Learning healthcare systems and rehabilitation: a scoping review protocol

Review question

How have “learning health systems” and “learning organisations” been conceptualised and operationalised in the field of rehabilitation?

The objectives of this review are:

•     To gather and note the prevalence of literature which applies LHS and LO concepts to the rehabilitation context.

•     From this literature, to collate different definitions of learning healthcare organisations and systems, and note core features of the concepts

•     To record how LO or LHS frameworks were applied in rehabilitation settings and the changes made within the organisation.

Screening protocol

The review will be guided by the methodological framework for scoping reviews outlined by Arksey and O'Malley (2005) and Levac et al. (2010), and the evidence synthesis guidelines from the Joanna Briggs Institute (JBI) Manual for Evidence Synthesis (Peters et al., 2020).

The review team

All co-authors actively participated in the development of this protocol. The full review team consists of three academic researchers, a nurse, a speech and language therapist, a physiotherapist, a rehabilitation hospital programme manager and a university librarian. Three members of the review team (CH, ZT and LC) will undertake the screening of the retrieved literature and the data extraction. Two reviewers will dual screen all abstracts and full texts (CH and ZT), and a third reviewer (LC) will resolve any conflict and decide on final inclusion. The entire process will be supervised by a senior academic researcher who is extensively experienced in scoping review methods (AC).

Electronic databases

The research team will undertake a comprehensive search of the literature within the following databases:

OVID MEDLINE (1946 – 9/2022)

EMBASE (1947 – 9/2022)

CINAHL Nursing and Allied Health (CINAHL Plus) (1961 – 9/2022)

APA PsycINFO (1967 – 9/2022)

COCHRANE Database of Systematic Reviews (1996 – 9/2022)

Key search terms and search strings

The key search concepts for this study are ‘learning health system’ AND ‘rehabilitation’. Alternative and related terms for each of the concepts will also be included, and exclusion terms will be added to eliminate results related to substance abuse and acute care rehabilitation. Table 1 contains the keywords and exclusion terms for the search strings. The search query will be adapted for each database using specific Boolean operators, truncation markers, and MeSH-, Emtree-, subject- and index terms and headings where necessary. The research team collaborated with an expert university librarian (D.S.) in designing and refining the search strategy. Table 2 contains the search strings for each database.

As recommended by the JBI Manual for Evidence Synthesis (Peters et al., 2020), a three-step process for applying a search strategy will be implemented. Step 1 has already been carried out and involved an initial limited search on multiple databases for relevant articles. This has been followed by analysis of the keywords and phrases contained in the titles and abstracts of the retrieved papers, and of the index terms used to describe the articles. Step 2 will involve the search across all included databases using the identified key words outlined in Table 1. Step 3 will involve the search of the reference lists of selected sources to identify any additional relevant studies, as well as hand search of two rehabilitation datahubs involved in application of LHSs, the International Spinal Cord Injury Community Survey (InSCI) (http://www.swisci.ch) and the LeaRRn | Learning Health Systems Rehabilitation Research Network (LeaRRn) (sites.brown.edu/learrn/).

Inclusion/exclusion criteria

In line with methodological framework by Arksey and O'Malley (2005), the final inclusion and exclusion criteria will be devised post hoc, based on increasing familiarity with the literature. Following the JBI guidelines, however, the scoping review question was guided by the PCC mnemonic (population, concept, and context) and it will inform inclusion and exclusion criteria and consequently the literature search strategy.

•     Population: health and social care professionals – nursing, medical, allied healthcare professionals/health and social care professionals, health and social care management.

•     Concept: studies whose primary focus is "learning" healthcare system or organisation.

•     Context: studies conducted in rehabilitation healthcare settings.

•     Types of evidence sources: peer-reviewed qualitative, quantitative or mixed-methods empirical studies, reviews, and grey literature in the English language. This includes conference proceedings and opinion pieces. Studies will be included if they define a “learning organisation” or “learning health system” and focus on or specifies relevance to a rehabilitation context. Alternatively, papers will be included if they describe an operating LHS (research focused on LHS data analysed) and/or translation of research evidence generated from LHS data into healthcare improvement, within a rehabilitation context.

Papers will be excluded if they contain no substantive discussion of the learning healthcare system or learning organisation concept or are conducted outside rehabilitation settings. Non-English language studies will be excluded due to time and resources required for translation. Considering the limited number of relevant papers from rehabilitation settings, the authors decided against limiting the search by date. Animal research and poster abstracts will also be excluded. In line with the guidelines from the Joanna Briggs Institute (JBI) Manual for Evidence Synthesis and Levac (Levac et al., 2010; Peters et al., 2020), at the beginning of the process, the team will meet to discuss decisions surrounding study inclusion and exclusion. Two reviewers will dual screen all abstracts and a third reviewer will resolve any conflict and decide on final inclusion.

Study selection

The found articles will be imported into the bibliographic reference management software Endnote (https://endnote.com/), and any duplicates removed. The systematic review software tool, Covidence (www.covidence.org), will be used for screening of the retrieved literature. A suitable free alternative to Covidence software is Rayyan (https://www.rayyan.ai/). For the pilot testing, a random sample of 25 titles/abstracts will be selected. These will be screened by the entire team using the eligibility criteria and definitions/elaboration document. The reviewers will meet to discuss discrepancies and make modifications to the eligibility criteria and definitions/elaboration document. Team will only start screening when 75% (or greater) agreement is achieved. The reviewers will meet at the beginning, midpoint, and final stages of the abstract review process to discuss challenges and uncertainties related to study selection and to the search strategy if needed. Any disagreements will be resolved by involvement of a third reviewer. The full text article review will be undertaken by the same reviewers using the same method, with the two reviewers reviewing the full texts independently and the third reviewer resolving any conflicts and deciding on final inclusion. The process of study selection, as well as the number of identified, screened, assessed, and included articles will be reported using a PRISMA-ScR flow diagram (Tricco et al., 2018).

Data extraction

A draft charting table – for extracting the data addressing the research question – based on JBI Manual for Evidence Synthesis (Peters et al., 2020) was drafted at the protocol stage (see Table 3). Descriptive data about the study documents will be extracted. This will include:

Data will be extracted under the following headings in order to address the research questions:

The charting table will be continually updated in an iterative process to capture other relevant data that the authors encounter during the process. In line with recommendation from Arksey & O’Malley and JBI Manual for Evidence Synthesis (Arksey & O'Malley, 2005; Peters et al., 2020), the review team will pilot the data extraction table on a sample of the included studies (10% of the complete list of retrieved studies) to ensure all relevant results are extracted.

Collating, summarising and reporting the results

A descriptive approach will be employed to address research questions. Frequency counts will be appropriate for example in demonstrating how many studies focused on operationalising a LO or LHS framework in an organisation, or in ascertaining which journal discipline was most prevalent in the collated studies. Basic coding can be done on definitions using NVivo (https://www.qsrinternational.com/nvivo-qualitative-data-analysis-software/) to identify different ways that the concepts have been defined, and key features of the concepts can be listed. The results pertaining to how the LHS or LO concept was applied to the rehabilitation setting will be descriptively summarised on a paper by paper basis.

The results of a scoping review will be presented as a map of the data extracted from the included papers (as charts) and in an aggregate tabular form (as tables) for better visualisation of the results, and in a descriptive format that aligns with the objectives and scope of the review. In line with JBI manual analysis will be descriptive in nature, as more in-depth analyses such as thematic analysis or synthesis are better suited to qualitative evidence syntheses (Peters et al., 2020). This approach will provide information on the body of research on the concept of “learning” organisations and systems in the context of rehabilitation care.