Meaningful Health Data Mining – How will we regulate consumer-driven research and advice?

At the National Committee on Vital and Health Statistics executive subcommittee hearing on “meaningful use” of health information technology, Carolyn Clancy, director of the Agency for Healthcare Research and Quality testified “We haven’t reached a system-based approach where the right thing to do is the easy thing to do.” The meaningful use of health information technology will free patients to organize to accelerate research and deliver advice independent of any particular doctor orhealth plan. Data mining opportunities traditionally restricted to doctors and health plans as a side-effect of their essential services will now be available to anyone that gains the trust of a patient-consumer including, for example, not-for-profits and Internet social networking groups.

Suggesting or confirming “the right thing to do” involves coordinating disparate information that includes mining patient data for decision support (to search and display guidelines), for comparative effectiveness research (to find and group similar cases), for bio-surveillance (to find cases that match a profile) and for informed consent (to quantify the risks of alternative treatments). The result of data mining is useful to the doctor, the patient and the investigator.

As with other things, the American Recovery and Reinvestment Act (ARRA) leaves much of the rulemaking and guidance regarding data mining to interpretation by the Secretary of Health and Human Services. To add to the uncertainty, at the recent Health 2.0 conference I learned from Ann Waldo, Esq. that health records not covered by ARRA are nonetheless covered by consumer protection laws. The law addresses the problem of inappropriate solicitation or misleading advice as a matter of privacy, consent, disclosure, role, identity (anonymity), transparency and accountability. I’m not enough of an amateur lawyer to dive into the details.

There is great value in access to a comprehensive Patient Controlled HealthRecord (PCHR) for second opinion, treatment, pay-for-performance, research and pharmaceutical marketing. Expect doctors, advocacy groups and other trusted entities to compete for the privilege of hosting and/or accessing the PCHR with patient consent.

So where do the cats and dogs come in? Doctors and health plans (the cats) provide necessary medical services to a more-or-less captive consumer and their activities are regulated by HIPAA, ARRA and institutional review boards (IRB) accordingly. Doctors and health plans are paid for services other than health records management (and it should not be surprising that they have traditionally done such a poor job of health records management).

The business models for patient controlled health record hosts, public registries, social networks and advocacy groups (the dogs) are all about health records management. As Deven McGraw JD, MPH of the Center for Democracy & Technology argues on the e-Care Management blog, these institutions and groups should not be subject to the same regulations as the covered entities. This will help provide maximum value to the consumer.

The Innovator’s Prescription traces the inefficiency of today’s medical practices to mixed business models.  Tethered, incomplete and ineffective (in the sense that they are inadequate for catching medical errors and seeking second-opinions) “personal health records” are evidence of this problem.

As ARRA eliminates the treatment exemption loophole in HIPAA while forcing the covered entities to put complete health records online, patient controlled health records hosts will compete for the right to mine and extract value on behalf of the consumer under ARRA and consumer protection laws. As with music, gambling and banking, regulations and services that do not take into account the consumer’s definition of meaningful use of the Internet will likely be bypassed or ignored by an empowered and impatient consumer. Competition to win the consumer’s trust, gain their attention, and mine their data is about to heat up.

Adrian Gropper, MD is a founder of MedCommons, with roots in patient-controlled and patient-centered health records that go back to MIT’s Guardian Angel project. AMICAS, a more recent radiology-focused venture, pioneered the clinical use of Web browsers and protocols. Adrian is driven by the vision of doctors and patients collaborating around shared health records on the Web.

Categories: Uncategorized

Tagged as:

9 replies »

  1. I would have to agree with you. The issue of privacy of information is becoming more important to consumers. THey want to be able to trust whoever they are working with that their personal information will not be taken.

  2. Thanks for this wonderful post. It triggered a lot of new thoughts for us in the areas of health informatics. We have something similar kind of a service called XTractor- which is very popular among biomedical researchers- we mine PubMed everyday for updates and provide them to users for FREE.. Please do check http://www.xtractor.in for more information..

  3. Adrian,
    I think I understand where this is going. I do agree that once we get over the hump of having clean, comprehensive, electronically maintained medical records, the data mining issue will need to be addressed. Frankly, today there’s not much to mine anyway.
    I can see organizations, other then the primary care providers, that will want to mine that data, and in some cases this mining will provide useful service to to the patient or the collective. In other cases it will not. You are correct in saying that this process needs to be regulated. I would think that regulation number one is patient consent for both data collection and its intended use.
    Data mining almost always has negative connotations when privacy is considered, even when the purpose of mining is ultimately beneficial to the collective, and I agree that we need to pay attention to this now because although liquid data is not yet available, its advent is imminent.
    As to the mechanism, I still think that the primary care facility, and every patient should have one, should be responsible for the aggregation on an individual level and, of course, reimbursed accordingly.
    On a population level, there may very well be organizations that engage in aggregation and analysis of data for all their members. The point-to-point paradigm should still work and the individual records should be obtainable from PCPs.
    You raised an interesting point regarding specialists in care coordination. I haven not thought about that possibility and the more I think about it the more it sounds like that specialist will really be the primary care provider. I guess the definition for primary care is a bit fluid in my mind. It should be whoever coordinates your care and aggregates the big picture of the entire patient, not necessarily the Family doctor, even though he would be the most likely candidate. I would still insist though that it should be a physician that performs that function.

  4. Margalit,
    I agree with you about paying for the value of curated, aggregated health records. Paying the primary care physician and medical home is quite reasonable but not the only way to achieve this. I can imagine specialists that focus on care coordination or a medically educated family member serving this role for some patients. I can also imagine tight and semi-private web support and advocacy groups helping their members as peers.
    The point of my post is to call attention to the opportunities around regulation of data mining which leverages complete and aggregated health records individually or as a population. Good regulation will raise the bar for quality and effectiveness of both licensed (hospital) and unlicensed (social group) advisers.
    With some payment reform or innovation, a small group practice can operate a medical home, but an independent practice with thousands of patients is hardly large enough to run significant data mining operations. Will such a practice offer their patients a choice of independent data mining services or will they honor the patient’s own choice even as they provide aggregation and curation services for the personally controlled health record of their individual patients?

  5. Adrian,
    Transparency is always a good thing and I can definitely see how costs are contained if physicians have access to up-to-date information from other care facilities, by avoiding duplication of tests and providing more appropriate care.
    I can also see how informing a patient of treatment choices available and having the patient participate in his/her treatment decisions may reduce costs to some degree.
    I just don’t see any opportunity for cost reduction and/or increased quality of care by dumping the responsibility of proper medical records aggregation and general management on the patient. I also don’t see how this alleviates privacy concerns, since the data will most likely be aggregated in some mega-storage, commercial vendor, which sooner or later will be attempting to monetize the investment in building these “free” PHRs.
    To paraphrase Dr. Halamka (but not his conclusion), the point-to-point data exchange system has much more appeal to me. There is no need for elaborate interfaces and most of the technology for such data transfer is already there. Yes, we may need a master patient index (even though surescripts is operating a rather large network without one) and we may need a sophisticated addressing scheme, but the internet protocols coupled with web services standards (and I don’t mean necessarily SOAP) should provide a very good start.
    Here is what I envision: patient walks into new care provider facility, provider’s web enabled technology sends request to patient’s primary care facility, primary care facility responds with complete records. This of course requires that the Primary Care facility, whether a physician or a Medical Home, is responsible for continuously aggregating the medical records from all sources. This should be part of the, so called, coordination of care.
    The way I see it, the chances of getting accurate, reliable records is much increased by this fully automated process, as compared to the patient being responsible for the aggregation. Some patients will do this well, most will come up with nothing or, even worse, incorrect or incomplete records.
    The question, of course is who is going to pay the Primary Care physician for this additional task. Since I think that record aggregation is integral to coordination of care and patient centric medicine, I would suggest that we change the reimbursement method for Primary Care physicians and actually pay for what we say we want, instead of what we say we don’t want – procedures and more procedures.

  6. Margalit,
    Paraphrasing the Congressional Budget Office, unwarranted care (procedures that the patient might not need or want if they were fully informed of the alternatives and risks) is a $800 B cost containment opportunity in the US.
    Introducing transparency into our system may not be a direct contribution to cost containment but it is likely to be the most economically significant contribution we can make. Many ARRA provisions are designed to increase transparency through health information exchange and this can be achieved through either clinician-to-clinician or patient-centered solutions. Comparing Patient Centered Health Records to clinician-to-clinician exchanges, John Halamka says “When the patient controls the data flows, the confidentiality issues are much easier to manage.” [ http://tinyurl.com/dlajf9 ]
    PCHR facilitate independent data mining and make second opinions more accessible. Informed patients, their families, and advocates will reduce unwarranted procedures such as aggressive treatment at the end of life that doctors are ill-positioned to manage themselves.
    See also https://thehealthcareblog.com/the_health_care_blog/2009/01/a-transparent-health-record.html

  7. Really(Exactly), to increased the number of cases, and we must see that type of incident will improve the way of helping the people, for it in findrxonline..com will find the solutions and other ways of offsetting it.

  8. I must be missing something here….
    I thought the ultimate goal of ARRA was to pull the economy out of this recession by creating jobs and, as far as HIT is concerned, by containing costs.
    So how are patient controlled medical records contributing to this goal so much that they are to be defined as “meaningful use” of HIT technology?
    Wouldn’t meaningful use be something that directly contributes to cost containment?

  9. One point I may have missed but what about using the health information technology system as replacement or supplement to the current FDA Adverse Event Reporting System for both drugs and medical devices?