This is good news here and will certainly help quite a bit. If you notice the wording here it will “decrease” the risk of linking two data sets together. I have done my share of queries and know how the processes works to attain data, so this is a good thing all the way around. 
The real trick here is to still have enough information available to evaluate and research without compromising identity. For the most part too, it is not the researchers as a whole that we need to be concerned over, it’s more or less those who use these data bases for profit as that is one big carrot hanging out there with tons of data to be analyzed.
I think we need a 12 step program for those who are addicted to “analysis”processes too as just like alcohol at times if the data is sitting there in front of them, they can’t help themselves and start writing SQL queries and searching for elements that can help cash in on some big bucks. Gee, that last comment reminds me a bit of Wall Street (grin).
I made this comment in jest but there really could be some substantiation to some of this as just like diagnosis in the clinical world leads to treatment, data in the IT world leads to analysis, more algorithms created.
Two words in the last couple of years have become very prominent in the world of journalism and blogging and those are “might” and “may”. We report on some of these analysis stories and are sure to use those words as we don’t know either on some of what gets presented. There’s the OMG stories like one that might say if you stand on your head more than once a week you “may” have a 25% greater chance of developing brain cancer (silly example) but indicative of what we see out there. There’s is is also good substantiated data information presented by all means, but we get a real mix in the news today.
Back on track here the folks at Vanderbilt have done some other neat things too, especially with a process of algorithms that help alert medical staff immediately to the onset of Sepsis. If you compare this latest project with working to anonimize data to what has been done in the past there, they may be on to something pretty substantial.
Vanderbilt University Makes Massive DNA Data Base Available for Researchers To Study Use With Medical Records
All DNA samples and medical records have been de-identified so use in research can take place. The research will be able to determine if genetic information in the medical records could help improve patient care via using personalized medicine, based on genomic information added to a medical chart. With a de-identified record, researchers will be able to study and evaluate a medical record and see if personalized genomic information added would have perhaps created a better or different outcome with treatment.
Vanderbilt University Sepsis Detection Technology Update – Added Management to Detection Process
After deployed the very first alert initiated an attending physician to begin antibiotic therapy for a patient. They average 3-4 alerts daily in a 25 bed ICU unit. Work has just started on adapting the language and knowledge to chronic heart failure. I originally posted this story back in March of 2008 with the start of the project.
It’s all about those algorithms in analyzing, reporting and making money for some in some areas of healthcare today. BD
Electronic medical records are being hailed as a tool to aggregate patient data and advance research, but questions remain about how the vast sharing and compiling of this critical medical/genetic information will remain de-identified to protect patients’ privacy and security.
Researchers at Vanderbilt University have found a unique algorithm to make electronic medical record information anonymous for genome-wide association studies (GWAS), according to a paper that recently appeared in the Proceedings of the National Academy of Sciences.
Their approach is called Utility Guided Anonymization of Clinical Profiles (UGACLIP), and it involves generalizing some of the diagnostic information from electronic medical records to make it more abstract and anonymous. The method was validated using data from nearly 3,000 patient electronic medical records from the Vanderbilt University Medical Center.
The team tested this UGACLIP algorithm on two real patient data sets from Vanderbilt University Medical Center’s electronic medical records system and found that the algorithm did decrease the risk of linking an individual to their GWAS data while also maintaining much of the clinical and diagnostic information needed to permit data sharing and follow up studies. Building upon this success, Vanderbilt researchers are continuing to improve the algorithm and plan to develop software so that other investigators can safely and securely extract electronic medical record data in future GWAS.
0 comments :
Post a Comment