3 ms·
Is it just me, or does even the 'bargain' price of $647/enrollment from the federal website... seem surprisingly bizarrely high? (by an order of magnitude or mo
by jrochkind1 13y ago
Is it just me, or does even the 'bargain' price of $647/enrollment from the federal website... seem surprisingly bizarrely high? (by an order of magnitude or more)?
But I'm not really sure exactly what to compare it to as 'typical' for website costs. Any ideas?
And of course, presumably (hopefully) the cost will continue to be amortized over the next few years -- more money will surely have to spent on it (continually), but the overall cost per enrollment should theoretically go down, as the initial work continues to be used for more and more enrollments. At least theoretically; in reality, who knows.
- _delirium 13y ago> But I'm not really sure exactly what to compare it to as 'typical' for website costs. Some kind of "enterprise" sales-management or HR/benefits-management system might be a reasonable comparison. I wonder if there are published per-user or per-sale numbers on how much a system like that in use at a big organization (IBM or Boeing or Siemens or whoever) would typically cost.
- HillRat 13y agoIt's really quite ridiculous for the linked report to use a capital:user ratio as a meaningful metric -- it's a bit like building an auto factory, building the first car, then arguing that the car cost $100m to build. Having said that, it's obvious that some states, like Hawaii, are going to fight an uphill battle to make a cost-effective case for their revenue, while states like Oregon failed spectacularly to make good on their requirements. OR's new CIO, Alex Pettit, just recommended killing the whole Oracle-belabored mess and moving the state onto the federal exchange. Pettit doesn't have any blood on his hands from the Cover Oregon debacle (he was hired away from OK after the exchange's failure), but it's still a gutsy move to kill it, since he's the one who has to deal with the politicians. (Plus he's not making any friends at Redwood Shores, and Oracle is going to outlast all the politicians.) I would take away from this the lesson that the political desire to use states as "laboratories of democracy" is sometimes the completely wrong move; one critical, massive, high-risk and poorly-managed rollout was enough, thanks. Adding fifteen more projects on the HHS dime to replicate functionality (leaving aside state-level Medicaid systems) ended up driving up costs and adding more failure points to the system. Ironically, the states that were most resistant to the ACA contributed the most to federalizing healthcare enrollment, a centralization that might be politically antithetical to those politicians but that will probably save time and money in the long run.
- fennecfoxen 13y ago> I would take away from this the lesson that the political desire to use states as "laboratories of democracy" is sometimes the completely wrong move; Well, in this case, since the mandate was essentially "implement this very specific federally-mandated policy monoculture!" there's really not a lot of gain you could hope to have from the Laboratory model. If the states had the freedom to vary health care policy in meaningful ways, it would be a different matter.
- r00fus 13y agoIs the requirement to setup an health insurance exchange (or grant residents access to the federal exchange) a "monoculture"? If you're informed about the situation, what parts of the state exchange requirement are "onerous" when compared with the status quo (i.e., what the insurance industry already requires)?
- anigbrowl 13y agothe political desire to use states as "laboratories of democracy" is sometimes the completely wrong move I would upgrade 'sometimes' to 'most of the time.' There's a thin line between responsiveness and political feudalism, with the attendant corruption.
- incision 13y ago>"And of course, presumably (hopefully) the cost will continue to be amortized over the next few years -- more money will surely have to spent on it (continually), but the overall cost per enrollment should theoretically go down, as the initial work continues to be used for more and more enrollments. At least theoretically; in reality, who knows." You'd think that. Even if the entire population of a given state were enrolled some of these costs remain bizarrely high and highly varied. DC: $211 Hawaii: $147 California: $28
- dllthomas 13y agoIt looks like these numbers are roughly consistent with a large, similar fixed cost spread over different population sizes.
- us0r 13y ago"s it just me, or does even the 'bargain' price of $647/enrollment from the federal website... seem surprisingly bizarrely high? (by an order of magnitude or more)?" Welcome to government contracting. The few who play the game basically get away with whatever they want. It is quite ridiculous. The cost is only going up now that states are essentially pot committed. After the initial disastrous implementations across the board I doubt anyone is in a hurry to try that again. I have the original CA proposal. $360,334,374.00. I’m sure there is at least another $50,000,000 in contract mods. The waste is just unreal. Creating local jobs? Not tech ones. The site was done in India. Something like this sounds like a very good use case or AWS or another cloud service given the enrollment periods. Of course there is no profit in that. From their proposal: “As discussed in Section 2.5.1 Hosting, we plan to implement a Private Cloud type infrastructure for CalHEERS. Why Private Cloud? Since standard cloud offerings, available from Amazon, Verizon, and a number of other service providers, offer sufficient capacity and are immediately available, they may seem like a good answer. However, standard enterprise-level cloud-based solutions currently focus on commoditized functions such as ERP, CRM and to a lesser extent HR and payroll. While SLAs are offered for security and availability, they do not meet the Exchange’s requirements for secure, SLA-based data intensive operations”. The solution? 1,714 cores, 9,394GB of RAM, 392,327GB of Storage over 249 virtual machines. Every single piece of software is Oracle with the exception of Adobe LiveCycle. Even with an average billing rate of $150/hr, more processing power then most Top 100 sites on the Internet…. they couldn’t keep a fail whale page online let alone any functionality. Once it is finally somewhat usable they are treated like a raving success and awarding with… the federal contract. Quite a business.