Later On

A blog written for those whose interests more or less match mine.

Tightly integrated vs. loosely linked

leave a comment »

A kind of “monument fever” afflicts many projects: a drift in the direction of building a tightly integrated and massive structure that will awe all who see it—wanting every project to be something like building Hoover Dam.

The sources of the impulse are various: overweening pride, for example, aka hubris. Or the thought that there’s a lot more money to be made in a massive project than in a simple solution. The massive single-answer project normally involves one (or a few) massive companies, who aim to make their budgets on the project, whereas loosely linked small solutions, though more effective and less costly, don’t allow for so much profit—for one thing, loosely linked projects end up eliciting multiple solutions so that competition becomes effective, and if there’s anything a big company hates, it’s competition.

When I was getting ready for the retina-repair surgery, the nurse took my medical history on a computer beside my gurney. She was seated so that The Wife (web-interaction designer) could watch over her shoulder. The Wife commented that entering each datum (e.g., list of current meds) required multiple screens and keystrokes—really, to her (professional, experienced, knowledgeable) eye, a total and utter mess of a system.

Plus, of course, I had already provided that same information repeatedly, to the other hospital for my cataract surgeries, to my various doctors, etc. Constantly re-entering and repeating the same data.

What we need, I realized, is a comprehensive integrated medical records system, so that any medical organization could tap into the system and get this info. Privacy might involve my having to give a password, maybe? I began having visions of this massive, majestic national medical-records system…

But of course, such a Hoover-Dam approach has substantial costs, and not just financial: a massive integrated single system would take years (decades?) to build, would be slow to implement new technologies (because of having the change the entire tightly-linked system). Those who have been involved in software development realize that this approach generally leads, after some years and massive expense, to a failed project that is never completed—or, if it is, is so clumsy, out-of-date, and difficult that it is not used. This approach is overkill, lead-footed, massive, difficult to implement, difficult to change or adapt, etc. However: the companies securing the contracts to build it will get a LOT of government money before they admit failure. (Cf. NSA/Thomas Drake scandal—see, for example, this summary.)

I learned that such a system is built into the new healthcare program and work is beginning. I bet you dollars to donuts that this requirement was pushed strongly by Big Business, looking for massive government contracts, and doubtless Congress was paid off along the way.

The Wife pointed out that the French took a much nimbler, less expensive, and (I would bet) more effective approach: a medical record “credit card” that carries in its magnetic stripe one’s full medical history. I show up at the hospital, give them my medical record card, they swipe it, and Bob’s your uncle.

This is NOT a tightly integrated system: very loosely linked, in fact. The key definition is the format and contents of the record and how it’s stored on the card. Once that is set, a thousand vendors can compete to produce the best (most cost-effective) solutions to all the other parts of the system: competing programs that can read, write, update the cards. The card is the fixed point, and these various little apps can readily be created and compete until the best solutions (based on user experience) emerge. Much cheaper, faster, more flexible, more likely to evolve quickly in the direction of increasing effectiveness—and much easier to modify because it’s not one giant, tightly linked system.

Really, if you think about the two approaches, it’s no contest: the French medical-history-on-a-card system is totally superior to the massive-national-tightly-integrated system (which, IMO, will either never be completed or will never work—too big and cumbersome to be modified quickly and easily, though such modifications are essential to the continued development and improvement of the overall system). The other system, with a swarm of smaller apps, can move much more quickly, with new, improved apps continually entering the fray and displacing the old, not-so-good earlier apps. And the card serenely sails though the on-going evolution of tools.

As empires age and grow, they become addicted to monument fever in their public works. The medical record information system now being developed by the Federal government is intended to be that sort of monument. It won’t work, but a good solution has in fact been found and we could even study and improve on it. But that is not the American Way. We are in the grip of monument fever (and pushed that way by Big Business, which wants the money to be spent on building the monument).

UPDATE: The medical-history-on-a-card approach could probably be pulled out with very little expenditure of public funds: First, the government funds the development of the format and the card. Then the government requires that medical records system must be able to read, update, and write such cards. The various systems developers, simply as part of competing, will take care of developing the programs to do that, and the resulting competition of systems and apps should push improvement. Card format and structure can be reviewed every 5 years, say, to accommodate updates necessary because of new knowledge from medical research.

Written by LeisureGuy

11 March 2012 at 8:11 am

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.