Data Privacy: Put ASHs in the bin

There’s a consultation about the regulations for protecting personally identifiable data. The government proposes allowing a number of local organisations to create secure Accredited Safe Havens (ASHs). They will have access to information from peoples’ personal care records, which could be used to identify an individual.

The consultation assumes a solution that there is a need to transfer such data into an ASH.  Quarkside suggest that an alternative is inherently safer. Instead of moving data to an ASH, it stays put in a Personal Data Store (PDS). A PDS resolves the problems of consent by only releasing data for analysis without personally identifiable data.  This could be controlled by Mydex.

The back of a beer mat design goes something like this:

• People control their own health and care records in a suitably encrypted data store.
• Data is held in 5* format in triple stores and using URIs appropriately (ask Sir Nigel Shadbolt how to do it).
• Explicit consent has to be given for the extraction (or viewing) of any attribute. This avoids any data which could lead to identification being stopped at source. The consent could also be given by an Accredited Data Attorney (ADA). An ADA could be the person or any single person who has been trusted to give consent to release data for sharing purposes.
• If an Accredited Data Processor (ADP) wishes to use anonymised data then temporary rights are given by the ADA. Data may be given an expiry period after which any copies of the source data are destroyed. The ADP would be allowed to stored summarised data for analytical purposes.
• Any joins of personal data are done within the domain of the PDS and the method of performing those joins and hidden from the ADP. The risks are reduced for loss of privacy. If you go back to the principles of FAME you will see the nine principles that can make this work. The Identity Management problem is solved at source. Sharing data from multiple agencies is logically performed in an infrastructure that is like a walled garden.
• Each time data is released to an ADP, then the source identity would be irreversibly hashed by the ADA. The regulations would be so much simpler to implement.
• The ADA can release personally identifiable data to multiple agencies, such as health and social care. Again this must be time limited and the agencies would be obliged to destroy data, without any rights to store archives that contain personally identifiable data. A PDS is the repository for health and social care records.

Big data technology has advanced to the stage where this has become possible. Give control of sharing to the citizen. Acknowledge that people have ownership rights to their data, even if it is collected and stored by the NHS (or any other ADP). If you don’t create ASHs, you don’t need to regulate them

Integrating Social Care and Health

The Government has mandated the integration of health and social care services by 2018, potentially imposing a financial penalties for siloed services.  Good.  There is a budget of about £1bn to ensure there are integrated projects in every part of the country by 2015.

Who is going to lead it in locations where services are delivered?  How is the funding going to be allocated?  These questions do not seem to be answered.  Quarkside suggests looking again at the Framework for Multi-agency Enviroments.  It does not give the answers – but it has a method for bringing all the issues together.  Perhaps the Local CIO Council may give some leadership.

The elderly and their families are most likely to benefit.  Two separate industries, for telehealth and telecare, are growing rapidly.  The new political direction now opens an opportunity for telehealthcare, extending the period older people can remain safely in their own homes.  An integrated service should support all levels of the Kaiser Pyramid.

Observe the need for technology that has to transition from Self Care to Professional Care.  Be aware of the interoperability requirements as more complex monitoring has to be added with increasing risk.  Families will need telehealthcare products that are simple to use.

QA may madden Maude

The 2011 Government ICT strategy preaches standards.   Tick box = Good.  People who bore for standards preach, ‘to do it properly you must define the standard and check later that the standard has been followed’.  This blog compares the strategy against a standard (standard with a small ‘s’) – in this case against the same set that was used to review the SOCITM ICT Strategy, released in draft last month.

The target for all public sector ICT is established in the introduction:

“6. Information and communications technology (ICT) is critical for the effective operation of government and the delivery of the services it provides to citizens and businesses. It offers key benefits by enabling:

• access to online transactional services, which makes life simpler and more convenient for citizens and businesses; and
• channels to collaborate and share information with citizens and business, which in turn enable the innovation of new online tools and services.”

Everybody must agree with this, and observe that sharing information across multiple agency boundaries is critical for citizens, businesses and agencies.  It has led to much discussion about shared infrastructure, shared services and the benefits this will bring.  Fortunately, we can use a standard for quality assuring the Strategy and highlighting any gaps that need to be addressed.  It has nine dimensions for assessing multi-agency information sharing partnerships.

• Business Scope and Plans
• Governance
• Legal Issues, Policies, Rights and Responsibilities
• Information Sharing
• Identity Management
• Federation
• Transactions, Events and Messages
• Infrastructure
• Sustainability

Overall these can be summarised into Process, Governance and Technology – the Quarkside mantra.  A quick traffic light assessment against the standard dimensions is as follows:

• Business Scope and Plans: Amber

The reasons are good and there is an aggressive, but risky timeplan.  Dependence on on word ‘Agile’, is a recipe for systemic obscuring of progress.  It provides opportunities for hiding problems that only emerge when the end-users in multiple location are expected to change time-honoured processes, and new systems are not interoperable with old systems.  The needs of 450 local authorities must not be ignored.

• Governance:  Amber

A structure has been developed, but it omits the input of local delivery agencies, such as local authorities.

• Legal Issues, Policies, Rights and Responsibilities: Amber

Apart from the Policy, other issues are not raised

• Information Sharing: Amber

Use of open standards and APIs will help at a programmatic level, but additional useful services, such as Master Data Management and systems interoperability standards are not mentioned.

• Identity Management:  Red

Avoidance of a cross public sector strategy for citizen, employee and agent identity management risks complete failure of the strategy and policy objectives will not be met.

• Federation: Red

Federated trust by all involved agencies is vital for both accuracy and efficiency.  Nowhere is this mentioned or implied.

• Transactions, Events and Messages:  Green

Operational systems usually find technical solution for inter-system data transfers.  The use of Web services on the Cloud should help.  Channel issues are addressed

• Infrastructure:  Greenish

The overwhelming weight of the document is technology and infrastructure, there are eleven actions planned.  However, one suspects that the thought process has ignored local government and external agencies in the calculations.  Are local authorities expected to reduce ICT costs by 35%?

• Sustainability:  Red

The standard means to ability to sustain a shared service for operation over many years, not reducing carbon usage.  Most shared services fail because of the inability to agree funding for operations, and all the development investment is wasted.  Central Government must agree a sustainable funding model at the very beginning of every information sharing project.  The Cabinet Office should feel responsible for the whole of the public sector, not just central government departments and agencies.

So how do you react to 3 Reds, 4 Ambers and 2 Greens?  It is low on Process and Governance and higher on Technology.  Quarkside thinks it is good enough for a first draft to get the ball rolling.  But if Francis Maude thinks this document is going to deliver all his policy objectives, then I fear that he, or his successor, is set for a big disappointment and some explaining to do.

Shared Services Swamp

There are fewer partnerships formed in the public sector than might be expected for maximum efficiency. Part of the problem is that partners have to start building partnerships from scratch. It is like a fresh learning exercise and many fail by missing some factors that are critical to success. People need to understand and document the risks from the very beginning.

It was a continuous thread in the latest Government ICT Conference (Jan 26th 2011).  Ample evidence from case studies from HMRC, DCMS, DCLG, Cabinet Office, Leeds, Kent and Herefordshire showed that different approaches can produce huge benefits.  But they all used ‘s’ word (standards with a small s).  We were informed that the Government is spending more time deprecating standards than publishing them. Improving outcomes is the target, interoperability is seen as key – and the Information Commissioner will not get in the way of data sharing.

Even though most agencies will accept technical infrastructure standards – they won’t co-operate on information governance standards.  There is no longer a central clearing house.  Organisations simply set up silo standards without consideration of how useful information might be in another silo.  Methods that would help, such as ISO 18876, are unknown or ignored.  ICT leadership, that is the Government CIO, should publish a Standards Policy.

A final example of head in the sand is that many years ago there was some research done on a standard for partnership formation.  At a cost of several million pounds from DCLG (ODPM), pilot projects completed and a simple Roadmap process was developed and tested. The findings are even more relevant today.  Look at this document to see how easy it should be.  It is only 4 pages long (plus appendices).  The executive summary is only half a page.

If all partners start singing from the the same, standard, hymn sheet – then progress towards efficient, sustainable, shared services could be much quicker.

Partnership Pie. Recipe Revealed

Independent review of performance and local accountability is needed to reduce the costs of internal public sector bureaucracy – or so suggests the Centre for Public Service Partnerships.   They responded to the HM Treasury and Cabinet Office call for evidence on public service reform.

Partnerships have to be the key to improving efficiency through shared and sharing services.  Quarkside hinted at this in November in relation to cooperation between the NHS and local authorities.  A key enabler is a process for building adaptable and sustainable partnerships.  The groundwork for a viable multi-agency framework was completed in a DCLG national project.  There’s also a roadmap to use at the initial stages of building a partnership, showing partners the critical success factors and levels of risk.  It can be done in a standard (ie common) way across any (or all) recipe(s) for partnership pie.

The full response is worth a read, but the last point supports Quarkside’s pressure for improving public sector governance:

18. What specific data or information would you like to see made publicly available in order to help individuals and local communities hold services to account?

There is a need for some standardised data for public service performance and costs which is comparable between agencies and services; verifiable; independently audited; and accessible and understandable to the public. This data should be available irrespective of the provider of the service and/or its commissioner or procurer. Prime accountability for public services should be either directly through an economic relationship between provider and user or through local government accountability

Yesterday’s blog was a prime example of how NHS bureaucracy only succeeds in baffling the public, hiding the facts and costing a fortune to operate.  It is hardly an easy model to follow in multi-agency shared services.

No excuses, LAs and NHS must talk

A good news story for a change. Simple connectivity between local authorities and the NHS has been needed for years.  I recall having to spend 15 months with the agreements and protocols to join 2 servers one foot apart in a server rack.

Joe Harley (DWP) and Christine Connelly (DH) have jointly published a letter encouraging local authorities to connecting to the NHS N3 network via their existing GCSX connection.    They explain that it is now possible for local authorities to access NHS Spine without the need to install a separate N3 connection.  The reverse is also true; NHS organisations can access local authority data via their existing N3 connections.

Everything these days come at a price, however.  The N3 Interconnect Service is an additional service charge to the GCSX service.  Expect to pay £5,760 for 10Mb.  There’s also value in sharing the service between LA partners: £7,200 for 4 local authorities (£1,800 each) then £1500 per additional local authority using the same aggregation route.

Now that the connectivity should no longer a barrier, we should revitalise the attempt at building information sharing partnerships and utilising systems interoperability standards, such as ISO 18876.   Exchange of data between Health and Social Services has long been the call of Enquiries into Victoria Climbié and Baby Peter.  One key excuse for non-cooperation has been removed.