By Michael Olenick, a regular contributor on Naked Capitalism. You can follow him on Twitter at @michael_olenick
This piece about my attempts to enroll on healthcare.gov runs the risk of being long, frustrating, and potentially repetitive, but that simply reflects the experience itself. But it also gives a taste of the nature of the problems and where the remedies might lie.
Much has been written about the rightfully maligned healthcare.gov website though little of that involves people, on the federal exchange, who actually try purchasing health insurance. Unlike contractors and federal employees who do not use the site, I have the misfortune of falling into this category. Since I’m also a software developer, I thought it worthwhile to document this strange experience.
I initially tried to sign up October 1, the day the exchange site opened, and every day after. Initially I’d sometimes get to the third screen of the account creation process – the one that asks security questions – but the questions were blank so I was thwarted from creating a login. One time the questions were there and I did create a login but I was never able to use it again.
Coming to the conclusion that the system reserved the username but lost the password – which later came out in the popular press but was never acknowledged by the government – I tried a new username. Finally, after weeks, I could log in. My sense of accomplishment was palpable: soon I could view unaffordable mandatory health insurance plans that cover 60-70% of my costs but include guaranteed insurance industry profits. This is progress!
After answering some odd questions to verify my identity the system told me I might not be who I say I am. My hopes were raised – maybe a random person wanted to purchase health insurance for my family and pay the full price – but I had to call a phone number to progress. After an hour on hold with Equifax a phone operator confirmed the system actually did verify my identity but, for whatever reason, told me to call in anyway. Apparently this is a common “glitch.” I’ll admit that the Equifax rep displayed the patience of Buddha, single-handedly increasing my feelings of goodwill toward Equifax, which was not hard because I had no goodwill toward Equifax before the call. However, I never figured out why a credit bureau is involved in taking applications for health insurance.
Being verified I marched onward in my quest for medical insurance.
Let’s take a breather for some background about how I found myself tangled in the governments half-billion dollar boondoggle of a website. I work as a research fellow for a European business school but I am based in the US. As an external contractor health benefits aren’t included so I get to buy health insurance for my family.
My state of Florida – led by Tea Party Governor and former healthcare executive Rick Scott – opted out of building a state-based insurance exchange so I use the federal exchange, or try to at least. Florida doesn’t have a state exchange because Scott, who paid the largest fine in Medicare history before being elected Governor, couldn’t find a way to make a buck for himself or his cronies so opted out of building a state exchange. Apparently, Scott didn’t think of creating an overpriced “consulting” firm, the latest rage in government scams.
Unlike many complainers, I largely support the Affordable Care Act. It should be a single-payer system, but I’m willing to accept the idea that a great journey begins with a single step and see the ACA as a start.
Getting back to the website, I eventually finished entering the pertinent information for my small family. I electronically signed my application, expecting to see plans, but instead was told the application is “pending.” Given our ages, medical conditions, and lack of eligibility for a subsidy. I half expected Obama himself to show up with a list of plans; we are the target audience. Instead I waited days, finally digging back into the application to be told me there was missing information. I went through the same process again, giving the same information about my family. This was easy because, for the most part, the “missing” fields were pre-populated from my last time through.
In my second try I was able to fix one especially weird glitch from my first attempt. That is my wife was listed as my spouse, and my daughter was listed as my daughter, but the system thought my daughter was also my wife’s spouse. Never mind that our daughter is seven years old and my wife already had a spouse (the system appears to be open minded, I suppose). In my second attempt, the fields were editable and I was able to correct this anomaly.
Again days passed with no emails, phone calls, knocks at the door, or any indication things were progressing. My wife asked me what’s happening and I showed her, submitting the same data a third time. Days later I tried a fourth time, from an iPad, with predictably dismal results. I thought I’d check again on the status while writing this piece but, on a Sunday at 11:18AM, I’m told that the system is down after I try logging in. Duh! .. there is obviously nobody who might want to try signing up for healthcare as part of their Sunday morning chores. [Follow-up while editing: it eventually came back up as dysfunctional as ever; “they” tell us Verizon can’t keep basic web servers alive, which is difficult to believe.]
The government apparently spent about a half billion dollar on this mess and, though it’s there seems to be an excuse for each and every dollar spent. It’s high volume, went the initial response. When Rep. Anna Eshoo, Democratic ACA supporting Representative from Silicon Valley pointed out that it’s much lower volume than many mainstream systems, the Illuminati of healthcare pivoted. Actually, beams the 1984-style government announcement, it was never really a volume problem but the data is very complex. This morning, during Congressional testimony, it sounds like the bureaucrats have pivoted again and that they are back to high volume.
All this struck me as odd, especially for the unsubsidized plans, because there really is nothing substantively complex about it. That’s the beauty of the ACA: preexisting conditions do not matter so there is actually much less information to collect than web-based systems insurance companies wrote that have existed for well over a decade. If people need a subsidy the system must check with the IRS, but banks do that routinely now when issuing mortgages. I read the system must also check with DHS, though – like the credit bureaus – can’t fathom why (if DHS is looking for terrorists in hospitals I’d suggest they skip patients and go to the billing department). In any event I don’t qualify for a subsidy so none of these checks should have been necessary – just show me the plans I’m expected to buy and let me purchase one – but the system still bombed out.
At this rate it might make sense to process people by hand, one at a time. If Fannie, Freddie, and the FHFA can process mortgages, which they can, those agencies implementing the ACA should be able to forward healthcare applications to private companies, which they clearly cannot.
Few people point it out but this fiasco is the predictable result of rigged government contracts. To put things in perspective I looked for some other outstanding government contracts to see how they rig “open” bidding; the methods quickly became obvious.
On the “RFP-EZ” marketplace there is one RFP to build a reservation system for the national park service, a straightforward project a small software business could bid for. Except there is one odd requirement: the system is still up for bidding on October 27, 2013, and it must launch between January 1 and February 1, 2014. That is, there will be no more than two months, including winter holiday, to build and test the system after contract approval. It’s almost as if somebody already built a system and now just needs to win an impossible bid in order to get paid. Almost…
Similarly there’s a project to provide a database of all publicly rated debt issues and issuers, including all historical ratings, credit migrations, and defaults. Successful vendors will not only keep up the data in real-time but will also keep the structure of the deals affected, and store the data so that it can be converted into many other programs. Finally tech support includes not only the use of technology but also “guiding OCC users through the use of the data.” Something tells me that the OCC – an agency that’s arguably the pinnacle of government incompetence, corruption, and waste – already has somebody in mind for this gig.
Lambert and NC readers questioned whether open source could have prevented this mess. In short .. I think so, or at least that open source methods may have helped (OK .. at the very least they wouldn’t have been harmful). On launch date Hacker News, the popular community for the geekdom, dissected the tiniest bit of healthcare.gov and found they were still using test code post-launch. By noon on launch day the tech community found specific problems, and suggested specific solutions, that the half-billion suited up DC contractors apparently did not notice for months or years. Granted it was not that hard because the files were marked “test,” but at least they took a moment to check, unlike the contactors paid handsomely to do the same.
Now the administration tells us that the misnamed Quality Software Services, Inc. (QSSI) – the firm that apparently botched both the signup and verification process – has been hired to fix the problems. This time will be different, I’m sure. Maybe they’ll try something new, like outsourcing the project. We know that did not happen before because they testified to Congress they did not outsource healthcare.gov. Of course, just like it only takes a moment to find the test code it doesn’t take much longer to find that QSSI applied for 23 H1B foreign worker visas from October 2011 through September, 2012. Most of these positions were for computer software engineers, and all but one are for workers from India, most to be paid $80,288/year. Maybe those are for a different project, though in that case QSSI miraculously could find people for healthcare.gov but couldn’t for a different project. Similarly, the other contractor CGI Federal applied for 13 positions during the same time period though with more geographic diversity: 10 from India, one from China, one from Bangladesh, and one from Venezuela.
Well, it’s not like there are out of work older American programmers who specialize in this field, except that there are. Unemployment for older programmers, oftentimes with backgrounds in large systems, is actually epidemic. However, that’s a different story though one deserving a lot more attention than it gets.
Since griping without solutions is not helpful I’ll offer some specific solution. Ask Amazon to build a healthcare sales portal, quickly. If Obama bet Bezo’s $100 Amazon couldn’t finish a functioning portal in a week I suspect they meet the challenge, in a week or less. They’d probably even be willing to process payments. Similarly Google and Microsoft might welcome the challenge, if nothing else just to show that they can. Government could have used healthcare.com (not .gov), an eerily similar health insurance marketplace that appears to work, though I’m assuming the owners did not know who to properly bribe (strike-out and replace with lobby). That is not because private business is always better than government, a widely misused argument for anti-government demagogues. But, in this very narrow and specific case – showing buying options from third-party businesses and completing a sale – the function really is the same and the private businesses have lots of practice.
One contractor during last weeks bloodbath in Congress said the government paid his firm $30 million and that he’s processed 3,000 paper applications. Given these economics, $10,000 per application, we could fly families to DC, put them up in style for a night, have a government and insurance company worker hand process an application, and still come out ahead. That’s facetious except, discussing the idea with Yves, maybe it should not be. There really is not much information needed to figure out the list of Obamacare plans. We know that census workers took much more complex information, at an accelerated pace, and did a good job. Maybe we should just switch to a paper system where helpers sign up every uninsured person in the US. That approach would have worked and might not end up being that expensive long-term, especially if the insurance companies would pay a referral fee to cover the temporary workers, like they do to brokers for new business.
The other virtue of a manual approach would be to understand the most efficient task sequence, which matters more than most people, including software developers, realize. From the New Yorker, emphasis mine:
The computers using Marchant calculators were working too slowly, so Stan Frankel, another graduate student in the computation group, recommended they try new I.B.M. tabulating machines, which used punch cards to perform calculations. Feynman agreed, but the I.B.M. machines and the technicians to maintain them were slow to arrive. In his 1975 lecture “Los Alamos From Below,” Feynman described his solution:
In this particular case, we worked out all the numerical steps that the machines were supposed to do—multiply this, and then do this, and subtract that. Then we worked out the program, but we didn’t have any machine to test it on. So we set up this room with girls in it. Each one had a Marchant. But she was the multiplier, and she was the adder, and this one cubed, and we had index cards, and all she did was cube this number and send it to the next one.
We went through our cycle this way until we got all the bugs out. Well, it turned out that the speed at which we were able to do it was a hell of a lot faster than the other way, where every single person did all the steps. We got speed with this system that was the predicted speed for the I.B.M. machine.
By organizing the human computers at Los Alamos as though they were I.B.M. tabulators, Feynman was able to match the speed of the I.B.M. machines. His “numerical steps” were literally a computer program, albeit a simple one carried out by people cranking noisy calculating machines. In 1945, Richard Feynman discovered something that remains true today: the problem of producing software is, first and foremost, a problem of organizing people.
Obviously a single payer system would have been much easier to implement: produce an SSN and proof of citizenship then you’re done. But that would cut out the layers of cronies. Accountability in the modern era being only for the little people this is obviously a non-starter. So we’ll wait to see if QSSI, who built a large part of the mess – and who botched testing it – can quickly repair it. Of course we could treat this like the census and encourage people to use paper, but that wouldn’t be much fun. I mean if a person called about “glitches” (read: serious software defects) rather than saying “come back in the next three years and we’ll fix them” tech support took their info and hand entered it. I’d rather not bet the health of my family to come up with this low-cost solution that is immediately available but that’s not my choice, is it?
All this would be funny in a snarky kind of way except that, of course, it is not. There will be people with preexisting conditions who cannot get insurance or who lost their insurance as a result of the ACA, and they will still not be able to purchase insurance by December 15 that will go into effect January 1 leaving a gap. They will be sick and end up in an emergency room and/or hospital, leaving with slightly better physical and exponentially worse financial health for no reason other than a lack of insurance due to the gross incompetence of the DC crony crowd. CGI’s Cheryl Campbell repeatedly crooned to Congress that she has insurance, as did overtly arrogant CMS employee Marilyn Tavenner. I guess since this does not affect them, or anybody who really matters, it does not matter what happens to the uninsured Americans paying for their latest escapade.
But the $10,000 per person cost to process 3,000 paper applications when one part of the system failed drives a key point home: a manual process being faster, more accurate, and cheaper than the computerized process shows that the Administration somehow saw technology as a magic bullet when it is just a tool. In fact, it’s becoming clear the project is rife with management problems; focusing on technical symptoms, although tempting, is trying to have the tail wag the dog. We’ll be returning to this topic in future posts, namely how well some more advanced (as well as not so advanced) approaches could have kept healthcare.gov out of this morass and which ones might help get it out faster.