Healthcare.com Would Have Worked Better

Is this any way to build a railroad?

By now you've heard that the "Obamacare exchanges" did not launch on October 1 so much as stumble out into public view, barely able to crawl.

Three weeks later, the federal version -- "healthcare.gov," which is actually the same exchange re-deployed 36 times in 36 states -- is still barely able to crawl. By contrast, most of the 15 exchanges operated by individual states and the District of Columbia are working more or less fine, for varying reasons we will explore in a moment.

Why the epic fail for healthcare.gov, estimated to have generated a health insurance enrollment rate of less than one-half of one percent among nearly 10 million visitors? Information technologists have identified lunk-headed flaws in its overall design, while pointing to the way the Federal government rolled it out, all at once, all across the nation -- as if it were a campaign commercial and not one of the most complex undertakings in the history of e-commerce.

Which would be for good reason: the federal exchange is a campaign commercial, one the Administration had no choice but to broadcast after its opponents went to war on every front against implementation of the Affordable Care Act.

The architects of the ACA expected that states would build their own exchanges. The federal exchange was supposed to be a failsafe -- a fallback for a few straggler states unable to build their own in time for the October 1 launch. For the rest, healthcare.gov was supposed to do two things: point people to their state's exchange; and handle the very complicated task of querying tax and other federal databases to verify people's eligibility. Instead, it found itself saddled with the entire e-commerce job for 36 refusenik states.

And then the really bad decisions began. Notwithstanding the existence of several private health insurance exchanges that have been working for years -- and numerous turn-key insurance exchange software systems -- healthcare.gov was hastily thrown together by CGI, a behemoth government systems integration contractor few had heard of until this week. CGI's work was overseen by a federal bureaucracy under political siege from all directions, including most of the states it was attempting to help -- while holding its breath as the very basis for the project moved its way through the bulk of the federal legal system. Obvious design problems were never addressed and worked out publicly, the administration's energies consumed instead by existential defenses of the law and a scramble for implementation funding.

Design review? Implementation simulation? User testing? Yeah -- we're getting to that next Monday.

This is exactly how not to build a railroad. Ask anyone in the working information technology industry -- or in biotech, nanotech, or alternative energy. Taking on grand projects with immense technical complexity and high risks of operational failure requires a culture obsessed with the task, rewarded for the risks, and shielded to every degree possible against uncertainty. Is that what it felt like for the people overseeing the creation of healthcare.gov the past three years? Of course not. Which is why it is important to remember that they never wanted much of that job in the first place.

The Devil is in the Politics

Back in 2010, when the health reform plan was moving from the drawing board to the floors of Congress, its architects had good reason to believe the states would build their own exchanges -- and do so competently -- because two very different states already had.

The actual Affordable Care Act -- as opposed to the hydra-headed political animal known as "Obamacare" -- deferred a great deal of the law's day-to-day workings to local markets and state-based insurance regulation. What is the obvious way to implement the ambitious national goal of universal coverage, while deferring to the unique conditions of hundreds of local marketplaces? With exchanges operating as close to those marketplaces as possible.

Long before the ACA was drafted, several private exchanges were up and running to do just that. There are private exchanges for large employers operated by benefits management companies like AonHewitt, Mercer, and TowersWatson; others for smaller employers and consumers operated by Silicon Valley companies like GetInsured.com and eHealth (full disclosure - I am an advisor to GetInsured); and for years, individuals have been able to sign up for Medicare plans via ExtendHealth, another Silicon Valley company acquired last year by TowersWatson.

The architects of the ACA anticipated that most individual states would choose from among these many competing options for their own exchange -- in some combination of buy, build, or lease - because they knew they were already up and working. They also took at face value the ferocious lobbying positions on various elements of the ACA, by champions of state rights, for more local control.

Deferring to individual states for the exchanges would also be consistent with the "Massachusetts model" at the heart of the ACA itself. But Massachusetts was neither the only inspiration for a state-based exchange, nor even the first. Utah was. In fact, Utah's pioneering health exchange introduced the very idea to national health policy conversation -- back during the George W. Bush Administration.

As a market-enabling mechanism for ending the dysfunctionality of local health insurance purchasing, exchanges were first championed by President Bush's Secretary of Health and Human Services, Michael Leavitt, Utah's former Governor. Not coincidentally, Utah is one of three states with a Republican governor operating its own exchange -- an adapted version of the one it built years ago.

The real-world, red-state inspiration for the health insurance exchanges included in the ACA may seem bizarre in 2013 -- but it is no more bizarre than the conservative origins of the individual insurance mandate -- another odd feature of the law that inspires as much fury from the right as it does denial. But bad partisan politics can do ugly things to good policy ideas.

Those first health exchanges were operational long before there was a Tea Party movement -- before the politics of health reform soured, the ACA turned into "Obamacare," and educated people with real political aspirations started equating the law with slavery. So much for state exchanges and all the other rational, market-oriented ideas in the ACA. Suddenly, it was Civil War.

That war erupted in 2012, when the Supreme Court upheld almost all of Obamacare -- except the provision that had, in effect, required states to expand their Medicaid programs. Most states responded in predictably red or blue fashion. At the same time, they also made their go/no-go decision about an exchange.

As a result, only 14 states (and DC) committed to setting up their own exchange, most of them solidly blue, though with a few odd exceptions: Kentucky went forward with its own; New Jersey did not.

Because we do not have reliable numbers yet, we cannot correlate which technology approaches by those states -- buy, build or lease -- are working best. But we do know from reporting around the country that they are working better than the federal exchange -- because most of them appear to be working and people are managing to sign up for health insurance.

Which brings us back to the well-publicized traffic problems and miserable enrollment rate of healthcare.gov. It turns out that users of the federal exchange have to run through the entire enrollment process before they can see any of their insurance buying options. It is a simple design flaw, but a potentially fatal one. Imagine a first-time visitor with Amazon.com having to create an entire secure e-commerce account before ever seeing a product, or with Match.com before ever seeing a profile of anyone, or with Expedia before seeing if it had any actual flights. If there is one product design principle known to anyone who has ever worked for an e-commerce company, people browse before they buy -- and many if not most of them are just browsing. The electronic traffic jams created by this terrible little design flaw multiply logarithmically.

This would not have happened if healthcare.gov had been undertaken by an organization with the wherewithal to treat the job like healthcare.com.

Some of the most mission-critical components of government programs are developed, built, tested and launched not by government employees or general systems contractors, but by highly specialized commercial enterprises. The federal government was not up for the task of building and overseeing a complicated e-commerce business with a significant data management back-end - any more than it has ever been up for building the military's fighter jets, running commercial airports, or administering the actual Medicare program. (Sorry, fans of "Medicare-for-All," but the program has always been administered by commercial insurers.)

As with national defense and Medicare, these critical programs should be mandated, funded and regulated by the government. But their execution is best left to private enterprises who compete for contracts, talent and capital, strive to foster innovation and efficiency, and maintain cultures of real accountability. If healthcare.gov had been designed, built and launched by that kind of culture -- rather than a behemoth government contractor supervised by a besieged bureaucracy -- we can be certain it would be running a whole better right now.

Yes, those who built healthcare.gov did so against a backdrop of enormous political conflict, and the failure of many states to step up and help is tantamount to sabotage. But that sabotage was aided and abetted by an administration that confused political and legal survival of Obamacare with implementation success. The administration apparently felt little need to push and pull on the product as it was being built; nor heed the very pointed advice of Bill Clinton back in February to "make it work;" nor rally the thousands of people working 24/7 on the project.

That is exactly not how things get done in Silicon Valley.

Healthcare.com would have worked better.