One of the recurring themes over the past decades of developments in Computers, Web, and Telecoms, is the persistent failure of government IT projects. The propensity for failure has endured over technology change, whether it is the old style mainframe greenscreen applications, windowing desktop computers, or the latest multi-mode mobile interactive app. Time and again we hear the familiar tale of cost overruns, delays, and functional inabilities, sometimes leading to highly expensive and embarrassing cancellations.
One of the largest failures to date is the UK NHS National Programme for IT (NPfIT), at £12billion it attempted to offer a large range of services and facilities of an all-encompassing e-records system of patients, available to NHS staff at all times. Unfortunately for such a large heavily interconnected health record system its specifications were ill defined and its contract system was too rigid and onerous. In June of this year, E-Health Insider reported that 98% of the estimated benefits of NPfIT were yet to be realised.
The private sector is not immune and has it share of such failures. J.P. Morgan Chase the banking and financial services firms, reputedly lost $6billion on a trade executed on advice from a badly constructed value-at-risk model, built in Excel spreadsheet.
The implementers of government contracts usually have a large proportion of private companies in their midst. However, what sets aside government IT from the private sector are two factors: scale and ambition. Governments are by far the largest single entity in a country with the largest range of services, and therefore have the largest user base. By default, government projects have the greatest issues of scale with system sizes outside of the experiences of the private sector. It is often said by defenders of the public sector that only governments can tackle very large programmes, but this might be another way of saying that the private sector does not undertake such large programmes, because they know the full extent of the risks. Government mandated technical projects are also driven by abstract legislative frameworks, whereas businesses have measurable and quantitative goals.
We know that the scale of the programmes is an issue, but the main factor which dominates is the motivation and extent of the ambition from the top levels of the hierarchies who request such programmes. What appears at government cabinet level as a highly desirable, easy to implement, voter attractive system with short time scales, can so easily turn out exactly the opposite in all cases, in the cold light of reality. In a political world driven by sound bites, distanced from reality, infused with wishful thinking, and immune from costs, the provision of some web screens and a few database computers behind them looks all to easy.
This is not helped by the evident lack of experience found in our career politicians to business issues, contracts, IT, real world demands of tight time scales, and tight budgets. Politicians are also easily swayed by short term events and political compromises. This is not a good combination if they are setting the goals for long term, expensive, and substantial national impact programmes. If the target government programme is also directed at open ended services, such as welfare or healthcare, where the individual usage may range form nothing at all to lifetime continuous draw-down, then top level well-meaning requirements are bound to be insufficient.
The Affordable Care Programme is a $30 billion program for doctors in the USA to adopt electronic health records by 2014. The aim is an interconnected system of electronic health records to improve safety and reduce medical costs. It is no coincidence that the recent launch of this programme has exhibited some of the same issues which beset the UK’s NHS health record system. Further, like the UK NHS programme, a great deal of political capital is riding on the outcome.
The care programme system has not failed, but its much heralded launch has been plagued by highly visible front-end failures and systemic inadequacies of the multiple back-end servers. It was expected to process 39,000 transactions a day, but the launch date saw just six completed.
It is not as if the warnings from the UK experience were unknown:
“If a country like Britain — which already has a national health system and is a fraction of the size of the US — had so many problems with electronic health records, imagine the problems America would face. Here, instead of four companies competing for contracts, we have dozens of vendors — most with proprietary software — vying for billions in stimulus funds. It will be virtually impossible to make their products compatible, therefore not allowing all doctors in different offices to see the same patient’s health information.”
– Stephen B. Soumerai, Professor of Population Medicine at Harvard Medical School, and Anthony Avery, Professor of Primary Care at the University of Nottingham Medical School, January 2010
The performance of government IT programmes are touchstones for the maladies which infest modern day governments. IT systems are particularly sensitive to incomplete specifications, remoteness of the specifiers from the intended users, and have inherent complexities which are not always in the province of the management to resolve based on known experience. They are fragile in the face of unrealistic time scales, and are difficult to change.
The political mindset required in a society increasingly influenced by ubiquitous and immediate computer media is not that which has hitherto prevailed. The instant dispersal of views and the permanent mass records of minutiae scrutinised in multiple dimensions, has never before been available to the citizen. Governments must have a deeper understanding of both the technicalities of IT and its societal impacts to avoid the many recorded misapplications and outright failures already seen.
If our political system does not provide due recognition of IT for its 21st century nation, the electorate may conclude that it is because the state of politics is also unfit for the 21st century.