This protocol describes a qualitative systematic review that will synthesize evidence using best-fit framework synthesis (BFFS), following the approach described by Carroll et al [21,22]. The review aims to identify, appraise, and synthesize evidence on the components and indicators of the ethics of care in nursing within an a priori framework comprising Gilligan’s [8] ethics of care as an overarching lens and Tronto’s [10] 5 elements of care as an operational structure. BFFS allows the existing evidence to be tested against the a priori framework and, where the evidence does not fit, allows that framework to be extended through inductive analysis. The review will be reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 guidelines, and the search will be reported in accordance with PRISMA-S (Preferred Reporting Items for Systematic Reviews and Meta-Analyses literature search extension) [23]; no meta-analysis will be conducted.

This review protocol has been registered with PROSPERO (CRD420251032024). The registration record was created on April 13, 2025, and is publicly available on the PROSPERO website. The protocol was developed in accordance with the PRISMA guidelines [24]. Ethics approval was waived, as no data will be collected directly from human participants. The findings will be disseminated through publication in a peer-reviewed academic journal.

Several amendments were made to the registered protocol before completion of the review. First, the review approach was changed from a systematic review with meta-analysis to a qualitative systematic review using BFFS because the data of interest are conceptual and behavioral rather than quantitative effect estimates. Second, the original date restriction (2020-2024) was removed so that earlier and foundational work relevant to the ethics of care would not be excluded. Third, the database set was revised according to institutional access and methodological suitability: Scopus was removed because institutional access was unavailable, EBSCO Discovery Service was removed because it is a discovery layer rather than a discrete database, and PubMed (MEDLINE) was added. Backward citation searching of the reference lists of included studies was also added. Fourth, the language restriction was limited to English-language publications. Fifth, owing to resource constraints, screening and quality appraisal were planned primarily with a single primary reviewer, with calibration, maintenance of an audit trail, and consultation with a second reviewer in cases of uncertainty used as safeguards.

An initial search had been performed under the originally registered protocol (a systematic review with meta-analysis, with the registered database set and a 2020-2024 date restriction). Following the first round of peer review, the RQs, review approach, search strategy, database set, and eligibility criteria were substantially revised. Because these revisions materially changed the search, the initial search was not carried forward; the search was rerun in full under the amended approach in May 2026, and all records reported in this manuscript were identified through this new search. The records retrieved under the original protocol were therefore superseded and are not included in the counts reported in this manuscript. These amendments are reported here for transparency; the PROSPERO registration will be updated to reflect them following the current round of revision.

The review is structured according to the population, concept, and context framework, with the population comprising registered nurses and nursing students, the concept focusing on the ethics of care as conceptualized by Gilligan [8] and Tronto [10], and the context encompassing nursing practice and nursing education.

For clarity, the term “components” refers to the principal conceptual dimensions of the ethics of care, namely, Tronto’s [10] 5 elements, whereas “indicators” refers to the observable behaviors or expressions that reflect each component.

A systematic search was conducted across 5 electronic databases: CINAHL Complete (via EBSCOhost), PubMed (MEDLINE), ScienceDirect (Elsevier), SpringerLink (Springer Nature), and JSTOR. Two concept blocks were combined with the Boolean operator AND. The first block captures the concept of the ethics of care (“ethics of care” OR “care ethics” OR “ethical caring” OR “caring ethics” OR “relational ethics” OR “nursing ethics” OR “moral reasoning”), and the second captures the population (nurs* OR health OR health care OR “health care”). The inclusion of the terms health and health care in the population block was intended to increase the sensitivity of the search so as not to miss nursing-related work that may not use the term nurse explicitly; eligibility screening will then be restricted to work relevant to nursing.

Where supported, controlled vocabulary was applied (CINAHL headings in CINAHL and MeSH terms in PubMed [MEDLINE]), while ScienceDirect, SpringerLink, and JSTOR were searched using free-text terms. In ScienceDirect, full words were spelled out rather than truncated because the platform limits the number of Boolean operators and does not support truncation. Searches were limited to English-language documents, with no date restriction, and automatic term expansion was disabled. The database searches were supplemented by backward citation searching of the reference lists of included studies. All searches were conducted in May 2026. The full, line-by-line search strategy for each database is provided in Multimedia Appendix 1 [23], reported in accordance with PRISMA-S. During the search process in May 2026, a research librarian was consulted, who reviewed the search terms and methods and advised on the search approach.

Studies will be included if they meet the following criteria: (1) population: registered nurses and nursing students, including interprofessional work only where data specific to nurses can be extracted separately; (2) concept: work that draws on Gilligan’s [8] or Tronto’s [10] ethics of care or that addresses any one of the framework’s components; (3) study design: qualitative, quantitative, and mixed methods research; (4) language and time frame: English-language publications, with no date restriction (from database inception to the search date); and (5) publication type: articles published in peer-reviewed journals (with the associated risk of publication bias acknowledged in the Discussion section).

Studies will be excluded if they are nonresearch documents, such as opinion pieces or editorials; do not address Gilligan’s [8] or Tronto’s [10] ethics of care in the nursing context; or are not available in full text for eligibility assessment. Here, “not available” refers to the inability to access the full text, not to incomplete information within an article. Quality appraisal using the JBI tools [25] will be used solely to inform the interpretation of findings and will not serve as a criterion for excluding studies.

References will be managed in EndNote 20 (Clarivate Analytics) and screened in Rayyan (Rayyan Systems Inc). Owing to resource constraints, screening will be carried out by a single primary reviewer. To reduce the risks associated with single-reviewer screening (such as missing relevant studies and selection bias), several safeguards will be applied: calibration with coreviewers on an initial sample of records, maintenance of an audit trail in Rayyan, and consultation with a second reviewer in cases of uncertainty. Single-reviewer screening will be acknowledged as a limitation in the Discussion section.

Data will be extracted using a standardized form that will first be piloted on a sample of included studies. The extracted data will include basic study information (title, authors, year, country, and setting), study details (design, objectives, population, and methods), and findings relevant to the review questions. In particular, data will be extracted on the components of the ethics of care addressed in each study, the conceptual definitions or descriptions of those components, the indicators or behavioral expressions reflecting each component, contextual factors related to their expression, and any additional indicators or aspects not fully captured by the frameworks of Gilligan [8] and Tronto [10]. For quantitative or mixed methods studies, narrative findings relevant to these components and indicators will also be extracted and mapped to the framework. The full line-by-line search strategies are provided in Multimedia Appendix 1, and the final data extraction form is provided in Multimedia Appendix 2.

The methodological quality of included studies will be appraised after inclusion using JBI’s critical appraisal tools appropriate to each study design, following the JBI Manual for Evidence Synthesis [25]. Depending on the design of the included studies, this may include, for example, the JBI Critical Appraisal Checklist for Qualitative Research or other relevant JBI tools. Each study will be assessed using the checklist appropriate to its design, and items rated as “unclear” or “not applicable” will be recorded separately. Owing to resource constraints, appraisal will be undertaken primarily by 1 reviewer, with discussion with a second reviewer in cases of uncertainty. Appraisal results will be used only to inform the interpretation of the findings and will not serve as a criterion for excluding studies; therefore, appraisal will not affect the counts in the PRISMA flow diagram.

Data will be synthesized using the BFFS, following the approach described by Carroll et al [21,22]. Because the review proceeds within the a priori frameworks of Gilligan [8] and Tronto [10], BFFS allows the evidence to be tested against the framework and the framework to be extended from the evidence within a single process. The data to be synthesized are qualitative text extracted from included studies, such as findings, conceptual descriptions, and passages describing caring behavior. All analysis will be conducted in NVivo (version 15; Lumivero) in 6 steps.

For included quantitative studies, findings will be interpreted narratively and mapped onto the relevant components and indicators; no statistical pooling or meta-analysis will be performed, as the review aims to synthesize qualitative definitions and indicators. The synthesis will be presented as a table of the characteristics of included studies; a table of synthesized indicators grouped under the 5 components, with operational definitions and frequency of occurrence; a diagram of the relationships among components and indicators; and an analysis of contextual variation.

On completion, the review is expected to yield clearer definitions and behavioral indicators for each of Tronto’s [10] 5 components, with Gilligan’s [8] ethics of care serving as an overarching lens, and to show how these indicators are expressed and how they differ when applied across various nursing contexts.

Most prior reviews have examined only particular aspects or components. Selvakumar and Kenny [15] reported that empirical work linking the ethics of care with moral resilience remains limited; Waterfield and Barnason [17] focused on nurses’ workload; and Klaver and Baart [27] examined a single component (attentiveness) and noted the absence of clear indicators for that aspect. Other reviews have tended to address negative experiences, such as moral distress and ethical climate [18] or moral resilience [19]. The present review differs by synthesizing the components and indicators of the ethics of care as a whole using the Gilligan [8] and Tronto [10] frameworks, which together encompass all relevant components. It is therefore expected to address a gap not addressed by previous reviews.

A strength of this review is its systematic integration of the Gilligan [8] and Tronto [10] frameworks with BFFS, together with reporting in accordance with PRISMA 2020 and PRISMA-S, which renders the search and selection process transparent and auditable. Several limitations should be considered in interpreting the findings: screening by a single reviewer, albeit with bias-reduction safeguards; the restriction to peer-reviewed journal articles, which may introduce publication bias; differences in health systems and nursing roles across countries, which may affect the comparability of findings; the heterogeneity of definitions of the ethics of care across studies, which may complicate synthesis; and limitations in the database set due to institutional access, which may be expanded as access to additional databases becomes available. Backward citation searching was incorporated to help mitigate the risk of missing relevant studies.

Several amendments were made to the registered protocol to improve conceptual alignment and methodological feasibility. These changes, particularly the removal of the original date restriction and the adoption of the BFFS, may broaden the range of included studies and increase heterogeneity, but they are expected to improve the conceptual coherence of the synthesis and the relevance of the review to its RQs.

The synthesized definitions and indicators can be extended in several directions: the development of evidence-based instruments for assessing the ethics of care, empirical studies to test the framework in real-world contexts, and nursing curricula designed to foster the ethics of care.