A reader asked me to write tonight about the Health Information Technology for Economic and Clinical Health Act, which is about as far from something I would like to write about as I can imagine, but this is a full service blog so what the heck. The idea behind the law is laudable — standardized and accessible electronic health records to allow any doctor to know what they need to know in order to treat you. There’s even money to pay for it — $30 billion from the 2009 economic stimulus that you’d think would have been spent back in 2009, right? Silly us. Now here’s the problem: we’re going to go through that $30 billion and end up with nothing useful. There has to be a better way. And I’m going to tell you what it is.

But first a word from my reader:

“My number one annoyance is that hospitals are still extremely dumb,” he writes. “My wife has Lupus (and other disorders) and every single freaking time we go in they always have to know the same information and fill out the same damn paper forms. I expect more, a lot more from my doctors and the systems that support them. Why are we still dealing with paper and paying people to just duplicate that effort by typing it in? Wouldn’t it be much easier to say, ‘Here’s what we have on file for you (hands the patient a tablet computer). Please review it, update any medications with dosing as well as any allergies.’ Boom, you make the changes directly in their system, you’re not repeating that effort every time and this system would be completely standards based with every other system in production. Problem is, we have no standard and it should have been laid out before any of these EHR systems were built. I’m not talking about creating another FBI virtual case file waste of taxpayer cash, but these professionals know exactly what kind of information they gather and how it’s used. There is no excuse for today’s hospitals.”

So we passed a law, appropriated a lot of money, and are now implementing a medical records system that the National Academy of Sciences says in a recent 273-page report isn’t going to do the job.

This is not the first such negative report, either.

Yet still the project moves forward because, well, there’s that $30 billion we should have spent back in 2009 and if we don’t spend it, heck, Congress might just take it back.

Use it or lose it is not the proper motivation for this or any IT project.

Remember that more than 50 percent of major IT projects fail outright producing nothing, so the chances going into this were that it would fail. From what I have read that is already happening since the whole concept of interoperability — the basic purpose of the system — seems to have been lost.

It’s not like there aren’t any good medical record systems. Back in 2007 I visited the Mayo Clinic in Rochester, Minnesota specifically to learn about their medical records system that had already been in development for more than a century. Mayo built the first standardized medical records system anywhere on Earth using lots and lots of paper and had over decades refined it into an amazingly useful bit of analog technology. When I visited, the clinic was rapidly turning six million paper patient records into electrons in a way they were sure would quickly lead to saving lives simply because they’d finally be able to correlate treatments and outcomes over thousands of similar patients over many decades. Computers were everywhere and caregivers could instantly find any data they needed including X-ray images. Mayo had a good system and they were making it even better. Problem solved, right?

Maybe if we could first kill all the administrators.

Every hospital administrator has his or her own idea about how medical records should be kept. This has resulted in many different systems being developed with something less than 50 percent of those actually working successfully. But no matter how many systems failed they didn’t all fail and some are probably pretty darned good. I was certainly impressed with what I saw at Mayo.

So here’s what we do. First we stop. Whatever we have so far is crap, believe me, so let’s throw it away. It was insanity to start building a whole new records system when there were already systems in operation that were close to the objective.

We do this really stupid thing over and over in IT, which is we find something pretty good and say, “Let‘s build something like that.” Which means, if you think about it, “let’s take this functional model and attempt to emulate it with what has historically been less than a 50 percent success rate.”

I’m with stupid.

So we stop; we throw away the work we’ve already done; then we take a look at all the electronic medical records systems that are already up and running at scale. I’m sure there are at least a dozen of them running in various places around America. Our goal here is to pick the best of these systems. We don’t even have to be good at selecting because if we were to make a mistake and choose the second best or the third best system it would still be better than the disaster we’d likely build from scratch.

We figure out which is the best presently operating system in terms of functionality, reliability, ease of use, and any other criteria you’d like to add. Then we do something that is never done in IT. We take this off-the-shelf product, spend a little of that $30 billion to buy it outright, then give it to every hospital, clinic, and doctors’ office in America.

And that’s it. No R&D, no development — just pick the best and give it to everyone for free. And if there are problems here and there, well it is easier to fix a problem than it is to build a system.

Now for a really radical idea: just point Google at the problem and let them buy the system and run it.

They’ll make it ad-supported and therefore totally free to doctors and patients alike.

This isn’t brain surgery you know.