Several news outlets have reported on a new finding published in the New England Journal of Medicine (vol. 357, no. 4, July 26, 2007) regarding the spread of obesity in social groups. The article, “The Spread of Obesity in a Large Social Network over 32 Years,” was written by Nicholas Christakis, MD, PhD, and James Fowler, PhD.
Their findings detailed the increased chances of an individual becoming obese if 1. a close friend became obese (57% increase); 2. an adult sibling became obese (40% increase); or, 3. a spouse became obese (37% increase). They also found that persons of the same gender also increased the chances of one obese person influencing another.
The data set used for the study comes from material collected for the Framingham Heart Study that began in 1948. A second cohort of the offspring of the first FHS was established in 1971 and a third cohort of offspring of the second generation began in 2002. The obesity study used only data from 5124 individuals in the second FHS cohort (1971). It tracked the social networks of the people by creating a database of information taken from handwritten administrative tracking sheets used to facilitate follow-ups with each participant. The tracking sheets included family names, relationships, addresses, and at least one close friend as a contact. This social information was not the basis for the Framingham study, but merely an administrative tool to be able to contact someone for their next appointment. The net result was 38,611 observable social networks among the participants.
This study demonstrates two interesting points in regards to archives. First, data collected for one reason can be creatively repurposed at a later time for another reason. This is one of the simplest arguments for retaining records in archives. Records are not kept merely to recreate the transactional nature of institutions and their activities, but instead provide for the use of information for other intentions.
The second point is the fragility of data. The information used to gather the findings was part of a three generation (and counting) study of heart disease. It was available because it was part of an active project that was well-documented. Due to the size of the FHS and its findings, it is likely that the data sets produced will be well-maintained after its conclusion at the National Library of Medicine or elsewhere. However, other data sets with the same potential to be repurposed into new studies are not always as lucky. And, it is likely it will become more precarious for them in the near future.
Issues concerning the privacy of patient and human subject information are part of the archival management process. The Privacy Rule regulations under HIPAA have done little to address the long-term preservation of PHI materials in archives. It is much easier for principle investigators, institutions, and archives to simply destroy the data as a sweeping act of safeguarding. True, there are few alternatives under the current regulations, but it comes at a cost of re-useable information and perhaps at the cost of our own health and well-being.