Create a National Healthcare Innovation Database?

Dr. Joseph Jasinski, Program Director, IBM Healthcare and Life Sciences Research

Mickey Kaus makes an interesting proposal that we should think about from the perspective of patient-centric longitudinal electronic health records:

“Open Source Health Studies? Here's an idea that undoubtedly is not original, and may or may not be good.** You tell me:…"

"...an open database of longitudinal medical histories, searchable by anyone with a modem and a hunch. Get a sufficiently large cohort of people to record everything about their health--what diseases they get, what procedures they have done, what allergies they have, medicines they take, lifestyles, basic socio-economic info, etc  ... Once that's done--and I'm sure such private research databases exist--it shouldn't be hard to construct a Web page that lets a visitor to the site compare, say, the percent of women who've had more than 8 mammograms who get lung cancer with the percentage of women in the whole sample who get lung cancer.”

I have been hearing a lot of similar types of proposals and seeing a lot of interest from people outside the medical profession who are realizing that their medical data is important to them and -- in anonymous form-- important in helping to understand health risks and find cures.

Does the information technology exist to do this – undoubtedly yes.  In fact, IBM is already pioneering some work in this field, our Interoperable Health Information Infrastructure is a prototype of how such sensitive data could be securely stored and sifted for such insights. And our Websphere middleware has become the platform for  the Healthcare Collaboration Network product we have developed.   

Moreover, we're working with Mayo Clinic, Cleveland Clinic and many other healthcare leaders on ways to transform research, drug discovery and innovations in patient treatment. The idea that information technology not only can help find long-term cures but can be an instrumental part of daily patient care is gaining steam.

The kind of database Mickey describes doesn't exist, yet, as far as I know.  But it's clear that the technical and social foundations for it are being laid. Consider how Wikipedia has emerged as a massive, self-organizing social collaboration. Clearly medical research is bound to follow simiilar models of open innovation. As a proof of concept, we can look at an existing project that involves genomics. 

IBM is partnering with the National Geographic Society to create the Genographic Project.  The Genographic project uses genomic data not for medical research, but to allow people to understand how their ancient ancestors migrated out of Africa. Each participant provides a sample of DNA from a cheek swab and some basic personal information and if they choose consent to put their anonymous data into the project database.  In return, participants get to find out in detail how their ancestral line migrated across the world tens’s of thousands of years ago from a website. 

So far, after only four months of operation, over sixty thousand people all over the world have participated.  In essence, thousands of people have volunteered their genetic information -- the biological core of who they are -- because they see personal value in doing so.  What's more, the more people who participate, the more accurate and detailed the information they receive. Sounds a bit off the topic, but think about it: this genetic databasing is a technology implementation of the kind of service Mickey is suggesting, and it demonstrates that people will participate - even pay to participate - if they see value in the project. 

I don't want to take the analogy too far, because the Genographic Project has nothing to do with medical research or healthcare, and the database contents are not open to the public, but hopefully you get the idea.

So let’s go back to the medical record idea…What if a foundation or a company or a medical research institution or a consortium put together the basic information technology infrastructure to allow volunteers to input their medical information, anonymously, and gave researchers access to the data over the web? 

The value of that database to medical research would be immense, but what would the value to individuals be to participate?  Perhaps if your data was available anonymously for research use but available to you as an archival repository of your personal, private health information and that of your family, that would create the individual incentive to participate in such a program.  How much might such a repository, if it existed, simplify the lives of evacuees from the U.S. Gulf Coast following Hurricane Katrina?

I don’t have an estimate of the economic or societal value an effort like this might create in terms of medical research or in helping individuals keep track of their medical care, but I do suspect it is significant.  I also suspect that it will only happen if companies, institutions and the public, with a shared interest, organize and participate - perhaps starting with a focus on a chronic or life threatening disease?

Comments

[Jack Mason]

Dr. Jasinski:
This idea is fascinating. How do you envision that data would remain anoymous and secure?

Jack - good question. Security and privacy will always be concerns. One advantage to the proposal we are discussing is that it is voluntary - like deciding whether or not to have a bank account. As a result, in addition to the normal technology protocols, you are in control of the information, you decide what goes in and what does not and the database can issue you a unique anonymous identifier - something that is not possible for the U.S. government to do, for example. JMJ

[Lei]

I suspect Kaiser Permanente has extensive data on their members; many of whom have probably been in their system since childhood. I'm sure some of my colleagues are mining their data right this minute.

[Janine Burbage]

Joe J - I like this idea. I think people would participate if it was secure and AND if they received summaries of the research findings where their data is being used. I'd participate. Who would own the data? Could you update the information in the data? Let's say you were cured of cancer (notice I didn't say you were diagnosed with cancer). Could your physicians use this? Do you think that this would be a way to advance the use of electronic medical records?

Janine

Janine - as far as I know, ownership of the data resides with the physician or institution that obtained it - BUT - as a consequence of HIPAA in the US at least you and I have the rights to a copy of our data and to use it as we choose. So yes, one could envision a range of services beign built up around the data including use for your own medical care. JMJ

[Erich Ruetsche]

Joe
I like the concept of an open patient owned database. Eventually the health data are owned by the patient at least in Europe. I think it would be worthwhile to think about a 'open source' model where the patient could add his data to the database and open it for selected access depending on his preference. Payers, providers and research organizations should contribute (finance) such a model because it might be the fastest and cheapest way to introduce make health records available and reduce cost in the industry. A reduction in premiums would be finally the best incentive for the patient to share his data.
Erich

[James Kaufman]

Joe,
This is a very interesting thread. Together with Janine's comments below it becomes really clear that any future infrastructure for support longitudinal ehr must implement the emerging standards from HL7 and the IHE. That is the only way to meet the needs of indidividuals and also support the requirements of public health. It the clinical affinity domains formed by RHIOs are all disconected and custom (one off) solutions, we will never achieve the ubiquity needed for a national health network.

[George Eisenberger]

Joe I think it's a great idea. There are some key technologies that will have to be part of the infrastructure. First, there will have to be a MED (Medical Entities Dictionary) that would ensure a common vocabulary for medical concepts. So that when a diagnosis or lab result is entered the same terms are used (e.g. ICD, LOINC, SNOMED, CPT) SNOMED-CT is the vocabulary of choice for the U.S. Govt and is making it available for use (in some cases at no charge). If we don't use a common vocabulary in healthcare analysis will be very difficult, nearly impossible. The other technology is a Universal Deidentificaion Plaform to ensure the privacy of the data and ensure that all the PHI (Personal Health Identifiers) are blinded from the Public.

The question is who pays for all this infrastructure. It has to be designed, built and managed/operated.