Unlocking the value of data - public engagement: literature review

Methodology

Search strategy and inclusion criteria

This was a rapid review, using an adapted systematic review approach. We conducted a systematic literature search of two electronic academic journal databases (Scopus and Web of Science) on 6^th of June 2021 using the following key word chain, tailored for each database:

(((public* OR patient* OR citizen* OR community*) W/2 (engagement OR involvement OR participat* OR perception* OR attitude* OR view* OR perspective* OR opinion* OR concern*)) AND ((public OR government* OR state OR administrat*) W/5 (data* OR record*)) AND (share* OR link* OR (private W/2 sector) OR commercial* OR (third W/2 part*) OR ((public-private OR private-public) AND partner)))

The key word chain was reviewed by and constructed with input from a University of Edinburgh academic support librarian. In addition, we citation traced the included articles and two topic relevant systematic reviews, scanning the references to identify further articles, and we conducted Google searches and grey literature database searches (Open Grey and gov.uk) to identify relevant grey literature. The searches were implemented on the 23^rd of June 2021, and the scope was limited to documents published after 2010 in the English language, but there were no geographical restrictions or restrictions on publication type.

After the exclusion of duplicates, the academic journal database search resulted in a total of 1463 articles. SE screened the titles, abstracts and, in some cases, full texts of the articles, of which 21 met the inclusion criteria and were included into the study. Additionally, 23 documents identified through citation tracing and grey literature searches were included after title, abstract and full text screening undertaken by SE, resulting in a final set of 44 included documents. The selection process is depicted in figure 1. Due to time restrictions, we did not conduct quality appraisal of the included studies. Documents that met the inclusion criteria, listed in table 1, were included. Notably, documents were included even if private sector use of publics sector data was a secondary rather than a primary topic, as long as this topic was an intentional part of the design of the public engagement activity or research that was reported. Studies where participants spontaneously expressed views on private sector use of publics sector data, even though this topic was not part of the study design, were excluded for two reasons: firstly, to manage the number of documents considering time limitations, and secondly, these studies did not report in depth findings on this topic and generally only included a mention of it.

Data extraction and analysis

Full texts for all included documents were downloaded and reviewed by SE. For each included document, main characteristics of the study of public engagement activity were extracted and entered into an Excel sheet, depicted in appendix 1. We applied a modified thematic synthesis approach to analyse the documents to enable rapid review. This, builds on but adapts the method described by Thomas and Harden (2008), which involves three stages: coding; developing descriptive themes; and, generating analytic themes. We applied a combination of (inductive) line-by-line and paragraph level coding, where SE coded lines or paragraphs of text from each included document according to their meaning and content, using the NVivo qualitative data analysis software. Documents where the primary focus or topic was private sector use of public sector data were coded in full, line-by-line. However, for documents where private sector use of public sector data was a secondar topic, coding was only undertaken at a paragraph level and for the sections of text where private sector use of public sector data was discussed.^[1] This generated a list of initial codes. SE then searched for similarities and differences between the codes, identified higher order descriptive themes, and organised these into analytical themes that relate to and address the research questions of this study.

Flow diagram of the selection process

Inclusion and exclusion criteria

Inclusion criteria:

• All articles / documents reporting public engagement / research on public sector data use by or with the private sector as a primary topic or focus of the public engagement / research
• All articles / documents reporting public engagement / research on public sector data use by or with the private sector as a secondary topic of the public engagement / research, where this was topic was explicitly designed into the engagement / study
• All articles / documents reporting primary evidence from public engagement activity or research on the topic

Exclusion criteria:

• All articles / documents not reporting public engagement or research on the topic of public sector data use by or with the private sector
• All articles / documents reporting public engagement / research on public sector data use by or with the private sector as a secondary topic of the public engagement / research, but where this topic arouse spontaneously and was not explicitly designed into the engagement / study
• All articles / documents based on secondary evidence (e.g. literature reviews, opinion or commentary articles)

Limitations

This was a rapid review, which presents some notable limitations. Firstly, the literature search, data extraction, and analysis were performed by a single researcher, SE, which means that there is a higher likelihood of the researcher’s personal and professional biases having influenced the inclusion and exclusion of documents as well as the analysis and interpretation. To mitigate this, early decisions were informed by consultation with an academic librarian and SCB. Secondly, this study was conducted in a restricted timeline, which entailed some comprises on rigour. Most notably, the fact that we did not conduct quality appraisal of the included studies means that it is possible that studies of low quality have been included into the analysis. Also, notably, with the exception of the 9 documents where private sector use of public sector data was the primary topic and line-by-line coding was applied to the whole document, only paragraph level coding of topic relevant sections of the text was undertaken for most documents. This may have influenced the construction of analytic themes. Additionally, the literature search was restricted to two academic journal databases, two grey literature databases, although these cover a very wide range of academic journals, and a limited number of Google key word searches. While citation tracing was used to identify further relevant documents, it is possible that more relevant documents would have been identified if a wider range of academic journal databases and grey literature databases were used to search for documents. Despite these limitations, this rapid review provides a transparent and robust analysis of the current state of the art with respect to public engagement on private sector use of public sector data.