Ezra Klein writes,

To this day, I’ve never read a compelling explanation of why the nation’s doctors and hospitals haven’t broadly adopted electronic medical records.

Has he not read my explanation, or does he not think it compelling? In case it is the former, let me state it.

I spent much of my career in business, and much of my focus was on use of information technology (IT). Among the lessons I learned were.

1. Within a company, every business area gets the IT it deserves. Chaotic, haphazard business areas get lousy IT (and blame the IT department). Organized, well-run business areas get great IT.

2. Data cannot be maintained unless there is clarity of ownership. It must be clear who is responsible for creating, maintaining, updating, and deleting the data.

Our health care system is highly fragmented. It is chaotic and haphazard, particularly when compared to a system of managed care or a government-run health care system (the latter being managed care carried to an extreme).

There are pros and cons of managed care. One of the cons is that Americans hate it–we would much rather be able to self-refer to a specialist rather than have to go through a managed-care process. Managed care can be difficult to reconcile with our preference for physician autonomy.

But one of the pros is that managed care is likely to use IT better. I am told that both the Veterans’ Administration and Kaiser Permanente are way ahead of the rest of the U.S. in their use of electronic medical records.

Without managed care, there is no clear owner of individual health records. With my checking account, my bank owns the data. They are responsible for creating, updating, deleting, and so on.

It does not work to have my personal computer be the definitive source of my checking account data–I am not reliable enough a manager of this data. Similarly, the individual is not a reliable enough data manager to own his or her own electronic medical records. But there is no one else in the process in a position to own those records, either. My doctor does not always know when I self-refer to a specialist. Even if I do not self-refer, the specialist does not have direct access to the primary doctor’s data records.

We could fix the existing system to create a definitive data owner. We could require each person to designate a primary physician, who in turn is the owner of that person’s medical records. However, I think it would be very difficult for the primary physician to manage these records in a secure manner when other providers get involved. If I am the primary care provider and you are a specialist, how does my system know which part of the medical records you are allowed to read and which part is confidential with respect to you? How does my system know which part of the records you can edit and update, and which part are “read-only” from your perspective?

These are tough issues. Again, I have a hard time seeing how they get resolved outside of a managed-care context.

In conclusion, don’t think that electronic medical records represent a mere technical challenge. They are part of a business process. As long as Americans prefer a fragmented health care process to a managed-care process, electronic medical records are unlikely to prove to be a panacea.

I would like the market to decide between managed care and fragmentation. I think that in a free market, more people would choose managed care, in part because of the cost advantages it enjoys, including greater ability to use electronic medical records. But the same employer-provided health benefits system that insulates consumers from the cost of health insurance also insulates them from the savings of managed care. If you as a consumer are not going to enjoy much of the savings from managed care, then you might as well indulge your preference for greater autonomy for yourself and your doctor.

As I’ve said before, the original sin in American health care is employer-provided health insurance.