Health Insurance Really Tried

05/12/2010 05:12 am ET | Updated May 25, 2011

I started a company in 1979. We sold very advanced information technology to large corporations and government agencies. In the early days, it was tough trying to convince potential customers that our complex, expensive, shiny new gadget would make them a ton of money.

I used to share a cartoon that depicted a medieval stone castle, with the yeoman of the guard in front of the moat talking to a mustachioed salesman. The yeoman was dressed in chain mail and carried a mace and a broadsword. The salesman, dressed in a three-piece suit, sat cross-legged in a machine gun nest, with the "whole nine yards" of ammo. The yeoman was saying to the salesman, "I don't have time to talk to any new-fangled gadget salesmen. I have a war to win."

Twenty or so years ago, the private health insurance industry was the gadget salesman, the health care industry was the castle. Private insurers believed that through a combination of automation, promoting best practices, and pooling risk, they would be able to offer quality health care affordable to just about everyone. While the health care debate rages towards some seemingly imminent conclusion, it may be well to contemplate why the serious effort of the private health care industry to implement so a good theory turned out so poorly.

Computers are incredibly fast, but they are also incredibly stupid - in the sense that they cannot apply judgment. If every third word of this post were mis-spelled (which the good people at the Huffington Post would never allow!), you would still be able to get the sense of it. But one character out of place flummoxes a computer. One of the biggest problems doctors have with Medicare is not the reimbursement rates, but that Medicare systems reject so many of their claims for tiny data entry errors that wouldn't bother a human, even if one noticed. Further, the rejections are mysterious, taking hours of effort to understand and unravel.

Medicare is a homogeneous system. Each private insurer, by contrast, offers a wide array of plans for different types of customers, each with its own terms. Automation is cost effective when the same thing is done the same way - exactly the same way - millions of times. It is not cost effective if identical rules apply only to a few transactions, and new rules need to be programmed all the time.

The result is that private health insurance has very high overhead: 20% to 40% of premiums collected. This contrasts to about 1% for Medicare. Some of the difference is scale, but most, I think, is the high cost of automating very complex processes with huge arrays of options requiring lots of judgment. Fleets of computer still have to be supplemented by armies of people, and the result is costly and error prone.

Private insurers also looked for automation to help develop databases of best practices: ways of treating medical conditions with the best outcome for the least expense. The problem is that insurance pays for activity, not for outcome. So, even with a massive amount of data about what was done, insurers collect amazingly little data about outcomes. All of the other parts of the health care industry wind up being paid by insurance for activity. The more things they do, the more they are paid. That simple.

As much as private insurance wants to promote best practices, the incentive of every provider is to do more, even if it is redundant, even if it is not effective. Fear of lawsuits does not cause this problem, although it undoubtedly accelerates it. Nobody can really measure how much is the "pay for service" dog and how much is the "defensive medicine" tail.

In a perverse way, automation undermined individual responsibility for health care. When my young adult children were little children, I would take them to the doctor, pay the bill on the spot, and accumulate enough bills until it was more than the insurance deductible. Then I'd send in a batch of paper, and get a check from the insurance company in due course. I'd hang onto the paper, because I could claim anything not reimbursed by insurance as a deduction at tax time. As a result, I knew exactly what services my family utilized and how much each cost, whether it was on my dime, the insurer, or Uncle Sam.

Today, the bill is settled directly between the provider and the insurer. The patient deals only with the residue in a complex, opaque, after-the-fact form. The tax code took away the incentive to keep and review the paper by requiring non-reimbursed medical expenses to exceed a percentage of income. Like most Americans, I now have no idea what my health care costs.

This is very important, because it means that market forces do not work in health care the way they do in most industries. Medical consumers never had much help keeping up on the science, but it's totally impossible to make informed decisions if you can't know either the products or the prices. Whatever it is, it is not a market. So nothing we know about markets helps us understand how it works, or how it fails. And in fairness, Medicare has exactly the same issue.

Private health insurance hoped to lower costs by creating large pools of customers, spreading the risks widely and thus lowering the average cost. The problem is that the very diversity of private health insurance works against this goal. Each plan, each group, each company, in each state forms a separate risk pool. This is particularly bad for small companies, and small companies account for almost all of the job creation in the economy over time.

So it is no surprise that private insurers raise their premiums. They need to cover their costs and make a profit, so they have no alternative. The problem is that this is creating what engineers call a "positive feedback loop:" as premiums go up, more people decide that they cannot afford health insurance. Since these tend to be the younger workers who are healthier, the insurers wind up with a customer base that is older, sicker, and thus more expensive. So they have to raise premiums further. And lose more of the remaining youngest and healthiest customers. And raise premiums even more ... until no one is left.

The medical term for this condition is "circling the drain." The private health insurance industry is in a death spiral, despite being profitable, despite trying to do the right things.

The debate on the health care bill has become all about political power, not policy. It will settle out one way or another. But the policy questions will remain. The private health insurance industry worked the problem for decades, but in the end did not have enough market power to drive a better solution. The challenge is to do better, understanding why private insurance failed to deliver on its once shining promise. We're betting our lives, as individuals and as a society, that we can.