Monday, January 21, 2008

Personalized Medicine

Folks who read my blog know that I am a strong believer in personal health records and personalized medicine. What is personalized medicine?

It's a lifetime transportable medical record that follows the patient everywhere to ensure they receive the safest and most effective care based on their history.

Additionally, personalized medicine is coordination of care among providers, payers, pharmacies and labs, respecting patient privacy preferences .

However, the most personalized medicine will occur when each of us has a transportable version of our fully sequenced genomes available to optimize care for the diseases we are likely to develop. The following is a description of my experience to date.

In 2007, I contributed blood, skin, and oral mucosa cells to the Personal Genome Project. Perpetual cell lines of my white blood cells were created to serve as a source of my DNA for analysis. By February 2008, project will fully sequence the 1% of my genome that is uniquely me, about 60 million base pairs.

As a first, step, my DNA was analyzed using Affymetrix technology to hybridize my genes to 12 million probes, identifying Single Nucleotide Polymorphisms and relating them to probabilities of disease through Genome Wide Association Studies . For those with a genetics background, my entire 20 megabyte SNP mapping is here and here

An analysis based on the Welcome Trust Genome Wide Association Study (GWAS) indicates that my risks for disease are:

Coronary Artery Disease - increased risk
Diabetes Type 2 - average risk
Rheumatoid Arthritis - average risk

Thus for me, becoming a vegan 7 years ago was truly a great idea. By reducing all my cardiac risk factors, I've likely negated my genetic risks.

This analysis used regions of my genome that are associated with diseases, not a full sequence. The full 60 million base pairs will be completed using the next generation technology described in this presentation.

In 2008, the US Healthcare Information Technology Standards Panel (HITSP), has been chartered with harmonizing the standards to securely transmit genome sequences. The work will be completed by October 2008.

The folks at the Personal Genome Project hope to expand their effort from the 10 pilot volunteers they have today to 100,000 volunteers. If they are successful, they may even win the X-prize.

In 2008 we'll have the technology to sequence humans in a fast affordable way, then transmit that data securely with patient consent to those who need to know it. The future of personalized medicine is around the corner, not a decade from now!


Ileana said...


This is very interesting, as usual.

May I suggest that you start adding labels and having a sidebar item with labels. It helps immensely with sorting things out and it helps Google find your blog on relevant searches. Thanks!

Unknown said...
This comment has been removed by the author.
Unknown said...

Wow, how far we've come.

About 7 years ago I participated in a DARPA-funded project with SAIC (lead), Hopkins, UPMC/PSC, UPenn and others, to work on the technical, policy and many other issues associated with a prototype of a regional medical image archive. We tried very hard to grapple with how to share medical images (and associated records) while "respecting patient privacy" but even in the very limited, controlled context of out project we encountered many intractable roadblocks - firewalls that throttled network throughput between sites; lawyers who didn't understand encryption, de-identification, etc.; lack of inter-institutional authentication mechanisms (much less clearly defined roles for fine-grained authorization for access to various parts and levels of a medical records); and, apart from DICOMM, a notable lack of standards for electronic medical information. So the project, while well-funded, did not make appreciable progress towards it's goals, although perhaps laid some groundwork for later efforts. (See 2000/hubs_04-26-2000.html for background info.)

So I'm immensely impressed when I read here about the progress that's been made in these directions with CAML, the HITSP & HISPC efforts and regional health information orgs like MA-SHARE. Frankly, I was astonished to peruse your online "Continuity of Care" document, John - what I would have given to be able to do that 7 years ago! And authentication frameworks like Shibboleth and Liberty Alliance are beginning to make it possible to share credentials across institutional boundries.

But, speaking as an traumatic aortic aneurysm survivor and a bioinformatics/genomics practitioner, who am I going to entrust with my entire "personal health record" - including years of critical CAT scans, MRIs and other digital images, not to mention my entire genome? Many of my baseline imaging studies were lost years ago, so now I insist on getting physical copies of the files or disks of the digital data (years ago when I first asked for a copy of my CT images, it was quite an interesting conversation, let me tell you - they ended up charging me $50 and handing me an optical disk).

My genome and a PDF or CAML-based PHR like you describe I could carry around my neck on a dongle or in my pocket on a smart card, but terabytes of digital images? Would I trust "Microsoft Health Vault" or Google or Wal-Mart (via Dossia) to preserve and secure them? Honestly, I don't think any of the current business models would get my vote.

Dr. Joseph Gitlin, one of the authors of the original DICOMM standard and a fellow of the Soc. for Imaging Informatics in Medicine, once proposed a community-based, non-profit "Shared Medical Archive" concept, one that, personally speaking, I'd like to see revived and a self-sustaining business model developed for.

John Halamka said...

All great points. In 2008, there will be a number of third party hosting possibilities for personal health records. I listed the categories at

The employer sponsored PHR at Dossia will be hosted by an independent third party using software called Indivo created at Boston Children's Hospital. For details, see this month's Journal of the American Medical Informatics Association "Early Experiences with Personal Health Records"

DNACousins said...


Could you explain the various column headers in your files with results? The SNP ID appears to be an Affymetrix catalog number, and the results appear to be "normalized" to an A version or a B version of the allele.

I've seen results from the consumer-oriented companies deCODEme and 23andMe that give the actual bases along with the rs (reference SNP numbers) as used in dbSNP.

Ann Turner, M.D.
co-author (with Megan Smolenyak) of "Trace Your Roots with DNA"

bestonline323 said...

As a nation, we have engaged in many debates and much discourse on how to improve America’s health care system and to me it seems to me that personalized health care is the key. I'm a big big fan of it. Good on you! Very very informative post!

Personalized Pens

Unknown said...

John, thank you for the intersting insight. Your background in medicine combined with your insight to technology is refreshing.

LEE mentions above:
Would I trust "Microsoft Health Vault" or Google or Wal-Mart (via Dossia) to preserve and secure them? Honestly, I don't think any of the current business models would get my vote.

I would love your take on this comment. I have certainly heard this comment before but in my opinion, the assumption here is that the current business models (not including Google and Walmart) do provide for a very "secure" environment. From a technical practioner's view, this could be a false sense of security.

If we continue to make this assumption, will this serve as a major obstable for Google and Walmart to move forward in this space?

Your take?

Anonymous said...



John said...

Can anyone here please provide me with this info which I am not able to figure out. What SNP databases are being used in GWAS as well as personal human genotyping. Is there a database against which the SNP data is being blasted to get the risks and diseases associated. Pls comment.