Reinvestment in Government-Funded Research

Blood Institute (NHLBI) sent a memo to his Division Directors instructing them to inform contractors and grantees about the Institute's data sharing policy. He wrote, " It is the policy of the NHLBI to make available detailed data from collaborative clinical trials, epidemiological studies, and other large-scale studies … with adequate protection of the confidentiality and privacy of research subjects. " 1 Dr Lenfant wanted assurance that the federal government's investments in research would achieve maximal potential by enabling researchers who were not direct beneficiaries of those investments to analyze shared data and in turn to share their findings, through publication, with the community. Indeed, one of us (MSL), as an early-career investigator, took advantage of NHLBI's policy of sharing to publish independent analyses based on data generated by an NHLBI-funded investigation. 2 And the importance of sharing has been recently reaffirmed as a critically important measure of the value of biomedical research. Since 1989, we have come a long way. At the end of the punched card era, the NHLBI Framingham Heart Study converted massive tabulations of collected variables into 38 monographs that were made freely available for outside researchers, many lacking computing resources. 4 As technology advanced, punched cards moved to reel-to-reel tapes, to big floppy discs, to smaller ones, to CDs, to secure FTP, enabling data transfer with increasing content and efficiency. Data sharing processes eventually moved from ad hoc requests to well defined and well-documented web-based systems including BioLINCC 5 and dbGaP. 6 Data sharing has also evolved from collaboration of a few studies 7 to massive collaborations that have included more than 50 studies. 8 Data sharing offered great opportunities for advancing research, but as Dr Lenfant intimated, also presented challenges that went beyond exploiting advances in computing technology. As responsible overseers of federally funded research, we needed to be certain that participants were aware that their data could be shared and that they provided clear consent for sharing. The consent form for Framingham Heart Study participants in 1950 – it was only one page long – looks very different from the one of today. 9 Participant consent now requires permission as to who should or should not receive their data, whether restricted by content or for-profit companies , and whether there are sensitive data that should not be shared with anyone beyond the first-line researchers. In the current issue of Circulation, the American Heart …