Abstract:
|
The democratization of source data and analytical tools has lowered the bar in terms of the difficulty involved in performing sophisticated analyses with minimal training or experience. An unintended consequence of this development is that we need to be increasingly vigilant about the nature of information that is made publicly available. The biomedical industry generates vast troves of sensitive patient level information, and needs to take appropriate steps to safeguard this information. While privacy of healthcare information is a pressing concern for many patients, there is a simultaneous and conflicting demand for industry in particular to be transparent about its research findings and in many cases make source level data available to the public. In this talk I will touch on a number of related topics of interest to industry, ranging from operationalizing public disclosure of de-identified clinical study data, to issues in differentiating between what is possible and what is permissible when combining data sets containing protected health information, to the merits of selecting analysis methods where the summary statistics are invariant to the process of de-identification.
|