Who Killed HealthCare.gov?

Scott Zuke

(12/2013) The Patient Protection and Affordable Care Act (ACA), also known as Obamacare, has been law for over three years, but has never felt settled. After more than 40 Republican attempts to repeal the law in Congress, a costly federal government shutdown during which a delay in implementation was demanded, and a Supreme Court ruling that upheld the law, but on obscure grounds, it has always felt like it's still up for debate, rather than being the established law of the land.

Part of the reason is also in the design of the law, which front-loaded popular items like letting children stay on their parent's plans longer and barring insurers from refusing to cover children with preexisting conditions, and delayed the more complicated portions of the law until now and into next year. This would normally be clever policy design, except that the ongoing fight over the law has prevented the usual public buy-in that develops over time. Even though support for Obamacare has not changed dramatically since it was passed, the fact that it was never allowed to settle into general acceptance during its gentle phasing in signals a rough path ahead.

Though not really relevant to the debate over the substance of the ACA, the botched rollout of HealthCare.gov has played into the hands of Obamacare critics. With the highly-visible launch of the critical, and embarrassingly dysfunctional website, they've been handed the perfect confirmation of their argument that the federal bureaucracy is incapable of administering the program.

President Obama has said that passing and implementing the healthcare law is the most important thing in his presidency, and the online exchanges have long been promoted as key to making health insurance more accessible and accommodating to peoples' needs. When people started logging on to the website on October 1, however, it buckled under the strain, revealing insufficient server capacity and a system riddled with incomplete features and programming errors that prevented all but a small fraction of users from registering accounts.

How and why HealthCare.gov's rollout went so disastrously is a case study still being written. Blame has been doled out in varying degrees to the president and his administration, the Department of Health and Human Services (DHHS), and the IT contractors who received millions of dollars to build the site.

The contractor drawing the most negative attention is CGI Federal, which won a $93.7 million contract in 2011 to build the health insurance exchanges. The Canadian-owned company has had a mixed record on its government contracts, finishing some projects on time and on budget and having others get so far off track that they were cancelled.

Despite being the lead developer of the ACA's key component, observers criticized CGI for understaffing the project, leading to missed deadlines and shoddy coding. Although CGI had raised concerns that there was insufficient time to test the system before launch date, their warnings were clearly understated, and their own expectations of what they could finish by the deadline were overly optimistic. By the launch, the system wasn't just untested; it was incomplete and defective.

Some blame may rest with DHHS, which took years to release the final specifications for the program and may have cut into the time CGI expected to have for testing. However, the responsibility still would have fallen to CGI or the federal project managers to revise the deadlines or reasonable expectations accordingly.

Another explanation for the system's failure is that there was poor coordination between the multitude of contractors who developed different portions of the website. There were "(at least) two sets of contracted developers, apparently in isolation from each other, working on two pieces of a system that had to run together perfectly," writes David Auerbach, a software engineer and writer for Slate.

Actually there were 55 contractors involved in the creation of the exchanges. Auerbach's analysis only looked at the disconnect between two of the most visible ones, and the consequences of that communication failure. The Centers for Medicare and Medicaid Services (CMS) is the federal agency charged with overseeing the website's creation, including coordination of the various contractors, so Auerbach's indictment would probably fall on them.

What's clear is that no one entity is solely responsible for the website's failed launch. But is it fair to say that Government itself, or at least the standard government procurement process for such projects, could be the underlying problem? It's tempting to think so since examples of costly, unattractive, dysfunctional government websites come quickly to mind.

The way accountability works in government tech projects, bare functionality is often the best that can be hoped for, since things like nicer appearance and better usability are expensive to produce and don't result in added revenue the way that they would for a commercial website. The reluctance to overspend tax dollars unfortunately also ends up frustrating citizens trying to use the sites their taxes paid for. HealthCare.gov was supposed to be a step in a new direction. Designed with the end-users in mind and with heavy emphasis on simplicity and usability, it turned out they created a beautiful website that didn't work.

This wasn't so much a matter of misplaced attention on form over function, but rather a result of needing a website of such size and complexity that weaving everything together was always going to be difficult. And in this, we have to come back to the law itself: a massive piece of legislation with many stakeholders and tough deadlines. In an ideal setting, more time would have been allotted in anticipation of the kinds of difficulties often faced by projects of this scale. This is far from an ideal setting, though. This is a political war zone.

From a technical standpoint, the problems with the ACA's insurance exchanges are fixable. If executed well and quickly, the political fallout may even be fairly limited. Nevertheless, we can expect the label of "the latest HealthCare.gov" to be applied to all bungled website rollouts for some time to come.