Consider the source in health care data integration

The Atlantic, not typically a technical rag, recently presented an article by business and economics editor Megan McArdle on health care data integration entitled “Paging Dr. Luddite”. The article brings to a mass audience an understanding of both the importance and difficulty of data integration, but the title and general anti-healthcare-professional tone seem counterproductive.

The article opens with a compelling example of data silos, detailing the challenges of integrating the many data feeds from sensors attached to a premature baby “as small as the hand that cradles” it. Each sensor produces a data stream, and, in a field trial at Toronto’s Hospital for Sick Children, IBM is integrating that data in near real time to identify early signs of infection. Early treatment of infection in preemies can help prevent “lifelong complications ranging from vision problems to permanent brain damage.” The article moves on from this point solution to the real objective: health care data integration will enable a shift from treatments to outcomes as data mining enables detection of “patterns in disease and treatment that we now discover only through painful trial and error.”

Today many factors are driving us toward more integrated and accessible health care data, including business initiatives like the IBM offering the article describes, and provisions in the recent stimulus and health care reform bills that mandate expanded data collection and analysis.

But these offerings and incentives are encountering resistance.  In my opinion this is where Ms. McArdle goes off the rails.  Sure, what she writes and quotes is true enough, but it seemed dismissive of the concerns of those involved, like the “fierce resistance from … patients concerned about privacy; doctors concerned about autonomy; people at every level of the system concerned about expense.”  She implies that the problem is that the “human beings” in the medical profession are “quite happy the way they are”.  So the reader comes to the conclusion that we’d be able to integrate health care data if it weren’t for all those Luddite doctors!

Success in data collection and integration is impossible without understanding the motivations and interests of those at the point of data entry.  It seems to me that physicians have two primary motivations: to help people and to make a living. Today they support financial goals by accurately documenting standard procedure and diagnosis codes for claims submission, which might be why claims data is of reasonably good quality. On the patient care side, “physicians will often jot down notes to be entered into the computer later” exactly because the systems they have access to are “poor substitutes for a pen and paper.” In a world where creating easy-to-use social networking apps for teenagers is top priority, we need to be able to provide the right device for doctors – there should be an app for that!  And that app should be able to provide the doctor with fingertip access to facts like whether or not a patient actually filled a prescription.  Of course there are other factors beyond data integration and ease of use for the physician.  Ms. McArdle notes patient privacy and potential increase in malpractice suits.  But these are reasonable concerns.  One who raises them is not necessarily a Luddite.

We in the data management / business intelligence community have learned the hard way about incentives for quality data collection: the act of collecting data should serve the interests of those who collect it. Since it ignores this basic guideline, does the disparaging tone of this article toward healthcare professionals itself indicate the primitive state of healthcare data integration?

Leave a Reply

Your email address will not be published. Required fields are marked *