Public Acceptability of Data Sharing Between the Public, Private and Third Sectors for Research Purposes

2 RESEARCH METHODOLOGY

2.1 The study was conducted using a combination of primary and secondary research methods, comprising:

• a desk-based literature review of international benefit-sharing models arising from the value of data sharing
• a desk-based literature review of different methods that have been used to empower citizens in decision making about how their data are used
• a series of deliberative events with members of the public

Desk-based literature reviews

2.2 The secondary research involved two desk-based literature reviews of published peer-review papers, written in English, and the grey literature. Relevant books/chapters were also included. The following methods were used to access appropriate literature:

• bibliographic searches were conducted through the following databases: ASSIA; IBSS; JSTOR; The Knowledge Network; BioMed Central; EMBASE and Web of Knowledge. Using several databases enabled confidence that all relevant articles were identified
• the titles and abstracts (or summaries) of the initial search results were reviewed in order to eliminate irrelevant references. This led to a refined list of relevant articles. Citation searches for each of these articles were then conducted in Google Scholar in order to identify additional relevant articles. The relevant references from each article were also reviewed
• each article was read in its entirety and analysed qualitatively to identify relevant themes which were integrated into a narrative review.
  

The desk-based literature reviews were undertaken before the deliberative research and informed the range of issues and examples put to participants.

Deliberative research

2.3 The primary research was conducted using qualitative deliberative techniques. Rather than attempting to identify pre-existing attitudes to data linkage and sharing between sectors, a concept which may not have been immediately relevant to participants nor fully understood, deliberation allowed for the sharing of information and expertise and the development of considered responses across different sessions within the deliberative events.

2.4 In accordance with the objectives outlined in the introduction, the research sought to explore the views of the general public as a whole, while also explicitly seeking the views of members of the public with protected characteristics - defined as sex, age, ethnicity, sexual orientation, religion, disability, pregnancy and maternity - which may, among other concerns, make inadvertent statistical disclosure more likely and/or more sensitive. To achieve this, the deliberative research comprised:

• four half-day events among a cross-section of the general public. At all of these events, larger equality groupings (women, different age groups, religious communities and people with a disability) were represented broadly proportionate to their representation in the population. Other minority equality groups (minority ethnic communities; Lesbian, Gay, Bisexual and Transgender (LGBT) people; and women who are pregnant or have a child under the age of 1) were over-represented to ensure their inclusion in the study (See Appendix A for a full breakdown of the profile of attendees).
• a separate, smaller scale event among LGBT people. This separate event was held to increase the likelihood of identifying any issues related to sexual identity by creating a forum in which participants might feel freer to talk openly about such issues. All participants were aware of the specific purpose of this event.

2.5 The general public events were held in Oban, Aberdeen, Glasgow and Galashiels between 29 June and 20 July 2013. The LGBT event was held in Edinburgh on 3 August 2013.

Recruitment of participants

2.6 The bulk of the recruitment was conducted face-to-face in participants' homes. However, given the relatively low penetration of LGBT people and pregnant women/women with a child under the age of 1 in the population, the face-to-face recruitment was supplemented with more targeted approaches for these groups.

2.7 LGBT participants were recruited partly through LGBT community groups and partly using on-street recruitment, with efforts concentrated in areas where there were gay bars and clubs. Women who were pregnant/had a child under the age of 1 were recruited on-street in areas surrounding GP surgeries and health clinics.

2.8 All participants were recruited between 17 June and 2 August 2013, using a questionnaire specially designed for this purpose.

2.9 To ensure that an appropriate range of people, including those with protected characteristics, were engaged in the research, quotas were set on: sex, age, working status, socio-economic position, ethnicity, disability, parental status and sexual orientation. Participants were also quota sampled based on their responses to a question about trust in public, private and third sector organisations^[4], which previous research shows can be crucial in shaping attitudes to data linkage for health research (see Davidson et al, 2012 and Aitken, 2011).

2.10 Individuals who worked in market research, media, advertising, PR, statistics/data analysis or for the Scottish Government; and those who had attended a group discussion or event in the previous 12 months were excluded from the research.

2.11 Thirty-two people were recruited to each of the general public events and 14 were recruited to the LGBT event, with the aim of ensuring that around 25 and 10 respectively attended on the day. Table 2.1, below, shows the number of people that attended each event.

Table 2.1: Number of participants attending each event

Location	Date	Number of participants
Oban	29 June	25
Aberdeen	6 July	25
Glasgow	13 July	28
Galashiels	20 July	27
Edinburgh	3 August	12

2.12 Attendees were representative of the broader pool of recruits for each event (Appendix A details the profile of attendees). All received £50 as a 'thank you' for their time and to cover any expenses incurred.

Structure of the events

2.13 Table 2.3 shows the structure of the events and summarises the purpose of each stage.

Table 2.3: Structure of the events

Session	Content/purpose
Plenary session - Welcome & orientation	To welcome participants and outline the scope of the study and key concepts that would be used throughout the day.
Plenary session - Introductory presentation on data sharing, linkage and the strategy for improving data access and analysis, plus Q&A	To provide participants with necessary contextual information, and the opportunity to ask questions.
Break-out groups	To explore unprompted attitudes towards data sharing between the public, private and third sectors, including: immediate reactions; whether attitudes varied depending on the type/role of particular organisations; the reason for sharing data; the type of data being shared.
Lunch break
Plenary session - Presentation on private sector involvement in data sharing and benefit-sharing, plus Q&A	To provide participants with a fuller understanding of the ways in which, and reasons why, the private and third sectors might be involved in data sharing; the types of organisations concerned; and the types of data that might be shared. The presentation also introduced the concept of benefit-sharing and gave examples of benefit-sharing models.
Break-out groups	To explore reaction to the second presentation, including: types and aspects of data sharing between sectors that participants felt positively/negatively about and why; measures that would mitigate any concerns; attitudes towards benefit-sharing and preferred models of benefit-sharing; and views on empowering citizens in decision making about how their data is used.
Plenary session - Closing remarks	A summing up of the key messages from the event and completion of a post-event questionnaire.

2.14 For the break-out groups, participants were divided into three groups on the basis of their age to allow for the identification of any variation in views by life stage. At each of the general public events, the intention was to divide participants into uniform age bandings (18 to 34 years, 35 to 49 years and 50 years and over). However, due to differential turnout by age, these bandings were adjusted at some of the events to ensure a roughly equal number of participants per break-out group. Table 2.2 below details the bandings used in each case.

Table 2.2: Age bandings used for breakout groups at the general public events

Location	Youngest	Middle	Oldest
Oban	18-34	35-49	50+
Aberdeen	18-24	25-49	50+
Glasgow	18-34	35-49	50+
Galashiels	18-29	30-43	44+

2.15 Participants at the LGBT event in Edinburgh were divided into two broad age groupings. The younger grouping included participants aged from 26 to 33 years and the older grouping included participants aged from 37-59 years.

Event materials

2.16 All materials used in the events were designed by the researchers at the University of Edinburgh and Ipsos MORI Scotland. The materials included two PowerPoint presentations with detailed information on topics for discussion, a discussion guide which was used by moderators to facilitate the break-out discussions, stimulus materials to promote discussion around different sectors/organisations, data types and data uses, and a post-event questionnaire (copies of all materials are provided in Appendix F to I).

Analysis

2.17 All discussions that took place at the events were recorded and then transcribed with the consent of participants. Group facilitators also compiled summary field notes at the end of each event. In addition to this, the project team held regular meetings, throughout the fieldwork process, in order to discuss findings and emerging themes.

2.18 At the end of the fieldwork, the researchers conducted a brainstorming session to identify the top-level findings and implications. The data was analysed using a thematic indexing and charting system often known as 'Framework'. This involved creating a code frame of substantive themes and sub-themes, and adding in key findings within each theme. Transcripts were systematically analysed for key points and illustrative verbatim comments. Any new sub-themes which emerged at this stage were integrated into the code frame.

2.19 This method ensured that analysis and reporting of the data was rigorous, balanced and accurate, and that key messages or concepts were brought out. It was also flexible enough to allow for links and connections across different themes or sub-themes to be made, and for moments of interpretive insight and inspiration to be recorded.

Interpretation of qualitative findings

2.20 The findings presented in this report were derived using qualitative data collection methods and analysis. Unlike large surveys, qualitative social research does not aim to produce a quantifiable or generalisable summary of population attitudes, but to develop a deeper understanding of the range of factors that shape views as well as identifying key attitudinal tendencies that are likely to be prevalent across society. Qualitative research is particularly useful when exploring complex or hard-to-understand areas, such as cross-sectoral data linkage; single or resolute responses are unlikely but the process of discussing the issues reveal the complexities of opinions and degrees of consensus. The integration of 'deliberative' approaches aids this process, since participants are given the opportunity to explore their feelings towards the topic alongside considering information provided to them. This process reveals a more nuanced and informed set of considerations and the influences underlying them, which can be useful for informing policy-making.