It started like this...

Your doctor knew you. He (and it was a "he") knew your family, your community, your history. That knowledge was stored in his head in narrative form where he understood your life, and thus your health history, in a longitudinal context annotated and detailed by the environment, events, and behaviors that defined you as an individual.

In other words, he understood your health history as a story, full of qualitative information running longitudinally through time.

As our nation grew so did the domain over which an individual physician roamed. Pressures, including professional and economic, drove doctors and their staffs to see more patients and to spend less time with each.

But an individual patient's history was still largely relegated to a narrative manner in the form of handwritten notes regarding observations and those handwritten notes becoming part of the patient's record to be reviewed. Like a kind of "John Smith for Dummies" whenever the doctor saw you. He had crib notes from which he could reconstruct your narrative, just prior to any visit.

In the second half of the twentieth century things began to change. The 1950s and 1960s saw the rise of the technocrat as a political force and with it the idea that the application of technology to virtually any problem, particularly social, would make things better. Why? Well, of course because, science. Because machines. Because labor factor productivity.

There wasn't any point in bucking the trend.

In medicine this meant two things. The introduction of technology for diagnostic purposes – imaging and biometric mostly – and the introduction of technology for practice management. And by practice management, I mean "billing."
The problem is: The Devil is in the details. We cannot move from transactional, acute, fee-based medicine to outcome and results based medicine without a change in our approach to problem solving and we cannot change our problem solving until we change the technology we use to help us solve our problems.

Now machines are very good at producing and processing quantitative data – adding numbers, doing monstrously repetitive tasks, etc. But they have really no facility for processing qualitative information.

We've heard much hype about the application of IBM's "Watson" technology to health care and diagnosis but when I went online and asked Watson the simple question "Does my wife have amyloidosis?"

It fell completely apart. HAL 9000 it ain't nor is it going to be, any time soon.

Machines produce data as quanta – numbers. And machines have a requirement that data, not information, be presented to them in quanta – numbers. So the entire practice of medicine moved from processing information as narratives, stories, to processing data in tiny, discreet, units.

Of which the two most familiar to us are: a) Billing Codes and b) Diagnosis codes.

This quantum approach to health care works OK for fee-based, face-to-face, acute care. We're really good at being presented with a severed limb, diagnosing it ("right leg is detached from torso"), fixing it ("stich leg to torso while controlling bleeding"), and most importantly, billing for it. There are individual codes for each and they come out of and go into machines with great facility.

The quantum approach absolutely falls apart when it is applied to long-term, chronic care. It even falls apart when applied to acute conditions that fall outside of the known and familiar – the ailments that are visible but inscrutable. Rare conditions or conditions presented in parts of the body that they do not usually present.

These are conditions where a qualitative, narrative-based, approach is necessary. But there is very little now in the American health care system structured to deal with qualitative information. Systemically, because of the requirements of the technology – data processing machinery not information processing machinery – we collect, we store, and we operate on health information atomically, not coherently.

As data, not information.

Case in point: the transition from ICD-9 to ICD-10. There are thirteen thousand individual codes in ICD-9. In ICD-10 there are nearly seventy thousand.

That's going the wrong direction. There is no justification for more codes. There is a tremendous cognitive benefit in the industry to go to fewer codes.

There are less than a thousand unique words in the Constitution of the United States. That document describes the entire legal and corporate form of the oldest and largest democracy on Earth. There are ninety-eight natural elements in the periodic table and together they describe every material thing in the Universe. But it takes seventy thousand codes to describe the taxonomy of the medical constellation?

The drive towards more codes is an attempt to encode qualitative information – narrative – within the constraints of a quantitative methodology. There is a feeling that if codes only become more specific – each code covering fewer variations of a case – that a side effect will be the encoding of qualitative, useful, information.

That having a code for "pulmonary hypertension and left-handed-ness" and a code for "pulmonary hypertension and right-handed-ness" will somehow improve care and diagnostic outcomes than if we just had a code for pulmonary hypertension. Or just hypertension in general.

And we do this, not because it's true, but because our machinery can easily absorb more codes that essentially all describe the same thing and since the machinery can, it should. And we feel better because should one machine exchange medical data with another that exchange will be more detailed and then that detail will magically turn into information.

The problem is: The Devil is in the details. We cannot move from transactional, acute, fee-based medicine to outcome and results based medicine without a change in our approach to problem solving and we cannot change our problem solving until we change the technology we use to help us solve our problems.

The technology we have today worked, to a degree, in a world much different than what we are moving to. That technology is also a virus which has infected the industry and constrained the way it does, can, and will do its job. For the challenges of this century the old cannot become the new – the transactional, cost-accounting, foundations on which it was built simply don't adapt to the qualitative longitudinal foundations needed now.

We can keep pretending otherwise, we can keep cramming more and more billing, diagnosis, and drug codes into the systems. We can come up with new but still linear and hierarchical models of relating those codes – like XML, CCDs, etc. But that just changes the arrangement of the deck chairs on the Titanic mess that is the status quo.

And it needlessly adds more deck chairs only because more chairs is better than coming to grips with the fact that the ship is sinking.