UK pins hopes on OSS to avoid Healthcare.govs sick site syndrome
A tale of two governments: UK healthcare tech system reforms highlight where America went wrong.
As demonstrated with last month’s case of UK government mangling
agile IT implementation, when bureaucrats get their fingers in
tech pies, the result can be less than sweet.
Case in point: When Obama’s controversial health
care reforms came into place, legions of opponents leaned back and
waited for some sort of apocalypse scenario. And lo, it came to
pass, Healthcare.gov, the site built to ensure millions of
Americans could finally gain access to affordable healthcare (boo,
debuted to a chorus of resounding glitches.
Within hours of going online, the internet was awash with
complaints that drop downs weren’t working, it was impossible to
log-in, and pages were failing to load.
It wasn’t necessarily the brutalist Soviet
nightmare doom mongers were hoping for, but it was enough to
generate headlines like,
“Obama Spent $634 Million to Build Broken Healthcare.gov
“Head fake! Is Healthcare.gov only an empty shell MOCKUP of a
working Obamacare exchange?”.
Not quite. Actually, what had happened was a
perfect example of what can go wrong when you outsource. As
Verge points out, this website comes from
the same people that staged one of the most technologically
sophisticated election campaigns of all time. Healthcare.gov
however, was not created by the elite team of technicians that
powered Obama to vicory. The bulk of the work was done by Canadian
firm CGI Group, who subcontracted up to
15 other firms for different aspects of the
It wasn’t all bad though. The front end of
Healthcare.gov, which launched in June, was built by Aquilent, who
then subcontracted Washington based startup Development Seed, who
are very much adherents of zeitgeisty Silicon Valley practices.
Working on agile principles, the team published their code on
GitHub, and ultimately created a product that married form and
function beautifully, whilst allowing for speed, scalability, all
at minimal cost.
In entrusting a closed-source proprieter to take
on the back end of what promised to be an incredibly complex
project, the Obama administration adroitly demonstrated exactly why
awarding contracts to a solitary bidder can be a terrible idea. The
best pitchers certainly aren’t always the best people for the job –
and there are an awful lot of companies out there who are great at
landing contracts, but not so skilled at following up on the actual
Any project on this size was guaranteed to have
some teething issues, regardless of how is was developed. But,
thanks to their poor decision making, the US government is now
saddled with a system riddled with bugs that only the CGI Group can
fix, and a system that is fundamentally poorly
designed - something that’s a lot mere serious
than a few initial glitches, and far harder to fix. And while
problems persist, the many fierce opponents to the project are
smugly nodding their heads and ruefully sighing, ‘I told you so’,
as what promises to be a game changing resource for so many
Americans fails to take flight.
Over in the UK on the other hand, the government
has learnt from the tech sector at large, and is embracing open
source. It was announced this week that when the National Health
Service main secure patient database and messaging platform Spine
upgrades, it will move its core from Oracle to a smorgasbord of
open source providers, including Riak datastore, Redis, Nginx,
Tornado, and RabbitMQ. Contractor BJSS will help manage the
rollout, in contrast to the US mega consultancy approach. The
primary development language will be Python, with some
Migrating Spine from a closed-source proprietary
project to an system centered on open source technologies is in
keeping with the UK Department of Health’s IT Strategy, which calls
for the adoption of open standards to increase interoperability and
interconnectivity between government sources.
Additionally, delegating infrastructure
responsibilities to primarily open-source companies over one IT
oligarch also means that costs won’t be solely absorbed by one
giant corporate sponge – minimising the risk of the astronomical
cost related heckles currently being thrown Obama’s way.
Before Spine 2 goes live, a complete redesign of
hardware, software and code will have to take place. However, after
the initial hump, in the long run this should allow for a more
limber service that can be easily tested against by third-party
suppliers, as well as reduced deployment costs.
With the throb of discontent humming from across
the pond, this was auspicious timing for the NHS to announce their
break with Oracle. And, should the UK government accidentally
follow past form and develop a crooked Spine 2, at least there will
be plenty of scope for fixes – which is sadly not the case for the