A Scotland-wide Data Linkage Framework for Statistics and Research: Consultation Analysis

Consultation Question 5: National Data Linkage Centre

The consultation paper set out a range of functions for a National Data Linkage Centre (NDLC) to provide and asked: Are the functions that will be led by the National Data Linkage Centre set out in section 3d the right ones?

Around two thirds of those that provided a response to the yes/no question felt that the functions of the National Data Linkage Centre as described in the consultation paper were the right ones.

Are the functions that will be led by the National Data Linkage Centre set out in section 3d the right ones?
Type of respondent	Yes, they are the right functions	No, they are not	No answer
Data custodian	7	1	3
Data user	20	8	2
Data subject	2	2	1
Multiple categories selected	2	2	8
No selection	0	0	3
Total count	31	13	17

General comments

The weight of responses regarding the NDLC were broadly supportive. One such response noted:

'The benefits of having a centre to capture the collective knowledge, skills and memory and to deliver efficiency savings but at the same time ensuring research excellence, is an exciting prospect.'(HEADLINES, University of Aberdeen)

However, a number of respondents suggested that whilst broadly reasonable the functions were expressed too generally. It was felt that a comprehensive and detailed list would allow an idea of how the centre would operate in practice and allow for a more informed response.

In the analysis of the comments on the NDLC, the type of respondent (data user, subject or custodian) was examined. The different types of respondents were found to hold similar views on the NDLC, the main points of which are outlined in the sections below.

Proposed additional functions for the NDLC

A number of functions additional to those in the consultation paper were suggested for the NDLC:

• promoting a better understanding of the constituent data sets
• quality assuring data and the linkage process and establishing and monitoring standards
• performing a screening role concerning the validity of requests for data linkage
• performing an oversight role concerning the uses of the linked data
• coordinating the maintenance of research-enabled national data ready for linkage – as separate from the technical function of linkage
• providing expert commentary on the linked datasets
• facilitating and encouraging the release of data for linkage purposes
• providing information and communication technology support
• acquiring and maintaining data linkage research equipment
• leading the standardisation of components of datasets that facilitate linkage
• providing data linkage and analytical workforce training
• providing a 'clearing house' role as a single portal for researchers to approach (this would be a positive differentiator from other jurisdictions when competing for research investment).

Staffing and resource issues
A number of respondents emphasised that members of NDLC staff would require specific technical skills with one respondent noting that the appropriate level of knowledge would have to be acquired rapidly. It was suggested that the NDLC would require sufficient resources to draw on the best available Scottish talent, some of which may be located in private data mining companies.

As well as ensuring adequate staffing, more than one respondent suggested that designing and building a sufficiently high performance infrastructure with comprehensive security features would be a challenge.

One respondent suggested that thought be given to how processes and procedures can be modernised, given that the model on which the NDLC is based is over 15 years old another called for 'innovative and forward thinking approaches'.

Additional bureaucracy
As with the Privacy Advisory Service a number of respondents were concerned that the NDLC might present further bureaucratic hurdles and lead to a lengthy process.

In particular, a number of respondents made clear that whilst they could see the merit of the NDLC, they did not want it to impact adversely on pre-existing, successful data linkage:

'this centre with these functions may help to facilitate future linkage. My concern is that where linkage is currently happening appropriately, securely and successfully...any requirement to go through this centre may result in long delays and may actually hinder future linkage.' (Anonymous)

Relationship with PAS and other bodies
Related to this, there were a significant number of responses discussing the relationship with the NDLC and other bodies

'It is not clear how this Centre would interact with existing bodies that currently deal with such data' (Medicines Monitoring Unit, University of Dundee)

In this regard, respondents urged that the NDLC to build on existing structures and best practice and several respondents commented on the importance of coordination, harmonisation and integration between all the bodies involved.

'We strongly urge that the National Data Linkage Centre is taken forward as an inclusive, collaborative network involving the regional safe havens and other experienced data management teams to build on the national strengths.'(HEADLINES, University of Aberdeen)

Respondents suggested variously that relevant stakeholder groups, the research community, and the Scottish Longitudinal Studies Centre should be involved with the NDLC.

As was the case for the PAS, the need for a NDLC at all was questioned.

'we question whether enough thought has been given to the need to create a new body when there may be existing organisations which could provide this service.'(Orkney Islands Council)

The relationship between the NDLC and PAS was again raised. Some responses suggested combining the two bodies. However, the counter-argument was also made:

'An important point to recognise is that privacy advisory service must be wholly independent and publicly recognised as such, from the data linkage centre'(Glasgow Housing Association)

Centralisation
A number of concerns were expressed relating to the centralisation implied by the formation of the NDLC and it was highlighted that the stated functions vested considerable power in one entity and this presented moral hazard and created an entity that is 'too important to fail.'

Equally it was argued that the NDLC must be implemented in such a way that it does not become a weak point in privacy protection. As one respondent noted:

'If one organisation ends up with access to a large collection of data sets, much of the protection that results from ensuring that data are not centralised would be lost. The design and implementation of such a centre requires great care.' (No2ID)

Additionally, there was a caution against increased centralisation on the grounds of 'efficiency':

'It would be much more efficient if the police and legal system were one body but there are good reasons to keep them separate and these reasons apply equally to data linkage systems.' (Violet Warwick)

Questions
Respondents also raised specific questions relating to the functions of the NDLC

• Would the NDLC hold linked datasets, if so for how long? Would it be an archive in any sense?
• What is meant by "satellite linkage units"?
• Would ISD be the source of all health data? Would Health Boards be consulted about access to datasets collected in their areas for their residents?
• What would the role of the NDLC (and PAS) be would in relation to data sharing by private organisations? Would the NDLC restrict the activities of commercial organisations regarding data linking within Scotland?