The launch of HealthCare.gov — the federal health insurance exchange website mandated by the Patient Protection and Affordable Care Act — went live Oct. 1. So many problems plagued the site that it prompted Congress to hold hearings about it in late October, with House GOP members demanding answers from involved parties that ranged from site contractors to Department of Health and Human Services Secretary Kathleen Sebelius.
Experts outlined many of these issues in the September Benefits Selling cover story “System Failure,” which predicted the project’s scheduling challenges, necessary system interfaces and user experiences could lead to problems. But the scope of them went far beyond the site’s inability to process complex individual applications, as evidenced by the software crashes that rendered HealthCare.gov inaccessible to users.
Not only that, but enrollment numbers recently released by HHS fell short of expectations. Nearly 27,000 people signed up for private health insurance thoruhg HealthCare.gov one month into enrollment.
So what went wrong? What, if anything, went right? And what’s being done to resolve these issues?
The most obvious issue was the website’s inability to manage a heavy use load: On the first day of the launch, it crashed due to the sheer number of potential plan enrollees hoping to register and shop online. Although the situation has improved, it didn’t instill faith in consumers or legislators that the site was ready to process the seven million applications the administration expects to log on before the end of March.
Last-minute software changes emerged as another pain point. For example, when the site first launched, users were required to register for an account before they could start shopping for plans — a design issue consumers found frustrating and experts said was antiquated.
“Anybody who has any type of website or ecommerce experience or understands anything about user experience knows that’s a big mistake,” says Andrew Carricarte, chief executive officer of IOS Health Systems, in Miami. “And I think the decision was made by somebody from the outside with no knowledge of how the technical aspects work — particularly that late in the cycle and that late in testing.”
The site’s decision-makers have since changed course, allowing consumers to browse what’s available before registering.
“That being said,” Carricarte adds, “that should not have led to the performance issues they experienced, even if you had everybody registering for an account. When they use these big numbers — 700,000 people visiting the site to purchase health insurance — to most people, it sounds like a big number, but it’s really not a high number in respects to most websites. When you look at companies like Amazon, Google or Facebook, which are in the billions. By comparison, it’s an extremely small number.”
The site’s ability to interface with other websites, including government sites such as the Internal Revenue Service or the Department of Veterans Affairs, probably made things worse, experts say.
“When you sit down and decide what the requirements are that this system has to do, somebody has to start mapping out the interfaces,” says John Sarich, vice president of strategy at VUE Software, in Omaha, Neb.
“Those systems weren’t designed and structured so hundreds and thousands of people can access them. You’re going to go to other systems from other government agencies to get and validate more information. So consequently, you have to interface to those systems. And those systems are old. Tapping into those systems was never intended to be this extensive, so consequently, they created bottlenecks, and they should have seen that very early on.”
Health care information technology attorney Tatiana Melnik says corrupt data is another issue that’s been reported already — and likely to get worse.
“The insurance carriers are receiving bad data,” she says. “There isn’t any specific information about how many applications have gone through to insurance providers, but from the reports I’ve seen, the ones that have gone through have had some problems in the classification of dependents and things of that nature. This is a real costly issue — and this is why the IT issues are so problematic, because many companies are relying on the data being correct.”
And with more than 50 different vendors working on different areas of the site, it was almost inevitable glitches would appear.
“You’ve got various disparate or siloed platforms — a federal data hub, Medicaid systems,” notes Michael Weiskirch, human relations and benefits technology consultant at HNI, in New Berlin, Wis. “It’s a hodgepodge of platforms as well as consultants and vendors working on this thing that’s supposed to come together, and it got really rushed. There’s an issue of lack of alignment not only from a project management standpoint, but also from a systems standpoint.”
Although information security hasn’t been an issue yet, experts say, it’s likely to become one as more consumers flock to HealthCare.gov to seek out insurance.
“I suspect that in the next three or four months, there’s going to be some type of security breach or issue with that website,” Carricarte says. “If they were that lax with what they needed to accomplish with basic testing and performance, one can only assume that there is some laxness and oversight with the security protocols, especially given all of the new changes they are trying to implement in a short period of time. It’s a high-profile target and this should have been their first priority.”
The biggest thing the government could have done to prevent these problems, experts say, is allot more time for testing.
“There are very extensive test criteria,” Sarich explains. “You come up with hundreds and hundreds of scenarios of somebody who’s accessing the system — maybe they forget something and their login times out, or they forgot their password. All the different cases you can get, you’ve got to test for them from the ground up.”
Aaron Mulder, an IT expert for Chariot Solutions in Philadelphia, says load tests would have helped site creators pinpoint problem areas before the site launch, which could have prevented the system crashes.
“We have to take it for granted that there are going to be hundreds of thousands of people trying to use this site,” he notes. “When we put the load on the system and see which part is breaking down, then we can put a fix in. And a combination of testing is what’s called for — you want to make the site more user-friendly and fix the underlying problems that make everything available and bring it up to speed.”
“There simply wasn’t enough time to adequately test these systems,” says Dan Schuyler, a director at Leavitt Partners who helps guide the firm’s exchange practice. “And as a result, we’re going to have problems like we’ve been seeing.
“I think the pieces are in place that need to be in place,” he adds, “However, because all the connections weren’t thoroughly tested, we’re seeing the problems we’re having.”
“There have been errors with the applications going to the carriers,” Schuyler says. “That’s another area that needs to be addressed and resolved quickly. Those applications, when submitted in large volumes, have to be 100 percent error-free because the carriers don’t have the time or resources to go through the applications one at a time to resolve those errors.”
Carricarte agrees, “If you get an incomplete application received by an insurer, it’s incomplete. Any errors on that application, and they can’t assume the risk. They’ll either deny it or send it back as incomplete. No one has said much about the data integiry issues, which I believe will become a much larger problem as the initiative continues.”
And these data issues are costly, wasting time and money across the spectrum.
“The escalating costs aren’t just for project-specific costs, they’re all down the line,” Melnik says. “The insurance company has to undertake additional steps checking and correcting data; and the consumer may need to go back and fill out the application again; and everyone is spending all this time doing duplicate work.”