Freeing health's data

Even before the Francis report into mortality at Mid Staffordshire trust, the quality of NHS data was emerging as a hot political topic. Over the past year, the government and the Conservatives (but, surprisingly, not the Liberal Democrats) have been trying to outdo each other with promises to empower patients with openly available data about the NHS. As yet, however, there are few signs of the promised consumer revolution. The centrepiece of the government's efforts is the website www.data.gov.uk, launched last November by worldwide web inventor Sir Tim Berners-Lee and Professor Nigel Shadbolt of Southampton University in their roles as the prime minister's cheerleaders for opening up the vast resource of public sector information (PSI). The idea, originally floated by the Cabinet Office in its 2007 Power of Information programme, is to create a central clearing house for access to PSI data sets and information about what use is being made of them. The site emulates a similar effort launched in May 2009 by the Obama administration in the US, where there is a long-held assumption that federal government data should be freely available for re-use. It is too early to say whether the same culture will take off in the UK. Officially, of data.gov.uk's 7,500 data sets, 400 relate to healthcare. They provide the basis for a handful of applications already listed on the site, including iPhone apps for finding GPs, pharmacies and dentists from supplier Elbatrop. Another recent offering is Best Care Home, which helps people find the right care home for themselves, family members or clients from Care Quality Commission data. The site includes extracts and the full text of all inspection reports from the regulator, and ranks 18,500 care homes in order of quality performance. The bulk of NHS datasets listed on data.gov.uk come from England's NHS Information Centre, including hospital episode statistics (HES) and national care quality indicators. The Welsh health service has also contributed data and access to a mapping visualisation system, and Scotland has also released a limited number of datasets on areas including abortions and alcohol interventions. The launch of data.gov.uk coincided with that of a new licence model for re-using government data by the National Archives, replacing its pioneering, but little used, "click-use" licence. Unlike the click-use licence, the new "non-transactional Creative Commons approach" explicitly allows data to be re-used both for commercial as well as non-commercial purposes. Right to open data If the Conservatives form the next government, a sizeable number of new datasets could be made available under these arrangements. David Cameron, who can now count among his advisers Tom Steinberg of the MySociety web activist group (and co-author of the 2007 Power of Information report) is relying heavily on free data as a tool for improving public services. While decrying the "Google government" tag, he says the web "allows us to make big change in the relationship between government and citizens, giving power to people on an unprecedented scale". Cameron's most eye-catching pledge is to publish in full every government contract worth over £25,000. But he has also promised that health data "currently locked away in vaults" will be published "in an open and standardised format". The promised explosion in the availability of NHS data raises two interesting problem for the next government. First, if a new generation of commercial and community websites springs up on the back of the new data, what role remains for the NHS's own citizen-facing presence on the web? One pillar of the Power of Information agenda is that the state should not duplicate or compete with independent web-based services. The government has not yet followed this logic through, but a Cameron government might. An obvious way to cut the cost of the NHS Choices website, for example, would be to trim it to a core of essential data feeds and rely on the private and third sector to create citizen-facing applications. The second problem concerns the quality of information. The more use is made of NHS data sets, the more their accuracy, currency and comparability will come under scrutiny. Commenters at the data.gov.uk website have already observed that the most recent available list of GPs dates from 2006 - and covers only England. If as simple a list as the up to date GP register is impossible to open up, what hope is there for more sophisticated and controversial data sets? One key finding of the Francis report into the Mid Staffordshire scandal is that a working group led by Professor Sir Bruce Keogh, the NHS's medical director, develop a "single, clearer" measure of hospital mortality ratios for use by the NHS and its patients. The problem is that clarity and simplicity don't always go together with accuracy.