Conceptual framework
According to Nilsen (2015), implementation is a multidimensional phenomenon, where implementation barriers and facilitators can be influenced by an interplay of multi-level factors ranging from individual to organisational factors(18). The Consolidated Framework for Implementation Research (CFIR) is a determinant framework in implementation science that was developed to guide the systematic assessment of multi-level contexts to identify determinants (i.e. factors) that can influence intervention implementation(19). The constructs in CFIR are derived from the synthesis of theories on dissemination, innovation, organisational change, implementation, knowledge translation and research uptake, and have received consensus from experts in this field(19). It comprises of 39 constructs grouped within five major domains - intervention, outer setting, inner setting, characteristics of individuals and implementation process - all of which interact to influence intervention implementation and implementation effectiveness. This framework has been previously applied to wide ranging fields of study, including eHealth technology(20), and provides a comprehensive approach to the investigation on barriers and facilitators that can affect implementation. Therefore, results of the extracted articles will be synthesized and integrated using the CFIR. Use of this framework will allow barriers and facilitators affecting implementation of interventions to be identified and presented in a structured manner. It will also enable findings from this review to be more readily comparable to other implementation studies and allow gaps in research to be identified.
PROTOCOL DEVELOPMENT
This study follows the Arksey and O’Malley’s (2005) framework for scoping reviews along with methodological enhancement by Levac et al (2010). The five stages within this framework are 1) identifying the research question, 2) identifying relevant studies, 3) selecting studies, 4) charting the data, and 5) collating, summarising and reporting the results. The study protocol was registered with the Open Science Framework on 16 May 2020 (https://osf.io/2x3y9/), under the Creative Commons Attribution Non Commercial (CC BY-NC-4.0) license. This means that others may distribute, remix, adapt and build on this work on a non-commercial basis, and license their derivative work using different terms, on the basis that the original basis is properly cited and the use is non-commercial (http://creativecommons.org/licenses/by-nc/4.0/). The structure and content of this protocol follows the Preferred Reporting Items for Systematic reviews and Meta-analyses extension for Scoping Reviews (PRISMA-ScR) checklist (21) (Additional File 1).
Stage 1: Identifying the research question
The main research question is defined as: “What are the barriers and facilitators that affect the implementation of social robots for older adults, including people with dementia?”.
Stage 2: Identifying relevant studies
Search strategy
Relevant published studies or literature will be identified by searching the following electronic databases: MEDLINE via Ovid, EMBASE, PsycINFO via Ovid, Scopus, Web of Science, Compendex and PubMed. Reference lists of selected studies will be hand searched to ensure that any additional literature that may be of relevance will be identified. To ensure that all relevant information is captured, grey literature sources (E.g.: Web of Science Conference Proceedings, Google Scholar) will also be searched to identify studies, reports and conference abstracts of relevance. Several terminologies have been used across the literature to describe the term or concept of implementation(22). Therefore, to improve the specificity of searches (23), the taxonomy of implementation outcomes that was developed by Procter and colleagues(24) will be used to guide the systematic search for articles relating to implementation. This taxonomy consists of eight constructs, which includes: acceptability, adoption, appropriateness, feasibility, fidelity, implementation cost, penetration and sustainability. Some of these concepts have been used in studies to describe implementation of SAHRs(15). Therefore, use of these concepts as part of the search strategy will likely yield relevant results and ensure thoroughness of the searches. The concepts in the taxonomy are solely to guide the literature search strategy, so that the search and selection of articles may be consistent, broad and unbiased. To improve the sensitivity of searches(23), the search strategy will not include terms such as ‘facilitators and barriers’, ‘factors’ or ‘determinants, because the authors anticipated that such terms are often not mentioned in the title and/or abstract. Instead, barriers and facilitators to implementation may only be discussed in the body of the text. Hence, this information will only be assessed through reading the full texts at a later phase of screening to ensure that no potentially relevant articles are omitted. These search strategies were developed in consultation with a research librarian to optimise the specificity and sensitivity of the searches, who will also provide support throughout the search process. No forward citation tracing will be conducted; however, the reference list of relevant reviews and included studies will be hand searched to identify other potentially relevant studies. A sample search strategy that has been developed in consultation with a research librarian can be found in Additional File 2.
Stage 3: Study selection
The titles and abstracts resulting from the search strategy focused on the barriers and/or facilitators that affect the implementation of social robots will be included for review. Articles will be imported into EndNote and be deduplicated. To determine eligibility, a two-phase screening process will be undertaken by two independent reviewers. Firstly, titles and abstracts of identified articles will be screened for eligibility by each reviewer as per the following inclusion criteria: (1) use of social robot(s) as an intervention, (2) involve older adults and/or people with dementia, (3) published in English language, and (4) provide information regarding factors affecting the implementation of social robots, based on any of the constructs listed in the taxonomy of implementation outcomes. This approach of evidence selection reduces the potential for evidence selection bias(25). All types of empirical research studies encompassing any types of methods and study designs will be included. No search limits will be applied to the year of publication, and all publications will be searched from inception Correspondingly, (1) non-empirical studies such as review articles commentaries or expert opinions, (2) does not involve older adults and/or people with dementia, (4) published in non-English language and (4) does not contain any terms relating to implementation will be excluded. Next, the full texts of relevant papers will be screened. At the end of each screening process, the reviewers will compare their decisions. Any non-consensus or ambiguity regarding eligibility for inclusion will be discussed and resolved among both reviewers and with a third independent reviewer if necessary.
Stage 4: Charting the data
A standardised charting sheet will be developed using Microsoft Excel to allow reviewers to chart the data to confirm the studies’ relevance and to extract their characteristics. Charting refers to the technique of sifting, mapping out and sorting of materials based on their key characteristics(26). Study characteristics to be extracted will include information such as authors’ name, year of publication, study design, country, participants’ demographics, study setting, construct or term used to describe implementation, key relevant results relating to the aim of the research question (i.e. barriers and facilitators affecting implementation). This charting sheet will be reviewed and pre-tested by both reviewers to ensure consistency in data extraction and that all the necessary information is captured from each study. Each reviewer will then independently extract data from the included studies, and comparisons will be made afterwards. Any incongruence will be discussed and resolved among both reviewers and with a third independent reviewer if necessary. Finally, the data will be combined into a single Microsoft Excel spreadsheet.
Stage 5: Collating, summarising and reporting the result
In this stage, findings will be collated, summarized and reported. First, terms that were used to describe implementation will be mapped onto Proctor’s taxonomy of implementation outcomes. Terms that are not described in the taxonomy will be identified as independent terms. The frequency in which these terms were used will be presented. Next, the types of social robots used will be categorised into three operational groups based on their functions: (1) socially assistive robots, (2) pet robots, and (3) telepresence robots (Additional File 3). Next, to synthesise the extracted data on barriers and facilitators, directed content analysis (27) will be applied to deductively using the CFIR (28, 29). Based on the extracted data, data synthesis will be conducted separately for older adults and people with dementia. Barriers and facilitators will be mapped onto one of the 39 constructs in the CFIR, based on a pre-established codebook of definitions that has been adapted to fit this study (Additional File 4). This synthesis will be verified by a second reviewer. Any disagreements will be discussed and resolved among both reviewers and with a third reviewer, as necessary. All data will be organised thematically according to the five domains in the CFIR to map and present implementation barriers and facilitators in a structured manner. This will show areas that have been under researched and may require further investigation. The findings of the study will then be presented narratively (30, 31), using the PRISMA-ScR checklist (21). Gaps in literature will be discussed, and areas for further research will be identified. A PRISMA flow chart will be used to present the methodological process in detail.