The barriers and facilitators to implementing screening in emergency departments: a qualitative evidence synthesis (QES) protocol exploring the experiences of healthcare workers

Protocol design

“Best Fit” framework synthesis (BFFS) produces context specific conceptual models which assist in explaining or describing the health behaviours or decision-making of patients or other groups using a pragmatic and transparent process (Dixon-Woods, 2011). In addition, this process can assist in generating programme theories relating to intervention effectiveness (Carroll et al., 2013). This method involves the creation of an a priori framework upon which thematic synthesis of primary research can be based. The framework is created utilising current models, theories and concepts that pertain to the topic under exploration; this framework is then utilised to inform the thematic synthesis of primary research studies identified during the review process (Carroll et al., 2013). This approach is often dependent on whether the framework is built upon emergent, established, refined or tentative theory and requires thoughtful consideration (Brunton et al., 2020). This interpretive methodology is deemed to be advantageous as it is reproducible, based on the current evidence base and, therefore, directly applicable to those who wish to inform practice or policy. This allows the reviewer to build upon existing models, from a potentially different but relevant population, and interrogate the testability, internal logic and fit within the evidence base (Carroll et al., 2011; Kelly et al., 2010). In addition, BFFS is well suited to improvement work as an activity rich in theories, where behavioural, social, organisational and implementation theories and frameworks might all be considered relevant (Booth & Carroll, 2015). Two separate sets of inclusion criteria, searches and study selections must be established, one to identify models, theories and frameworks and one for populating the systematic review of primary qualitative research studies (Carroll et al., 2013) (Table 1) Both searches are conducted simultaneously but independently. At the framework synthesis stage, the two “strands” then join together. This process will be reviewed and supported by an independent reviewer and any conflicts will be reviewed by the wider review team.

Identification of relevant models or theories

Initially, a framework of a priori themes needs to be created. This will be achieved by employing the BeHEMoth strategy to identify relevant models and theories (Booth et al., 2013) (Table 2). CINAHL, MEDLINE, PsycINFO and PubMed will be interrogated using a combination of free text and database thesaurus/subject terms for the behaviour of interest and health context with terms for models and theories. Reviews, regression models or integrative models will be excluded by using database filters. Results will be dual screened (title and abstract) and the full text of potentially relevant publications will be retrieved and checked for relevance. Additional relevant citations may be sourced in the reference lists of all papers satisfying the inclusion criteria. Reporting of the search strategies will be represented on PRISMA flowcharts (Moher et al., 2009). The a priori framework will be formed through secondary thematic analysis of findings from studies identified through the systematic search and screening process. The Braun & Clarke (2006) framework will be utilised. This form of inductive analysis is grounded in the data, interpretive and consistent with the process for the “best-fit” methodology. In addition, it more accurately reflects the data assisting in the identification of differences and similarities between models or theories and name them as themes (Carroll et al., 2013). Definitions, based on the elements of the original papers, will support these themes (Carroll et al., 2013). This process will then form “concepts”.

Inclusion and exclusion criteria primary research studies

To ensure the relevance and specificity of included articles, the formulation of inclusion and exclusion criteria was completed collaboratively by LB, PM and RG and is reflected in Table 3. Studies which explore the experience of HCWs (e.g. doctors, nurses, midwives, allied health professional, pharmacists) in ED settings (e.g. EDs and acute assessment units) which pertain to the barriers and facilitators of screening and the implementation of screening are suitable for inclusion in the review. Internationally, the ED screening and referral process varied with the inclusion of acute assessment units that were integrated into ED’s in some areas or part of the screening/referral process. Therefore, to ensure that all possible relevant publications were included, acute assessment units were included in both the inclusion/exclusion criteria and search strategy. The focus of this synthesis is on routine screening for physical illness and functional decline, with regards to the other types of screening that were excluded e.g. Screening for Domestic Violence, these did not appear to be routinely used and were quite case specific. To add, triage screening is a requirement in most ED’s and is consistent across most populations, our review pertains to those types of screening that have varying levels of uptake and require effective implementation and integration to ensure usage, therefore this type of screening was also excluded. As highlighted above, for the specific search for primary research studies, qualitative or mixed-methods will be eligible for inclusion. Studies must have used qualitative data collection (e.g. semi-structured interviews, observation) and analyses methods (e.g. thematic analysis, grounded theory). Peer reviewed journal articles or non-peer reviewed items including unpublished research articles and theses may be included. Grey literature sources including guidelines, reports and theses are also deemed suitable for inclusion and sourcing of these materials will be included in the search strategy. Quantitative studies and literature reviews are not deemed eligible for inclusion. Studies that pertain to the assessment/screening of adults (>18 years) will only be considered. Included studies must explicitly discuss factors that can impact on screening or the implementation of screening within the ED. If qualitative results can clearly be extracted from quantitative results, mixed-method/multiple-method studies can be included.

Search strategy primary and grey literature

In collaboration with a medical librarian, a systematic search strategy for six databases Scopus, CINAHL, Medline, Embase, Pubmed and Cochrane will be formulated. Grey literature sources will also be included and are Open Grey, Google Scholar, Lenus Irish Health Repository, Science.Gov and Embase Grey Literature sources. A multistep approach will be used to source primary literature. This will include keyword searching of electronic databases, using medical subject headings (MeSH) and specific database headings to further identify search terms, using truncation to broaden the search and ensuring all appropriate key words are used (Booth, 2016). Literature published in journals between 2009 and 2020 will be included. These dates were chosen to ensure the most up to date salient literature is sourced given the changes in healthcare and technology in recent years. In addition, our scoping search indicated that there are a number of contemporary studies related to screening processes in the ED. Therefore, we are focusing our search on sources published in the last 10 years to reflect the contemporary approach to screening in the ED. Certain terms will be truncated to ensure all spellings are captured. Specific database features will be utilised to enhance the search strategy e.g. refining the search using CINAHL, search queries and adjacency searching in Cochrane. A scoping search will be conducted to refine the search methods, identity all possible key terms, inform the formation of MeSH terms and ensure truncated terms are inclusive of all possible applicable terms/spellings are captured. For example, the terms screening and assessment are not interchangeable, however, a scoping search of the literature suggested that a search for both screening and assessment tools was warranted to ensure that all suitable publications were included in the search results. Please see Table 4 for sample Medline search string that may require refinement as the search strategy progresses. The Boolean terms of AND, OR and NOT will be utilised to expand or specify the search as required. The search string below was formed after refinement of a scoping search where initial searches utilising AND yielded large numbers of results with a limited number of relevant results. The use of OR to provide more relevant results as indicated has been the most successful search string to date but this will require further testing and adaptation. Any changes will be indicated in the QES. The methodological filter was added in consultation with the medical librarian as a sensitive and specific filter and is a search filter resource provided under library guides under their systematic review search filters (https://dal.ca.libguides.com/systematicreviews/searchfilters).

Screening search results

All references will be imported into Rayyan (Ouzzani et al., 2016) and duplicates removed. LB and a second independent reviewer (PM) will screen titles and abstracts independently and in duplicate. Where conflicts emerge LB and PM will resolve these in a consensus meeting, RG will assist in decision-making should conflicts fail to be resolved. Full texts of the articles that remain will be retrieved for screening by LB and PM against the pre-defined inclusion/exclusion criteria. Full texts will also be screened independently and in duplicate. If required, authors will be contacted for further clarification and information. Similarly, any conflicts will be discussed and RG can assist with decision-making if conflicts cannot be resolved. This independent dual screening process will make up part of the rigorous quality appraisal process where each article will be appraised using CASP and confidence in synthesized findings will be assessed using GRADE CERQual.

Quality appraisal and data extraction

The quality of all included studies will be assessed using the Critical Appraisal Skills Programme (CASP, 2018). Noyes et al. (2019) offered guidance on the critical aspects of the appraisal process from the Cochrane Qualitative and Implementation Methods Group. They advised that appraisal should include clarity with regards to research aims and questions, congruence between the research design and methodology and the research aim and question, rigor of case and/or participant identification, sampling and data collection to answer the research question and appropriate application of the methods, which includes research reflexivity and richness/conceptualisation of the depth of findings. The appraisal will be carried out independently and in duplicate by RG and LB. Data for analysis will be extracted by either verbatim quotation from study participants or findings reported by authors that are clearly supported by study data. A link to the data extraction table is included below. Google forms will be utilised and contain a section pertaining to the aim, background, setting, population and type of screening being explored. This will offer implementation and contextual details. These data will be coded against the a priori concepts. Therefore, new themes are generated from evidence that was not already captured by this framework. These new themes will be based on reviewer’s interpretation of the evidence and comparison across studies. Dual checking for consistency of extraction across studies and coding of results data for all papers against the a priori concepts derived from the relevant conceptual model will also be undertaken by LB and PM. New themes for evidence or findings that cannot be accommodated will be analysed using the Braun & Clarke (2006) thematic analysis framework. Following quality appraisal, data from low quality studies will also be extracted as, although they be of lower quality, they may still contribute data to the formation of themes where first and second order constructs from other more credible sources may enhance confidence in these thematic findings along with the evidence from these lower quality studies

Link to Data Extraction Google Form

https://docs.google.com/forms/d/1G4JxueDpcyy0B2Qv_31RYS5Ce6AmrCFoyj9sbnkSV-4/edit.

Synthesis and conceptual model

The final list of concepts will be synthesized with reference to the extracted data from the included studies, to construct a new evidence based conceptual model regarding the barriers and facilitators to screening in the emergency department (Carroll et al., 2013). Data extraction will be facilitated within Google forms, where the a priori framework will be represented in individual categories. Themes and representative quotations from the primary research studies and grey literature sources will be categorised along with the a priori foundational themes. Firstly, a simple list of defined themes, underpinned by the evidence from the included studies, and any new themes generated by the thematic analysis of any primary research that falls outside of the a priori framework will form a conceptual framework. Relationships between individual concepts will then be explored with reference to the evidence; this will then lead to a clustering of concepts and the creation of a new conceptual model describing and reflecting the behaviour of interest, representing the foundational model and theory in conjunction with the themes and concepts extracted from the primary data (Carroll et al., 2013).

Purposive sampling of included studies

A QES aims for a greater variation of concepts through analysis and synthesis versus an exhaustive sample that avoids bias (Ames et al., 2017). Therefore, larger study numbers can negatively impact on the quality of analysis and synthesis in a QES (Ames et al., 2017). In addition, data saturation may be associated with the stage when the inclusion of further evidence provides little in terms of further themes, insights, perspectives or information in a qualitative research synthesis (Suri, 2011). Therefore, if larger numbers of studies are deemed suitable for inclusion after screening, a purposive sample from eligible studies may be taken after extraction of the relevant data. This can be undertaken collaboratively by three review team members, PM, LB and RG. If required, maximum variation sampling will be undertaken with the aim of achieving the broadest possible variation within the included studies (Suri, 2011). The sampling criteria and purposive sampling frame will then be formulated based on the results of the screening and data extraction. The ultimate goal is to ensure that rich data is captured and, consequently, that review objectives and aims are met. Employing maximum variation sampling can assist research synthesists in identifying essential and variable features of a phenomenon, as experienced by diverse stakeholders among varied contexts a QES (Suri, 2011).

Confidence in QES findings

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) Confidence in Evidence from Reviews of Qualitative Research (CERQual) approach will be used to enhance transparency and confidence in reporting of QES findings. Its use is evidence-based and there are extensive resources to support reviewers in using this approach. GRADE CERQual supports the use of QES findings in decision-making, guideline and protocol development and to inform further research (Houghton et al., 2017). Four components are assessed during the GRADE CERQual process: Methodological Limitations, Coherence, Relevance and Adequacy.

The methodological limitations component will be satisfied using the CASP appraisal tool. This approach also assesses the coherence and relevance of individual review findings and in this case will indicate how major themes and sub-themes reported are grounded in the data included from primary studies and is applicable to the overall review aim (McGrath, 2019). The quantity and richness of the data supporting the review findings will also be assessed using this approach to ensure adequacy. Therefore, at this stage the contribution of lower quality papers, to the formation of themes, will be assessed in conjunction with that of more credible findings from higher quality studies. This will result in the assessment of overall confidence in review findings, taking into account the studies which informed each finding and their strengths and weaknesses. The GRADE CERQual process will be undertaken by LB and PM and reviewed by RG, and areas of concern will be pin-pointed for each component (Table 5). RG will check individual review findings for adequacy, relevance and coherence. Each phase will be discussed by the reviewers and conflict resolution will be attained through discussion and consensus.

CERQual summary of findings tables will be populated for each review finding and will include the specific review finding under scrutiny, the assessment rating (very low, low, moderate or high confidence), rationale for this assessment and the number of studies that contributed to this review finding or theme (Colvin et al., 2018; Glenton et al., 2018). The applicable QES objective and the perspective taken in the synthesis will also be included in the table to give context to the analysis of each individual finding/theme. All findings are recommended to start off with high confidence and are rated down if there is any concern re CERQual components (Lewin et al., 2015). The reviewers will remain mindful of possible interactions across components and look iteratively to make a final assessment (Lewin et al., 2015).

Reflexivity

For the purposes of this synthesis, the findings will be considered in the context of research team members’ views and experiences (Larkin et al., 2019). LB, PM, RG and SMT have backgrounds in health science and health services research. The authors have operated within the Irish and other healthcare contexts, one in allied health (RG) and three in nursing and midwifery (LB, PM, SMT). All four of these authors have experience pertaining to qualitative evidence synthesis, with one author having particular expertise pertaining to QES (PM). RG has expertise pertaining to screening in the ED. In relation to analysis, the lead researcher conducting the analysis (LB) also has experience pertaining to screening in the ED but works outside of the ED as a nurse researcher. Two authors have expertise and experience in ED clinical assessment, development of screening protocols and change management within the ED (DR, MOC). As reflected in this protocol, authors will discuss, examine and consider the significance of their beliefs, attitudes and preconceptions surrounding the research question and methodology during each stage of analysis (Larkin et al., 2019). At regular intervals throughout the process, team members will meet to discuss the review process and findings in relation to their contextual significance and changes will be made by consensus to ensure that individual bias will not impact on the process. Strict adherence to the methodological process outlined will limit bias along with this approach to reflexivity.

Dissemination of QES findings

Findings will be submitted to a peer-reviewed journal for publication. Dissemination of results among stakeholders and members of the research project team will be via oral presentation. Relevant international academic conferences in the areas of emergency and acute care and the care of older people will also be targeted. Dissemination of research findings to local and national health service management will be achieved through online platforms, video conferencing and targeted dissemination to key stakeholders via e-mail where appropriate.

Review status

Database searching for primary research studies is ongoing in conjunction with searching to formulate the a priori framework.

Underlying data

No underlying data are associated with this article.

Reporting guidelines

DANS-EASY: PRISMA-P checklist for “The barriers and facilitators to screening in emergency departments: a qualitative evidence synthesis (QES) protocol”. https://doi.org/10.17026/dans-zyc-9ega (Barry, 2020)

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).