Words Can Be a Window into Fuzzy Thinking about EHRs

Recently David Blumenthal, writing in the New England Journal of Medicine, presented a summary of the health IT (HIT) schemes of the federal government. Since he was the national coordinator of HIT during the time that the new healthcare reform schemes were conceived and instituted, it is reasonable to assume that his statements reflect the rationale and motivations of those who gave them their current form and, from the absence of critical comments in his summary, it is reasonable to conclude that he agrees with the rationale and goals of the programs.

Some would say, these programs are a "done deal," so why dwell on them? There are at least two answers: 1) Unlike communism in the USSR, participation in these programs is (at least partially) optional; and 2) Just like communism, their mere existence does not make them good or guarantee their viability in the millennia to come. They will certainly end at some point, either by obsolescence, replacement, or collapse. In any case, after they end there will still be sick people and doctors and life will go on. The degree of difficulty that will be encountered in “going on” will be determined in part by whether patients’ medical records "go down with the ship" and whether enough insight was gained to do a better, longer-lasting job the next time around.

As I suggested last time, there is an almost universal wish that computers be applied to every problem that has been identified with both medicine and healthcare delivery. This is much like the hopes that were held out in 1900 for electricity and radiation. The wish is so strong that the things that are wished for have been commanded to materialize, even though there is little reason to expect that such is likely, or even possible. The wish is so strong that I believe it has led those formulating the schemes to choose, either subconsciously or intentionally, words to describe them that appeal to the "Fast System" of Daniel Kahneman, words that are likely to trigger favorable associations and biases, thereby reducing the chance that people might think deeply and analytically and ask hard questions. This is a trap into which both the speaker and the listener may fall easily; the words are all very familiar and so few have either hands-on experience or theoretical knowledge that could counteract the immediate associations made by the fast system.

I want to begin with Blumenthal's lead item: "In the health field, one ... program involved a commitment to digitizing the U.S. health information system. The HITECH Act set aside up to $29 billion over 10 years to support the adoption and 'meaningful use' of electronic health records (EHRs) (i.e., use intended to improve health and health care) and other types of health information technology." First he invokes the word digitize, with which few are really conversant, uses it to evoke an image of activities that are beyond its actual definition, and then conflates the conjured image with health records and technology-related changes in the process of delivering healthcare.

He then refers to "the U.S. health information system." Does he mean the paper records that most facilities still use or the computer information systems that are currently in use at a minority of facilities? The term "information system" is so closely associated with computers nowadays that an author who proposes to use it in some other sense should make that explicit. Absent such a qualification, most will assume that the reference is to computer systems - but wait ... modern computers ARE digital, that's why they are called digital computers and any information stored within them is digital.

"The term digitization is often used when diverse forms of information, such as text, sound, image or voice, are converted into [digital] code," according to the “Digitizing” entry on Wikipedia. If Blumenthal meant to refer to existing paper records, the only way to "digitize" them is by scanning, which most organizations already do. If document imaging is the primary goal (first one he mentioned) it is already attained or was well under way, if not, what exactly WAS intended? One could develop computer application into which information from paper records could be entered (abstracted) but that computerized data would not be a "medical record" either functionally or medico-legally. Any current or future care that might be documented directly via a computer application would be digital de facto but the process would not be digitization (which, by the way, carries an added implication of automation,) it would simply mean that the work done by the practitioner would be easier or harder but certainly different and it would produce records that are either better or worse than before, but different.

One further point about digitization of old records, if they are needed at all they are going to require direct inspection by the practitioner. No one acting prudently could or would make clinical decisions on the basis of abstracted data, especially if that abstraction was done by machine or by individuals not medically trained - except for a limited circumstance that I will take up next week as well as the question: Why not continue to digitize records in the future?

Find out more about Daniel Essin and our other Practice Notes bloggers.