Objectives 1 and 2 will follow the Joanna Briggs Institute (JBI) review methods [33]. The protocol was registered on PROSPERO in January 2021. Any changes or updates to the protocol (e.g., eligibility criteria, search strategy, data extraction, or data analysis) will be documented in the final manuscripts.

Eligibility criteriaPopulation: older adults

Studies of older adults, with or without cognitive impairment or dementia. The focus is on older adults as they comprise the majority of LTC residents [8, 9]; studies will not be restricted to only those of older adults. However, studies will only be included if the mean age of participants is 65 years or older (or at least 2/3 of participants are 65 years or older).

Concept: social connection

Studies of measures assessing any objective or subjective aspects of social connection (Fig. 1), defined according to working definitions provided in Additional file 1: Appendix I [1, 27,28,29,30,31, 34]. Measures with subscales and items assessing aspects of social connection will be included if these subscales/items are reported separately. However, studies of measures that provide only the summary scores of other concepts (e.g., overall quality of life) will be excluded.

Context: LTC homes

Populations must reside in a LTC home setting defined according to the international definition of a nursing home [35]. Studies must report that at least 2/3 of participants were LTC home residents or present results for LTC home residents separately. Studies conducted exclusively in other congregate settings (e.g., assisted living, hospice, independent living, retirement homes) will be excluded.

Types of sources

This review will consider primary research publications with no language restrictions. Relevant reviews will also be scanned for eligible publications. Consistent with COSMIN recommendations, secondary texts, literature reviews, conference abstracts, editorials, and dissertations will be excluded as they do not have sufficient detail regarding study design. We worked with an experienced information specialist to develop comprehensive, inclusive search strategies that were employed in multiple bibliographic databases (see Additional file 2: Appendix II). Grey literature will not be included.

Search strategy (objectives 1 and 2)

The search strategy was adapted from a previous study [36] and updated by an experienced information specialist (see Additional file 2: Appendix II). Electronic databases that will be searched include MEDLINE ALL (Ovid), Embase Classic and Embase (Ovid), Emcare Nursing (Ovid), APA PsycInfo (Ovid), Scopus, CINAHL Complete (EBSCOhost), AgeLine (EBSCOhost), and Sociological Abstracts (ProQuest). Databases will be searched for published research studies that quantify any aspect of social connection that has been identified for use in research in LTC homes. The working definitions of the key aspects of social connection, relevant to the conceptual model (Fig. 1), are provided in Additional file 1: Appendix I. A broad range of search terms will be used to acknowledge relevant studies that might use alternative terminology. To meet the context requirements of this study, terms including long-term care, nursing home, care home, residential home, and home for the elderly will be used to reflect the variance in terminology sometimes used to describe LTC homes [35]. To search for studies that report measurement properties, the COSMIN filters, which have a sensitivity of 97.4% and a precision (akin to positive predictive value) of 4.4%, will be applied [37]. We will also scan reference lists of included studies and contact experts to seek additional eligible studies.

Study/source of evidence selection (objectives 1 and 2)

Citations will be exported from each database, uploaded to Endnote X9.1 (Clarivate Analytics, PA, USA), and imported into Covidence for duplicate removal. A pilot test screening for 15 papers (titles and abstracts) will be conducted to familiarize reviewers with eligibility criteria. Following the pilot test, titles and abstracts will be screened by two reviewers. For full-text review, publications will be imported into Covidence and assessed by two reviewers. Non-English papers will be assessed by additional reviewers with relevant language and research expertise. Reasons for exclusion will be stated as (1) types of sources (e.g., no measurement properties reported), (2) setting: not LTC, (3) concept: not social connection, and (4) population (e.g., not older adults, not LTC residents). Reasons for exclusion will be recorded in Covidence. Any disagreements that arise throughout the screening of abstracts/titles or full text will be resolved through a discussion with a third reviewer. The results of the search and study eligibility will be reported in the final review and presented in a Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram [33]. We used the PRISMA-P checklist for this protocol [38].

Data extraction (objectives 1 and 2)

Data will be extracted independently by two researchers using standardized data extraction instructions.

For each included study, the data extraction will include population (country, race/ethnicity, inclusion criteria, exclusion criteria, sample size—number of residents and homes, gender/sex, age), social connection concept, name of measure, response options, mode of administration, and observation period. A draft of the extraction form is provided in Additional file 3: Appendix III. The extraction form will be modified as necessary during data extraction and analysis. Modifications will be detailed in the final reviews. Any disagreements between reviewers will be resolved through discussion or by a third reviewer.

Objective 1

Additional data that will be extracted are the individual items (questions and statements) that are used to assess social connection.

Objective 2

Data will be extracted using the predefined data collection template provided by COSMIN (available at www.cosmin.nl). This includes data on each measure’s measurement properties (i.e., content validity, structural validity, internal consistency, cross-cultural validity/measurement invariance, reliability, measurement error, criterion validity, construct validity, and responsiveness).

Data analysis

Selected studies and measures will be summarized in tables reporting frequency and percentage statistics to describe the studies (author, year of publication, population) and measure (social connection aspects, name of measure, response options, mode of administration, and observation period).

Objective 1

Objective 1 will be analyzed using an adapted framework method [39].

Stage 1.1: transcription

A copy of each measure will be obtained from the original study, online search, or contact with study authors. Studies that report on the same measure will be grouped; measures will be identified as dementia-specific or generic (non-dementia-specific). The name of each measure and the terms it uses to describe social connection will be identified.

Stage 1.2: familiarization

The items within each measure will be reviewed. Thoughts and impressions concerning how items describe social connection aspects will be documented in a journal. Familiar aspects of social connection (Fig. 1) will have their items reviewed alongside the working definitions (Additional file 1: Appendix I). Unfamiliar aspects of social connection will have their items reviewed so that a working definition can be created. Thoughts and impressions on how these unfamiliar aspects fit into the preliminary conceptual model (Fig. 1) will be recorded in a journal.

Stage 1.3: coding

A hybrid deductive-inductive coding approach will be used [40] where codes are based on the working definitions of social connection aspects [39]. Using a sample of five to ten randomly selected measures, two researchers will code the items in these measures separately.

Stage 1.4: developing a working analytical framework

Following coding of the sample measures, researchers will meet and compare the coding labels. Researchers will agree on a set of codes to apply to subsequent measures. Codes may be grouped together if both researchers agree their working definitions overlap. Any disagreements will be resolved through discussion with the wider research team. The preliminary conceptual model for social connection (Fig. 1) will be revised to include any new aspects of social connection. This revised conceptual model will form an analytical framework for social connection in measurement.

Stage 1.5: applying the analytical framework

Each measure’s items will be sorted into the analytical framework for social connection in measurement. An electronic software such as Computer Assisted Qualitative Data Analysis Software (CAQDAS) may be used to complete this process.

Stage 1.6: charting the data into the framework matrix

The data from the analytical framework will be used to generate a matrix (table) whereby the rows represent “cases” (measures) and the columns represent the codes (aspects of social connection) developed in stage 1.4. The items in each measure will be charted into the appropriate column (i.e., according to the aspect of social connection).

Stage 1.7: interpreting the data

Data will be interpreted in three steps. First, each measure will be summarized according to the codes to which it was mapped; dementia and non-dementia-specific measures will be compared by tabulating the presence of codes. Next, inductive coding will be applied to identify themes within codes and compare them across dementia and non-dementia-specific measures.

Objective 2

Objective 2 will be conducted and reported following the COnsensus-based Standards for the selection of health Measurement Instruments (COSMIN) guidelines [41]. COSMIN outlines rigorous, transparent methodology and standards used to assess measurement properties, feasibility, and interpretability of a measure, and the methodological quality of the studies which report measurement properties.

Stage 2.1: assessment of measurement properties

Two researchers will extract data on each study’s measurement properties, with measurement properties being defined using the COSMIN taxonomy [42], including content validity, structural validity, internal consistency, cross-cultural validity, reliability, measurement error, criterion validity, construct validity, and responsiveness.

Measurement properties will be rated as sufficient ( +), insufficient ( −), or indeterminate (?) [26]. COSMIN’s [43] modified criteria for good measurement properties (adapted from Terwee et al. [44] (Additional file 4: Appendix IV) will be used to provide evidence of the quality of measurement properties.

Stage 2.2: evaluate measurement properties

Two researchers will independently and systematically evaluate each measure of social connection by providing summarized results of each measure’s measurement properties. Completed versions of the COSMIN risk of bias checklists on patient-reported outcome measure development, content validity, structural validity, internal consistency, cross-cultural validity/measurement invariance, reliability, measurement error, criterion validity, hypotheses testing for construct validity, and responsiveness will be used to compute the overall ratings for summarized results. Overall ratings for these results will be labeled as sufficient ( +), insufficient ( −), or indeterminate (?) (Additional file 4: Appendix IV).

Stage 2.3: grading the quality of the evidence

The research team will summarize the evidence for each measure (and each measurement property) and provide an overall rating using a modified Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach [45]. This approach grades the quality of the evidence as high, moderate, low, or very low.

Patient and public involvement (objectives 1 and 2)

During this review, people living with dementia and friends, family or current/former caregivers of people with dementia and/or in LTC homes will be involved through the Canadian Consortium on Neurodegeneration in Aging’s Engagement of People with Lived Experience of Dementia (EPLED) program (www.epled.ca) [46] and the Alzheimer’s Society (UK) Research Network. During the preparatory phase [47], EPLED members were involved in identifying the research topic. During the execution phase, people with lived experience will contribute to interpreting results and knowledge translation. In particular, preliminary results will be presented to panel members in workshops where they will be asked to provide feedback (oral or written) and assist in identifying relevant themes and patterns surrounding important aspects of social connection. As part of the SONNET study, people with lived experience will also assist in knowledge translation by disseminating results to academic and non-academic audiences.