well enough for sharing.
Though there are arguments from both sides of the table, it is clear that through careful guidelines for data sharing, this is a chance for evidence based clinical trials, better prognostic models, and will result in enhanced biomedical innovation.