Open - The Scottish Government

advertisement
Joined-up data for better decisions :
A Strategy for improving data access and analysis
REVIEW: The Data Linkage Framework - April 2014
1. INTRODUCTION
Following extensive consultation with stakeholders and deliberative research into
public opinions, Joined Up Data for Better Decisions: A strategy for improving data
access and analysis was published in November 2012. This document set out how
the strengths and successes of the Scottish Informatics Programme (SHIP) would be
expanded beyond the health sector. The Strategy was published alongside the
Guiding Principles for Data Linkage which were recommend as the basis for all
decision making relating to data linkage for research and statistics.
Four main deliverables – or ‘blocks’ – were outlined within the Strategy. Collectively,
these are termed the Data Linkage Framework and comprise:
1.
2.
3.
4.
The Guiding Principles for Data Linkage
A Privacy Advisory and Ethics Committee
Information Gateway and Linkage Service
The Steering Group
The original Steering Group which developed the Joined Up Data for Better
Decisions Strategy evolved into the Data Management Board (responsible for
providing strategic direction) and the Data Linkage Framework Programme Board
(responsible for operational delivery). As part of its winding-down phase, the
Programme Board committed to undertake a review, drawing together the views of
its members to assess the outcomes produced against the original four deliverables
of the Data Linkage Framework outlined above. This document reports on the
findings.
2. PROGRESS ON THE DATA LINKAGE FRAMEWORK
2.1
Overview
Over the course of the review, a number of recurrent and relevant points were raised
with respect to progression of the Data Linkage Framework. These include:

the general view that the Data Linkage Framework Programme Board was
well-run and the commitment of colleagues contributed to progressing work.

The landscape within which this work was originally set has changed
significantly.

The work to-date has been heavily dependent on the experience and
knowledge of the Health sector.

There was a lack of clarity around the roles, responsibilities and functions of
the different partners involved in the Framework.
There is little doubt that the landscape within which the Framework was originally
developed has changed significantly. The ability of colleagues to adapt to the new
environment and opportunities was applauded. However, it has meant that the
appropriateness of some of the original four Framework deliverables has diminished.
Given that the Strategy was based on how the strengths and successes of the
Scottish Informatics Programme (SHIP) would be expanded beyond, and given the
knowledge and resources in the health sector, it is understandable that the work todate continues to be heavily reliant on this area. However, this may act as an
obstacle to the broader progress of data linkage across other sectors. How the role
and influence of other sectors is increased with respect to data linkage is an
important strategic question which must be addressed.
From the outset there appears to have been a lack of clarity around the direction of
work and programme management following publication of the Strategy and the
Guiding Principles. This coincided with a change in all three Directors involved.
However, significant changes were made in how the Project functioned in mid-2013
with the establishment of the Data Linkage Framework Programme Board. The
Programme Board was identified as a delivery mechanism for the Strategy and
formalised the work. An Operations Group was established, providing a greater
sense of accountability to the Data Management Board (charged with strategic
oversight) and to each partner. However, some question whether the work of the
Data Linkage Framework Programme Board is greater than the sum of its parts.
A similar lack of clarity was raised regarding financing of the Framework.
Specifically, the relationship between the National Records of Scotland (NRS) and
the Data Sharing and Linkage Service (DSLS) is unclear, as is how the Office of the
Chief Statistician and Performance (OCSP) and DSLS functioned given that fiscal
responsibility sat with NRS (and was linked to the Census thinking for Beyond 2020)
but OCSP were looking for DSLS (within NRS) to drive work forward.
2.2
Outputs versus Deliverables
As previously outlined, the Data Linkage Framework is composed of four ‘blocks’
which were to be delivered and supported through a set of work streams. During the
Review process, members of the Programme Board were asked to give their opinion
on the relative success of each of these deliverables. These findings are outlined
below.
2.2.1 The Guiding Principles for Data Linkage were developed to assist data
controllers and decision makers when taking a proportionate approach to managing
the risks inherent in any data linkage. Three main points emerged during the review
process:

The Guiding Principles for Data Linkage were published at the same time as
the Joined Up Data for Better Decisions Strategy.

There is no evidence that the Guiding Principles are being used.

The Pathfinder projects have not made use of the Guiding Principles.
The Guiding Principles were based on and supposed to replace those developed for
SHIP. The continuing reference to SHIP is perceived as being unhelpful moving
forward. There are concerns as to whether wider UK interests (with respect to the
development on the Scottish Informatics and Linkage Collaboration (SILC)) can align
with the Principles. It is accepted that these are Principles, not requirements and this
may cause a lack of public confidence that data users will adhere to them.
The responsibility for promoting and implementing the Guiding Principles sits with
individual organisations across the public sector. The Information Commissioner’s
Office (ICO) have used and disseminated these at a UK level, supporting a cohesive
approach to data linkage. Anecdotal evidence suggests that those referring to the
Principles have been central to the development of the Data Linkage Framework.
However, the Principles are not being systematically applied by organisations.
It is anticipated that the electronic Data Research and Innovation Service (eDRIS),
Farr Institute and DSLS will use the Guiding Principles through the SILC
collaboration. There is also an assumption that going through certain organisations
(for example eDRIS) results in the ‘similar’ SHIP approach being applied, and
therefore - by default - the Guiding Principles.
Awareness around the Guiding Principles needs to be increased. OCSP is leading
this work, with sources of evidence and advice on using the Principles available on
their website. However, staffing vacancies has delayed this work. There are
concerns regarding a lack of engagement with existing large data linkage Institutes.
Similarly, the Health Informatics Research Advisory Group (HIRAG) reports to Chief
Scientist Office (CSO) and links to the National Research Service, which is also
promoting approaches to data linkage. It is not clear whether this overlaps with or
supports the work of DLFPB/DMB.
The Pathfinder Projects were established to demonstrate different aspects of data
linkage and the use of the Principles. Governance arrangements around the
Pathfinders was weak. The DSLS (as part of NRS) was tasked with delivering and
funding these projects. It was believed that their customers would conform to the
Principles by going through DSLS. However, the projects appear to have gone
through SHIP or SLS approaches, as DSLS paid eDRIS to undertake the linkage.
Thus they have not helped to progress the wider work on data linkage (e.g. as case
studies). It is not clear how (or whether) the DSLS encouraged or supported the use
of the Principles or helped partners to understand these. Moreover, the Principles
promote transparency, however this is not reflected in the Pathfinder work with little
information available on projects. Pathfinder researchers have been asked to review
whether the approach they took is similar to the requirements under the Principles. It
is unclear whether this has happened or the conclusions reached. It appears that
whilst the DSLS has been in place, few actual outputs have been produced.
2.2.2 A Privacy Advisory and Ethics Committee was proposed to advise on data
sharing and linkage and associated ethical, legal and social issues. Three main
points emerged during the review process:

The Privacy Advisory and Ethics Committee (PAEC) was not set-up.

The lack of a PAEC is considered to be a gap.

A PAEC is still seen as a useful approach to help enable a more collective
and supportive approach to data sharing.
It is understood that the decision not to progress with a PAEC was taken by the Data
Linkage Programme Board. This was seen as a practical decision which reflected
the complexity of establishing a single oversight body when there were already a
number of different regulatory and advisory bodies in place.
The lack of a PAEC is however considered to be a gap. Initial work from partners
had taken place under the assumption that a PAEC would be established and the
expectation is that applicants seeking data access will have gone through an
appropriate and identifiable process.
A PAEC is still seen as a useful approach to help enable a more collective and
supportive approach to data sharing. There were concerns that without a PAEC,
progress with data sharing may be too dependent on individual decisions by a wide
range of responsible colleagues. It was noted that progress is being made in some
social policy areas, but that this wasn’t co-ordinated between sectors and this lack of
co-ordination could significantly reduce the ability of the Programme to deliver in the
longer term. Colleagues recognised the legal responsibilities in place, but broadly felt
an overarching PAEC would be valuable.
There were a range of views on what a PAEC should do; from responsibilities
underpinned by a (new) legal basis, through to an advisory body to help inform and
support decisions of other data owners. The need for clarity of roles is considered
beneficial to data owners, those seeking access to data and also to the general
public who have an interest in understanding who is agreeing the use of data.
2.2.3 Information Gateway and Data Linkage Service was proposed to support
researchers and data custodians develop legal, ethical and practical linkage projects,
provide separated indexing and linking functions for cross-sectoral data and secure
access for analysis. During the review process, four overarching views emerged:

The landscape has evolved significantly.

The establishment of the Scottish Informatics and Linkage Collaboration
(SILC) has overtaken this work.

There is a lack of clarity regarding roles and responsibilities of those involved
in SILC in terms of work, funding, governance and legal accountability.

The funding issues for SILC are complex and may cause problems going
forward.
Work has progressed in a manner which has taken advantage of the developing
environment (e.g. Farr and UK Research Council funding). The initial intention was
to provide a single point of access or ‘gateway’ service where users could get
information on all the available datasets (e.g. through a website) or a person-person
‘linkage’ advisory service (e.g. eDRIS) approach. Changes in this work have made it
difficult to ascertain roles and responsibilities, accountability and whether anything
substantial has been delivered. Work on access to data has been focused on
meeting the needs of academics, and may be overlooking the needs of others (e.g.
SG, NRS, LAC). These developments fit with wider UK policy, ensuring a joined-up
UK approach, best use of resources and increasing the chances of success.
The DSLS was originally meant to bid for UK resources to run the Gateway Service,
but this was progressed by Administrative Data Research Centre ((ADRC) via the
Economic and Social Research Council (ESRC)). ESRC decided to fund one
Administrative Data Service. This is now Essex-based and provides much of the
Gateway function. The approach has evolved to bring together ADRC (academics),
eDRIS (Health), DSLS (public and voluntary sectors).
There is a Gateway: eDRIS will provide this function for the Essex Centre to support
access into the DSLS, ADRC and Farr (MRC) systems and vice versa. The Scottish
Government is putting £1m p/a into SILC for capacity building in this area. DSLS is
using eDRIS people/ATOS storage. ADRC and DSLS are operating under similar
(but different) Principles and processes which needs to be addressed. There is an
assumption that the Principles will be delivered through SILC. Regarding
engagement across public sector organisations (e.g. local authorities), researchers
using the service are also likely to be data suppliers. This work will be guided by
eDRIS, who will also advise where the data requested is and how it can be
accessed. Further clarity is required around the gateway, for example how this links
to Essex (via different websites or single SILC website with signposting). There is a
view that OCSP should take this and the public reassurance and information work
forward.
Researchers will have a single point of access to SILC via eDRIS (via phone/email)
followed by triaging for Farr/ADRC where a research co-ordinator will be allocated.
Both DSLS/ADRC and Farr have set-up joint projects to support collaborative
working and are working together to be used as a service, realising efficiencies
across funders and public engagement. There is a clear separation of indexing and
functions for cross-sectoral data, with NRS supporting DSLS and PHI supporting
Farr. Work needs to be done so that service users have appropriate access to linked
data; a model that looks beyond analysts visiting the site (e.g. via remote access).
Funding has recently been agreed by Ministers to fund 30 data linkage projects
supported by the Scottish government, and 30 for the public/voluntary sector (not
academic and not health). An internal decision making process will be employed for
these projects, which the Scottish Government Analytical Leadership Group has
outlined. Scaling up or down of the projects will be easier under SILC. It is
understood that the Scottish Government Chief Statistician is leading on this work.
SILC has developed by building on opportunities as they arise leading to the way it
has been set-up. Issues will include how individual organisations will collaborate so
that skills and ideas are shared. There is a view that it is not enough to simply colocate teams. Farr, ADRC, eDRIS and DSLS all appear to have some accountability
to the Governance groups, but what this is and what it means is far from clear. The
relationship between these organisations also needs to be clear (and clearly
communicated), especially regarding the different funding streams and the
(unintended or intended) impact of the sum of the analyses producing an outcome
that may undermine trust in the Programme. This could be managed through the
SILC Management Committee who will oversee approval. Strong direction and
support across SILC is essential.
Infrastructure must be established to make datasets available to researchers. There
are concerns around individual organisations having the resources (money, staff,
skills) to share data or the capability or interest in doing so, if there are costs. There
are questions about who eDRIS (as the co-ordinating lead for SILC) will report to.
Funding of SILC will have to be addressed in a few years’ time and there may be
issues on skills shortage regarding uptake of data scientists. Collaboration across
e.g. local government and third sector must be actively considered. The role of Farr
in progressing Health analysis is not clear, as is progress outside of health. There
must be proactive stakeholder engagement and transparency around SILC.
The funding issues are complex and are linked to other issues including Beyond
2020, with some apparent contradictions in the funding arrangements around the
three partners which may impact on access to organisations within SILC.
2.2.4 The Steering Group was to be developed to oversee the strategic direction
of the Linkage Centre and Advisory Committee, ensuring a joined-up approach with
other major developments in Scotland. During the review process, the following
views emerged:

The original Steering Group was split into the Data Management Board
(providing strategic overview) and the Data Linkage Framework Programme
Board (supporting operational delivery).

The roles and responsibilities of the various organisations and groups
involved in the Data Linkage Framework need to be clarified.
The Data Management Board is chaired by Paul Gray (DG Health and Social Care)
and the Data Linkage Framework Programme Board by Andrew Morris (Chief
Scientist). The Programme Board appears to have a limited awareness of the work
and role of the Data Management Board. With regard to the Framework Programme
Board, the views were that it was successful in focusing attention on the operational
aspects of work of partners and helping to put in place organisations and procedures
to allow work to progress. Views were expressed that work had been too centred on
the approach to sharing data in those areas already making progress and insufficient
work had been done to move work forward in other areas.
It was broadly felt that the roles and responsibilities of the various organisations and
groups involved in the Data Linkage Framework need to be clarified. The
Programme Board tended to agree papers from those leading strands of work, rather
than operating as an active discussion and decision body. This might have been
appropriate but there was a lack of clarity over its role. The context of different
organisations involved in providing data, linking data and holding data, different
funding streams, different decision bodies, different UK and Scottish legal
underpinnings and different paths to access data all mean that clarity of roles and
responsibilities is important. This is relevant to ensuring/enabling work to progress
and also in explaining to the public how data are used.
3. POINTS FOR DISCUSSION
The review has raised a number of issues which the Programme Board should
consider. The recommendations have fallen out the of the discussions which
underpinned the review and if implemented will help to ensure that the operational
aspects of the Data Linkage Framework will be delivered – particularly as the work
moves forward following cessation of the Programme Board.
The development and publication of the Guiding Principles for Data Linkage and the
establishment of the Scottish Informatics and Linkage Collaboration (SILC) are two
positive areas of work to have emerged from the Data Linkage Framework. If
Scotland is to fully realise the value of its rich data and strengths in informatics, it is
essential that the Data Linkage Framework provides a solid foundation, supporting
researchers and the public through better informed policy and outcomes delivery.
The key messages emerging from the review process are outlined below with
respect to each of the Data Linkage Framework deliverables or ‘blocks’ and should
be addressed if we are to successfully progress the data linkage agenda.
The Data Linkage Framework Programme Board is asked to consider these key
messages and discuss how these should be progressed following cessation of the
Board.
3.1
The Guiding Principles
3.1.1 There is a need to have a single set of Guiding Principles which everybody
uses, across all sectors, reducing complexity across the landscape. Following
consultation with Graeme Lawrie (who was fundamental in the development
of the SHIP Principles) to ensure that a move by SHIP-based (health)
organisations to the Guiding Principles is feasible, it is recommended that the
SHIP principles should be retired.
3.1.2 Better use of the Guiding Principles should be promoted, with public sector
organisations involved in their inception leading by example. The lack of use
by Pathfinder projects and other key projects has contributed to the perceived
lack of relevance of the specific Guiding Principles.
3.2
PAC and Ethics Committee
3.2.1 It is unclear why this work was not progressed and the lack of a PAC / ethics
committee is seen as a gap in the landscape. Support was expressed during
the review for a single PAC / Ethics Committee which would offer a
streamlined approach with a single form, helping to reduce complexity across
data linkage work.
3.2.2 The Programme Board should commission a report to clarify why the intended
approach didn’t materialise.
3.2.3 The Review Team understand that Scottish Government’s eHealth Directorate
is working on data linkage across health. It would be of value to understand
their plans surrounding this. If this work seems appropriate, the Programme
Board should support Health in driving the PAEC forward, allowing for roll-out
across all sectors. A timetable for putting in place a PAEC should be agreed.
3.3
Information Gateway and Data Linkage Service
3.3.1 The establishment of SILC has been regarded by the Board as a significant
success, where all involved have responded to funding opportunities arising
across a number of sectors within the data linkage landscape. This has
generated significant income for Scotland. However, the structure and
operational aspects of SILC continue to look very confused and need to be
firmly established for both internal and external parties. A clear description of
how the future structure of SILC might look is essential as is the need for a
clear Governance Plan and a clear Business Plan.
3.3.2 It is important to note that SILC has UK responsibilities. However, those
involved in SILC should deploy the Guiding Principles for data linkage, or
justify why they have not. The Guiding Principles have been established to
support researchers to undertake data linkage responsibly, thereby instilling
public confidence and trust. The Guiding Principles are not prescriptive and a
degree of flexibility may be appropriate as long as it is not to the detriment of
public confidence and trust.
3.3.3 Collaboration of organisations and individuals (particularly those contributing
to SILC) must be actively managed and promoted. Appropriate progress is
unlikely to be realised without a clear and active focus.
3.4
Steering Board
3.4.1 As the largest and arguably most significant area of work moving forward, the
Programme Board should prioritise the Governance of SILC and how it works
with Farr, ADRC, DSLS. This includes identifying who is best placed to lead
SILC, and is likely to include representation from NRS, NSS and OCSP.
3.4.2 The role of the Data Management Board needs to be clearer – both in regard
of its remit, but also how it will work with other parts of the Governance
system.
3.4.3 To maximise success in data linkage and demonstrate value, the Scottish
Government must stimulate activity in local authorities.
Mairead Wood – OCSP, Scottish Government
Pete Whitehouse – Health AS, Scottish Government
7th April 2014
Download