He should really be wearing latex gloves.

The key driver for the Affordable Care Act was a statistic that showed that 47 million in the US (of a population of 300 million) were uninsured, many of these children.  In order for the Affordable Care Act to be financially sustainable, 30 million will need to sign up.  The goal is to have 7 million enrolled by the end of March, 2014.  How are we doing?

So far, we are just over 20 days into the enrollment process, and official numbers state that 476,000 citizens have begun the enrollment process, filling in a form partially at least.  Estimates state that 70% or more exit the site without completing a form, which makes rough sense given that 19 million visits to the site have been logged.  A phone number alternative has also been provided for those finding the site hard to navigate.

Web Design 101

Many problems with the site have been reported:  slow response times, errors, being kicked out of the site after getting so far, bad data being sent to healthcare providers (duplicate forms, missing fields and spouses being listed as children are just a few), inability to link with some healthcare providers, and state sites that differ in quality from the national site (the worst reported is Hawaii).  State web site quality correlates closely with the state’s support of ACA.  What went wrong?

The first answer to this question was pretty unsatisfying:  that due to overwhelming demand for this popular product, the site couldn’t handle the high volumes.[1] Given the goal to enroll 7 million by end of March 2014, the site should be able to handle at least double the volume that has been presented so far.  Assuming 182 days of enrollment if equally distributed (which never happens) would mean that to date, 770,000 citizens would have already enrolled, not the 467,000 who have made some partial attempt to do so.  Also, bear in mind that Amazon handles 70 million customers per day.  There are plenty of tech companies that know how to handle volume.  As Donald Trump rightly pointed out, there are many US tech companies who want this to succeed who would probably even fix it for free.

The first and most obvious scapegoat is the web developer, a French Canadian company called CGI, based in Montreal.  CGI has done a lot of government contract work, the majority of their burgeoning business since 2008.  They are a mixed bag when it comes to healthcare work.  Predictably, and probably rightly, CGI points to the fact that they met all their deadlines, and that it was the late (and perhaps inaccurate or incomplete) specs that are the culprit.  Of course, I never met a tech person who didn’t say “It’s working as designed.”  Additionally, CGI points to the difficulty of retaining top talent, a weak excuse I have used myself in business on many occasions, always unconvincingly.  CGI also pointed out the problem of changing requirements.

Just what was their responsibility?  They were the top contract (of 55 total contracts) awarded at $88 million of the total $394 million budgeted; the next highest was awarded $55 million.  A recap of CGI’s role:

Along with building sites for the states that elected to participate in the federal marketplace, it’s in charge of knitting all the pieces together, making Quality Software Services’ data hub work seamlessly with Development Seed’s sleek user interface and Oracle’s identity management software — or “designing an IT solution that is adaptable and modular to accommodate the implementation of additional functional requirements and services,” as CGI Senior Vice President Cheryl Campbell put it in a September hearing on the progress of the site.

Even Disneyland ain’t cheap.

The site serves citizens in the 36 states that opted not to do their own sites (including the populous TX and FL).  In the first week of October, 24 of the 36 states’ sites received error messages.  The remaining 14 states have varying levels of quality in their own state exchange sites.  There are plenty of potential scapegoats on this project:

  • Underfunded.  Donald Berwick, who administered the federal Centers for Medicare and Medicaid Services in 2010 and 2011, says the site failed because they didn’t have enough money to build it from the start.
  • Late specs.  Others point fingers at the Department of Health and Human Services, which took years to issue final specifications, preventing CGI from really getting started until this spring.  CGI seems to confirm this, stating that requirements were dynamic (changing) during the design process.
  • Bad build.  IT specialists who evaluated the site said it was “built on a sloppy software foundation,” potentially due to the haste with which code was written.
  • Project complexity.  Slate described just how difficult it is to coordinate multiple contractors creating different components of a complex Web portal.

One staffer at CGI complained that the pressure within their company to address these issues has been enormous:  “There’s a lot of frustration.  People are getting sick, fainting in conference calls.” [2]  Hopefully, they have good health insurance to help these distressed employees out.

If We Build It, Will They Come?

Harry Potter? He’s also not an American, so he doesn’t need it. Plus, he’s a wizard.

When polled, an estimated half (47%) of the uninsured said they were “not at all familiar” with the health-care exchanges (71% said they were “not too familiar” or “not at all familiar” when combined).  Those numbers have been static since before the Oct. 1 start date.

There are some among the uninsured who prefer to pay the fine rather than enroll, banking on continued good health.  A Gallup poll from September showed that 65% of the uninsured planned to get insurance vs. 25% who planned to pay the fine.  That poll is going the wrong way with October results showing only 56% now plan to get insurance while 34% are prepared to pay the fine.  But perhaps this will change over the coming months:

The best evidence we have about enrollment patterns comes from a New England Journal of Medicine paper written by economists Amitabh Chandra, Jonathan Gruber, and Robin McKnight. They examined the Massachusetts reforms, during the first year, to see who signed up and when. Most people waited until the last minute—the point at which delay meant incurring financial penalties under the mandate. Relatively speaking, the latecomers were more likely to be young and healthy people, the ones whose support the system needs to function properly. That’s good news, because it suggests the most tentative consumers—the ones least likely to wait out website delays—haven’t started shopping yet.

My own experiences abroad have varied when it comes to health care.  I found other countries to be competent and reliable, but wait times to be longer with more onus on the patient and less expectation that money can cheat death.  Medications and services are generally incredibly cheap compared to the US, based on my time in Spain and Singapore, although one medication I bought in Singapore for S$60 a tube was  available in both Nepal and India over the counter for less than a dollar!  I have found care to be generally quite good in the US with more open discussion of issues and more options given to the patient, but overall incredibly expensive.  However, at least one UK woman found US care to be sub-par compared to her expectations in a hilarious and horrifying article.

  • Do you plan to use the health-site exchange to get a better rate?
  • Do you know people who plan to pay the fine rather than get insurance?  Do you agree with their decision or disagree?
  • Will the ACA improve health care overall in the US or merely address the issue of the uninsured?
  • Could these web site issues have been avoided?  Who bears the responsibility?

Discuss.

_____________________________________________

[1]   I’ve worked in 20+ years running call center operations, and I know this line.  It’s pulled out to deflect valid criticism of your faulty systems by pretending it indicates high interest in your product, never mind the fact that the faulty systems are often generating a lot of the volume through repeat attempts.

[2]   Honestly, I’ve never heard this “employees are fainting” excuse in the US, but I have heard it plenty in India and Canada; I am always skeptical.  I had colleagues who watched planes crash into the twin towers who calmly gathered their belongings and left the building.  Fainting because you built a crappy web site?  Really?  Drink a Red Bull, people.  While you’re at it, get over yourself.