A Software Developer's Take on Healthcare.gov

As a software developer I have some experience with projects like healthcare.gov. I worked for companies that provided computerized testing.  They had worldwide registration systems and transmitted exam databases and software to a network of autonomous testing centers across the world, where high-stakes exams were delivered under the supervision of proctors and the results transmitted back to our central hub.

The tragedy of software development is that the launch of the health care website is all too typical. That's not an excuse: it's a simple statement of fact.

When Target came out with its site in 2011 it had a ton of problems. Best Buy also had serious issues with its site. When United Airlines and Continental Airlines merged their reservation systems they had terrible problems (United was still having problems as recently as September, when it was charging customers $0 for tickets). Pretty much every online game has launch-day blues, ranging from minor issues to major meltdowns that can take weeks or months to fix. Who can forget the problems that Microsoft has had with every major revision of Windows, including Vista (which had a redo with Windows 7) and Windows 8 (which just came out with Windows 8.1).

And of course, there are the problems that George Bush's signature program, Medicare Part D, had when it came out. Instead of jeering on the sidelines, Democrats helped the Bush administration resolve those issues.

The reasons for these problems are usually the same: late starts, uncertain scope and scale, shifting requirements and inadequate testing. The healthcare.gov site suffered from all of these problems, many of them foisted on the project by its detractors.

Late Starts, Cast-in-Concrete Deadlines
Most software projects don't get started on time, but the deadline rarely changes. This was especially true of the ACA, which had been delayed by legal action for years. The people who claim the developers had four years to write the software are flat-out wrong. Lawyers were still duking it out in the Supreme Court in June of last year, which meant that the software developers didn't even know whether the project would really happen until 16 months before launch date. At that time the justices threw out a major portion of the law -- the Medicaid expansion -- which  affected the design and implementation.

Shifting Requirements
Then there are imprecise requirements. The health care exchange was clouded by uncertainty over the exact scope of the website: the law gives the states the right to create their own exchanges, and many did so. If all the states had created their own sites, that would have dictated a much smaller design for healthcare.gov, relegating it to a simple entry portal to the state exchanges and back-end data processor and validator.

But the developers of healthcare.gov had to wait for all the states to decide what they were going to do -- and many of them didn't make a decision until December of 2012, after the Republican Governor's Association asked the Obama administration multiple times to delay the deadline. That gave healthcare.gov only 10 months after finally learning the true scope of the project. At that point they learned that they had to set up exchanges for 38 states. And only 10 months to do it. Imagine how difficult it would be to design a skyscraper without knowing how many floors it was going to have.

It's rather hypocritical that the Republicans who forced delays in the development of healthcare.gov -- through lawsuits and dillydallying on deciding whether they would create their own exchanges -- are now complaining about its rocky start.

Full-Bore Launch vs. Incremental Development
Websites that people compare healthcare.gov to -- Amazon.com, for example -- have been developed incrementally, and have been up and running for more than a decade. When Amazon began they had few customers and little visibility. They had low volume on launch day and were able to fix their problems in obscurity. They had no deadline other than their own internally imposed one. They could push the launch back without any repercussions. They could grow slowly and incrementally.

The healthcare.gov website was a huge deal that everyone was watching. Its start date was dictated by law. Millions of people and thousands of companies are depending on it. The site got hammered by millions of hits on day one, and it collapsed -- just like thousands of smaller-scale private software projects before it.

Inherently Messy Project
The job this site is trying to accomplish is much messier than ordering a coffee maker. It's not trivial to find out whether people qualify for tax credits, and you can't find out the real cost of a policy without that information. You have to make sure that the person really is who they say they are, and you have to be concerned with privacy and security. Making something simple to use and making it confidential and secure are competing -- almost contradictory -- requirements.


I'm not making excuses for the problems. I'm just saying that we've seen them before, in thousands of private and government software launches. But we've got time -- two more months before any insurance policies issued under the ACA actually go live, and three more months after that for people to sign up.

People should give Jeff Zients his month to get the problems fixed. If we're still having these problems on Dec. 1, then we can go into full histrionics mode. Until then, healthcare.gov isn't all that different from many other software launches that limp along in their early days. Except that the project is beset by millions of people hoping desperately for it to fail and actively trying to sabotage it.

The real reason that this health care rollout is so messy is that it's basically the Heritage Institute and Mitt Romney's plan to keep the insurance industry profitable. Obama didn't originally want an individual mandate; he preferred a public option that would have looked a lot like Medicare, which would have eliminated many of the problems with healthcare.gov.

Thus, the left criticizes Obama for selling out to insurance execs, and the right criticizes him for a socialist takeover of health care.

I guess that's the definition of compromise: nobody is happy.