Study registration and Protocol
This review was registered with the International Prospective Register of Systematic Reviews or PROSPERO (CRD42020196301). We will also follow the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) in our research procedures [32] to further safeguard research rigor.
Search Strategy
The following databases will be searched for potential articles: PubMed, PsycINFO CINAHL, and Scopus. The search will be limited to original articles published in English from the database inception to the end of March, 2021. Searches incorporated medical subject heading (MeSH) and keyword terms in three categories: cancer, caregivers, and technology platforms. Our search strategy will be developed in consultation with an academic librarian experienced in systematic review studies to ensure research rigor. Snowballing (manual searching reference lists) of included studies will be conducted to obtain additional eligible articles. Furthermore, reverse tracing potential eligible manuscripts that cited included papers will be administered via Google Scholar. An example search string is listed in Table 2.
Table 2. Example PubMed search string
Concept
|
Search string
|
Cancer
|
cancer*[MeSH] OR cancer*[TIAB] OR tumor*[MeSH] OR tumor*[TIAB] OR tumour*[MeSH] OR tumour*[TIAB] OR neoplasms[MeSH] OR neoplasms[TIAB]
|
Caregivers
|
caregiver*[MeSH] OR caregiver*[TIAB] OR famil*[MeSH] OR famil*[TIAB] OR spous*[MeSH] OR spous*[TIAB] OR dyad* [MeSH] dyad* [TIAB] OR partner*[MeSH] OR partner*[TIAB] OR couple*[MeSH] OR couple*[TIAB]
|
Technology-based Interventions
|
“technology”[MeSH] OR “technology”[TIAB] OR “eHealth”[TIAB] OR “telemedicine”[MeSH] OR “telemedicine”[TIAB] OR “tele-medicine”[MeSH] OR “tele-medicine”[TIAB] OR “telehealth”[TIAB] OR “tele-health”[TIAB] OR “connected health”[TIAB] OR “digital health”[TIAB] OR “mHealth”[TIAB] OR “mobile health”[TIAB]
|
Inclusion and exclusion criteria
In the context of this study, caregivers are defined as patients' family or friends who may offer mostly long-term care to patients, often with little or no financial compensation of any form. This paper broadly defines interventions as stimuli or mechanisms that are aimed to produce changes in outcome variables (e.g., self-care abilities increased). Studies will be included if they are published in English, have relevant information on technology-based interventions for cancer caregiving, with detailed inclusion criteria listed in Table 3. Ensuring data quality, comments, editorials, gray literature, and reviews will be excluded from the review. Overall, articles will be excluded if they: (1) did not include a cancer caregiver population, (2) did not provide information on intervention, and (3) did not describe how technology is integrated into the intervention strategy.
Table 3. Study inclusion criteria
Data type
|
Inclusion criteria
|
Participants
|
Cancer caregivers (≥ 18 years old)
|
Language
|
English
|
Study type
|
Original journal article
|
Study design
|
Focus on technology-based intervention that aim to improve cancer caregivers' health and wellbeing
|
Intervention
|
Technology-based interventions for cancer caregivers
|
Outcome
|
Report empirical findings on intervention outcomes
|
Selection of studies and data extraction
Search results will be managed using Rayyan [33], a free web application that allows sorting and storing articles will be used to remove duplicate records and screen articles. Both two phrases of screening, title-abstract screening and full-text screening, will be conducted by two primary reviewers (ZS and XL) independently. Discrepancies will be solved by consensus, and when needed, with input from the rest of the research team. Data will be extracted independently by the reviewers (ZS and XL) based on the research aim and selection criteria adopted in this study. Specifically, for studies that met the inclusion criteria, the primary reviewers (ZS and XL) will extract the following information from the final included studies: study and participant characteristics (e.g., study aim), intervention characteristics (e.g., the use of technology in interventions), and details on study outcomes (e.g., intervention outcomes).
Data synthesis and analysis
If eligible studies are found to be heterogeneous, we will conduct a narrative synthesis instead, as opposed to a meta-analysis, to summarize key insights from the data. A summary of the key information extracted will be utilized to synthesize the main research findings. Descriptive analysis will be performed on categorical variables. If there are enough similarities in eligible papers to be pooled, a meta-analysis will be conducted to obtain further insights from the data. The Cochrane's Q test and I2 test will be adopted to calculate heterogeneity within studies— I2 values of 25%, 50%, and 75% suggest low, medium, and high heterogeneity, respectively [34]. If much heterogeneity is present, random-effect models will be adopted (as opposed to fixed-effect models), as these models are more robust in detecting large variations in studies [35]. Review manager 5.3 (Cochrane Collaboration software) will be utilized to conduct the random effect model. Visual inspection of funnel plots will be used to evaluate publication bias if needed. If enough data are available for meaningful investigation (e.g., theoretically or clinically meaningful), subgroup analyses will be conducted for diverse types of cancer types, stages, interventions, follow-up duration, caregiver types, and country. A sensitivity analysis will be conducted by sequentially removing one study periodically and reanalyzing the data to evaluate the impact of individual studies on overall outcomes. This process will allow potentially methodologically flawed research included in the review; a study will be considered acceptable if it affects the meta-estimate of less than ± 5%.
Risk of bias and quality of evidence
The Cochrane Collaboration evaluation framework will be utilized to investigate the risk of bias of articles that met the inclusion criteria, including selection, performance, detection, attrition, reporting, and other biases [36]. The framework has seven domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and any other source of bias. Two primary reviewers (ZS and XL) will evaluate the included articles' quality independently, and give the articles a score (high, medium, or low) based on results of the evaluation. These reviewers will also independently adopt the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) framework to assess the overall quality of evidence of the eligible articles. The GRADE guideline has five domains: risk of bias, indirectness, inconsistency, imprecision, and publication bias [37]. Any discrepancies will be resolved via group discussions till a consensus is reached.