The methods were informed by the JBI approach [7, 31], which builds on Arksey and O’Malley [32] and Levac et al. [33], and by the hermeneutic approach of Boell and Cecez-Kecmanovic [34, 35]. We noted that scoping review methods are increasingly being clarified and refined [7, 36], with recent attention to the involvement of patients in reviews [37].

The team consisted of a patient [PZ] with extensive life experience including implementing programs; academic researchers from several disciplines [IG, MM, GM, DS, EW, LZ], with expertise in hermeneutics and/or implementation; a healthcare leader responsible for health programs and services [CU]; a health research librarian [TF]; and research trainees [SJ, EK]. Unlike many scoping reviews, where research trainees implement the review protocol, all members of the research team, including the patient, were fully involved in all aspects of the review.

An a priori scoping review protocol was registered with the Centre for Open Science (osf.io/eac37). Deviations from the protocol are detailed in the sections below.

Due to a commitment to the hermeneutic approach of remaining open to new possibilities [8, 9, 12, 34], coupled with the unexplored area of what could count as a hermeneutic approach, we used flexible and iterative processes [7, 38] in (1) formulating the research question; (2) screening and identifying relevant literature; (3) selecting relevant studies; (4) charting data into tables; and (5) synthesizing, summarizing, and reporting results.

Formulating the research question

The formulation of the research question provided a good example of how the research team worked together throughout the review. When the initial search results were inconclusive, the whole team, including the patient, trainees, librarian, and academic researchers, all with differing experiences and disciplines, met to determine the next steps. The discussion was inclusive and iterative, and revolved around each team member articulating their perspectives. The team took time to develop a shared common understanding. That is, the team had a hermeneutic conversation, characterized by listening, dialogue, and coming to new shared horizons. The patient, in this conversation and throughout, asked for clarification and raised the question, what does this mean? He helped others in the group to be confident in expressing what they knew and did not know. As a service user without insider knowledge of the healthcare system, the patient kept bringing often highly conceptual discussions back to concrete realities, asking about implications for those who experience implementation at the clinical interface. There were many such extensive discussions over the course of the review.

The question was formulated around PCC: Population or Participants, Concept and Context [7, 31]. The team kept the question broad to remain open to potential links between hermeneutics and implementation. In a preliminary search of research databases and the grey literature, the librarian found no research literature that explicitly linked the two concepts. Although hermeneutics was mentioned in the health literature, the term, hermeneutics, was not found in the implementation literature.

The original question identified in the protocol was, “How is philosophical hermeneutics currently taken up in implementation science or the context of implementing?” We changed the term to “hermeneutics” because the term “philosophical hermeneutics” is often associated with the work of a specific philosopher [8].

We therefore identified the population or participants as any participants; the concepts as hermeneutics and implementing; and the context as health programs, services, or practices. The research question became,

What constitutes a hermeneutic approach to the process of implementing health programs, services or practices?

Screening and identifying relevant literatureInclusion/exclusion criteria

In keeping with a hermeneutic approach [8, 9, 34, 35], precise a priori inclusion and exclusion criteria were not identified. We proceeded with concepts that were sensitizing rather than operational, to avoid pre-determining or limiting the inclusion of articles that could deepen or extend our understanding of the nexus of hermeneutics and implementation.

Through multiple conversations among team members which drew on disciplinary knowledge, other literature, and experience, we sharpened inclusion criteria in the process of selecting articles. We sought to balance feasibility, breadth, and comprehensiveness. In reviewing articles, we always asked, what meaning does this particular article bring in relation to our understanding of the intersection of implementation and hermeneutics?

Detailed documentation and frequent communication among the whole team assisted in maintaining consistent interpretation of the criteria.

Population or participants

All populations or participants concerned with health were included, with anticipation that primary populations or participants would be patients, healthcare providers, communities, and decision makers.

Concept

Each publication had to include both hermeneutics and a component of implementing or implementation.

Hermeneutics

For an article to be considered hermeneutic it needed an explicit theoretical statement of being based in hermeneutics. It also had to explicitly address context, temporality, dialogue, and personal understanding. That is, there was an indication of where implementation was happening, who was involved and in what ways they were involved; there was a mention of change over time; there was an indication of conversation or dialogue among individuals intended to foster understanding; and it was evident that there was a change in participants’ understanding. The articles needed to state that interpretation was used and reflect interpretation in how the article was written. It needed to be evident that links had been made beyond the immediate situation to new understandings of theory, or a new articulation of experience or to gaining a new point of view. Articles needed to state interpretation and links to the philosophical hermeneutic underpinnings, as well as demonstrate these aspects. The way in which hermeneutics was expressed in the article was used as an inclusion criteria. If an article was not “sufficiently hermeneutic” it was dropped from consideration [39].

Implementing

As the goal was to study the actual process of implementing, that is adaption, adoption, improvement, decision-making, or communication that affects practices and/or behaviors [40, 41], articles needed to include such features as what the authors/researchers did, how they changed practice, etc. Articles needed to report more than just a change in attitude, but also could include decision-making, a change in behavior, or descriptions of applying action, plan, or recommendations for change. The studies did not need to be primary studies of implementing.

Context

We did not focus on a particular concept of health, but noted that implementation could occur within health programs, health services, or practices, in any setting, and with any population or participants. We included all health settings, including community, acute care, long-term care, primary care, as well as some settings less frequently associated with health, such as education of health professionals. No exclusions were based on geographical or locational factors, cultural factors, specific race, gender, or sex-based interests.

Article types, study designs, and language

Primary studies and reviews were eligible. Included were academic journal papers and brief reports from the health sector. Conference abstracts, editorials, opinion pieces, commentaries, and philosophical or theoretical papers were excluded. Grey literature was also excluded following the preliminary search that resulted in no relevant documents. Included was literature published in five languages: English, Icelandic, Norwegian, Swedish, and Danish, because all research team members were fluent in English and one was fluent in the other four languages.

Identifying the literature

As a result of the iterative process, the librarian performed a simple search of the literature using subject headings, if available, for hermeneutics and a search for the term hermeneutic in the title (ti) and/or abstract (ab). We then situated hermeneutics within a search of simple implementation language in the title and/or abstract: implement* or adapt* or adopt* or improve*. Examples of the search are provided in Table 1.

Table 1 Literature search strategy

We conducted the search across eight health-related databases: MEDLINE OVID, Embase OVID, EBM Reviews, PubMed, CINAHL EBSCO, PsycInfo EBSCO, Web of Science, and JBI EBP Database. The search included published literature from conception of databases to June 11, 2021. The original search was completed on February 27, 2019 and updated searches following the same process with the same databases were completed on April 17, 2020 and June 11, 2021. We did not pursue a formal peer review of the search [42] due to the simple search strategy, the team’s expertise in the subject area, and the team’s iterative and thoughtful discussion regarding the hermeneutic and implementation language included in the search.

Selecting relevant publications

In order to include a broad range of literature, titles and abstracts were reviewed at the same time. Refining or sharpening the inclusion decisions happened through the full involvement of the whole team in all stages of review and through the extensive discussion of what was meant by hermeneutics and implementing. Once articles were identified, titles and abstracts were uploaded into Evidence Partner’s DistillerSR systematic review software for duplicate removal, article screening, selection, and data extraction. Members of the research team: a patient, academic researchers, the healthcare leader, and a research trainee, were split into five pairs to first screen titles and abstracts of articles, and second, to screen the full texts of articles.

Pilot screening

Pairs of reviewers completed pilot screening with 1% of the articles identified in the search. Reviewers independently considered the titles and abstracts for the concepts of hermeneutics and implementing within a health context. Each pair came to consensus. Reviewers provided a text response for their primary reason for inclusion, exclusion, or uncertainty (“cannot tell”). Following whole team discussion, the text responses were grouped into reasons for inclusion or exclusion and categorized. In addition to type of article, lack of abstract, and language, decisions to exclude occurred in this order: (1) the title or abstract did not concern a health context; (2) hermeneutics was not mentioned; (3) a process of implementing was not identified. As some articles about decision-making had implications for implementation, decision-making was created as a criterion for inclusion.

Title and abstract screening

The same pairs individually screened titles and abstracts to determine whether articles should proceed to full text review using the same process and priorities as in the pilot screening. Discussions among the whole team clarified questions and confirmed reasons to include articles. Reasons for inclusion were refined. Within the concept of implementing, articles were included that were either about implementing a practice, service, or program; could inform implementation; or were about making decisions relevant to implementing. Rationales for excluding the articles were that they were “not implementation” or “not sufficiently implementing”. Reviewers could select “other” as a reason to include or exclude and provide free text comments. An individual reviewer who was uncertain about inclusion could mark the article as “cannot tell”. Any discrepancies, where one reviewer indicated “include,” resulted in full text screening.

Full-text screening

Three steps were taken for full text screening. At the first step, each pair individually and independently reviewed each article and decided to include or exclude an article. The primary reasons for inclusion and exclusion at this stage remained the same as for title and abstract screening. Following this review, two other reviewers independently categorized comments in the free text “other” category into existing categories. The whole team reviewed and reached full consensus, agreeing to this categorization, with no additional categories needed.

During the regularly scheduled full research team discussions, as each person contributed their perspective and insights, it became apparent that reviewers picked up different features in the articles and that greater consistency of review was needed as well as equivalent attention paid to hermeneutics and implementing. During those discussions, each person contributed what they found to be indicative of hermeneutics and implementing. Through this open dialogue, with the goal of hearing each other and understanding differences and similarities in perspectives and meaning, the team gained increasing clarity and came to consensus about the features of hermeneutics that needed to be present for the articles to be included for further examination [39].

Following the first full-text review step, experts in hermeneutics and implementation reviewed the articles in a second full-text review step. MM reviewed all included full text articles to confirm the presence of hermeneutics and LZ reviewed the full text articles that MM had reviewed at an earlier stage. If hermeneutics was present, IG reviewed to confirm the presence of implementing.

In the third full text screening step, the whole team made final decisions about inclusion of the remaining articles during a two-day meeting. Research team members read these articles in depth and discussed them in detail for the presence of both hermeneutics and implementing. Through dialogue we reached consensus and gained increased understanding and insights [35] about the nexus between hermeneutics and implementing across included studies.

Consistent with scoping review methods, we did not assess methodological quality [31] or risk of bias [7].

These repeated in-depth discussions led the research team to further articulate what needed to be present to be named as a hermeneutic approach. Articles were not sufficient for inclusion when there was the following: results were presented as descriptions of experiences and/or through simply naming and illustrating themes [43]; the presentation, stated philosophy, or philosophical references were not linked to the approach or findings; hermeneutics was talked about but not reflected in the article. The team also clarified what needed to be in place about implementing. The processes of implementing were present when authors depicted how they actually went through the steps or processes of implementing an intervention or decision-making. We did not consider implementing to be present when the article only depicted the following: participants’ experience of a service, an evaluation of a service, the result of an implementation, or the need for future implementation.

Charting the data

Three persons charted data from the final remaining articles. The study characteristics as outlined in the protocol were charted by a research trainee (SJ) and confirmed by the research team (Table 2). Two research team members with expertise in hermeneutics (MM, LZ) extracted details on hermeneutic approach and implementation. The categories evolved during discussion of the results, and the whole team refined the charted data accordingly in Tables 3 and 4.

Table 2 Characteristics of included studiesTable 3 Hermeneutic features of included studiesTable 4 Implementation components of included studiesSynthesizing, summarizing, and reporting results

Through the extended discussions held in person and subsequently in three multi-hour teleconferences, team members engaged in a hermeneutic interpretation of the results, achieved a common horizon, and answered the research question. The discussions were reflexive and free-flowing, where we shared and questioned our pre-judgements [8, 9], assumptions, and interpretations. Team members used the detailed notes that were taken of all team meetings and teleconferences to reflect on their own and others’ insights and revise and/or deepen their own and our collective understanding. Although stated in the registered protocol, we did not use NVivo to manage data.