The protocol for this review was registered on the Open Science Framework (JCWTD) and reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) checklist.

We conducted a systematic review in 5 major literature databases, which included MEDLINE, Embase, CINAHL, PsycINFO, and Web of Science. No restriction was imposed on the start date of the search, and the review was current as of February 28, 2022. The search strategy used key terminologies related to T2DM, caregivers or caregiving, and remote management of patients (Multimedia Appendix 1 [17,25-34]). The search terms used were adapted from systematic reviews that evaluated remote management or caregiving in other patient populations [23,35-37].

We included full-text articles in English that assessed the role of caregivers or caregiving in remote management of patients with T2DM. Using the “patient, intervention, comparator, and outcomes” (PICO) guide, the patient population of interest was patients with T2DM, while interventions included the provision or inclusion of caregivers or caregiving in the remote monitoring of patients with T2DM. Comparator groups for the included studies included patients with T2DM who received usual care or other remote monitoring interventions without caregivers. Studies that evaluated only the role of caregivers or remote management of patients with T2DM separately, as well as studies that included patients with type 1 diabetes mellitus or maturity-onset diabetes in the young, were excluded. We also excluded case reports, series, irrelevant systematic reviews, and meta-analyses.

All references and abstracts extracted from the databases were exported to the EndNote (version X9; Clarivate). Duplicate citations were removed, and the screening of the titles, abstracts, and full text of the retrieved citations was performed by 2 independent reviewers (JJBS and MFRG) to identify relevant articles for inclusion. All disagreements during the initial screening process were discussed, and any unresolved disagreements were arbitrated by a third independent reviewer (MHAY). Additionally, hand searching of references within the included studies was performed to identify other relevant studies.

Information from included citations was collated in a standardized Excel (Microsoft Corp) worksheet by 2 independent reviewers. We conducted an initial pilot data extraction for the first 15 citations to ensure the accuracy of data extraction. The study details extracted included the study title, publication year, sample size, characteristics of the patient population, details related to caregivers, and modality of remote monitoring. With regard to outcomes of interest, the Economic, Clinical and Humanistic Outcome model proposed by Kozma et al [42] was adapted. It is an integrated approach that considers economic, clinical, and humanistic outcomes in the allocation of health care resources. The outcomes from the included studies were broadly classified into 3 main categories: clinical, humanistic, and economic outcomes based on the model. Clinical outcomes were further subdivided into T2DM related, non-T2DM related, and other health outcomes. Humanistic outcomes encompassed measures related to the patient’s health-related quality of life, other patient-reported outcomes, and behavioral changes [42]. Economic outcomes included health care–related expenditures and health resource use.

In this review, the study quality assessment tools from the National Heart, Lung, and Blood Institute were used to evaluate the risk of bias among the included studies [43]. A total of 2 independent reviewers (JJBS and MFRG) evaluated the methodological quality and risk of bias among the included studies. The checklists consist of items that assess each individual study’s risk of bias associated with the research question, patient population recruitment, outcome evaluation, and patient dropout percentages. All disagreements in the risk of bias assessment were arbitrated with a third independent reviewer (MHAY).

Details related to the characteristics of all included studies are available in Multimedia Appendix 1 [17,25-34]. Descriptive statistics were used to summarize the characteristics of the included studies. Continuous variables were reported as the mean (SD), while categorical variables were reported as n (%). To improve the completeness of this review, the authors of studies with missing information were contacted. A total of 2 separate email reminders were sent out 2 weeks apart, and information that could not be retrieved was labeled as unavailable. Data imputation was not performed in this review.

To evaluate if meta-analyses could be performed for this review, 2 independent reviewers (JJBS and MFRG) examined the clinical and methodological heterogeneity of the included studies. Clinical heterogeneity looks at the differences in the characteristics of patient population and outcomes. On the other hand, methodological heterogeneity refers to variation in study design and risk of bias. Due to the expected heterogeneity of the included studies, meta-analyses were not performed. A narrative synthesis was provided for the role of caregivers in the remote management of patients with T2DM.

Table 1 shows the characteristics of the studies in this review. The majority of the studies (6/11, 55%) were conducted between 2001 and 2010. Most of the studies were conducted in North America (7/11, 64%), and the most common study design was a randomized controlled trial (5/11, 46%). Most studies were conducted in a primary health care setting (8/11, 73%), and all studies used primary data sources (n=11). The majority of the caregivers in the studies involved family members or friends (10/11, 91%), while the most common remote monitoring modalities were interactive voice response (5/11, 46%) and phone consultations (4/11, 36%). With regard to the risk of bias, 3 studies were noted to have a moderate risk of bias, while the remaining studies were assessed to have a low risk of bias (Multimedia Appendix 1 [17,25-34]).

Table S1 in Multimedia Appendix 2 [17,25-34] shows a summary of study details related to the 11 studies included. The follow-up period across studies ranged from 3 months to 3 years.

A study explored the impact on patients’ health care use. The study showed significantly higher rates of 12-month hospitalizations compared with the control group (P≤.05) [34].

To the best of our knowledge, this is the first systematic review that has evaluated the role of caregivers in the remote management of T2DM. Current evidence in the literature appears to suggest a positive impact of caregivers on T2DM with regards to clinical, economic, and humanistic outcomes ranging from diabetes-related parameters to patients’ compliance to monitoring.

The positive impact of caregivers on remote monitoring of T2DM is likely multifactorial. Studies have shown that common barriers to remote monitoring include patient-related factors such as health literacy, technology-related barriers, and challenging patient experiences with remote monitoring technologies [44]. Through the partnership with caregivers, knowledge gaps related to technical aspects of remote monitoring technologies and T2DM can be bridged, which in turn aids in improving patients’ adherence to medication and remote monitoring services [45]. This is important in the context of varied levels of electronic health literacy among older patients, which often impair their ability to comply with remote monitoring technologies [46]. In addition, patients with poor language skills, poor digital literacy skills, and equipment deficiencies are often excluded from remote monitoring programs and studies due to technological and logistical limitations [47]. With caregivers serving as potential avenues to bridge some of the modifiable gaps, there is a potential role to allow for the expansion of remote monitoring technologies to these previously excluded patient populations if suitable caregivers can be identified to empower patients.

With regard to the mixed evidence related to improvements in T2DM-related laboratory parameters, there are several reasons that may account for the lack of statistical differences in T2DM parameters in half of the studies. For example, the study by Burner et al [29] was limited by the study sample size as it was intended to be a feasibility study and was not adequately powered to evaluate differences in glycemic control between the intervention and control groups. Likewise, the study by Gambling and Long [30] sought to qualitatively evaluate telephone-based support for patients with T2DM and recruited only 9 patients. Another postulated reason could be due to underestimation of the effects of an intervention when it is offered alongside routine care, which has been seen in other studies evaluating the education of patients with T2DM [48]. Hence, larger studies are required to confirm and evaluate the impact caregivers have on the clinical outcomes of patients with T2DM.

Interestingly, the study by Wakefield Vaughan-Sarrazin [34] showed an increased number of hospitalizations among patients with T2DM on remote monitoring with caregivers compared with the control group. In this study, one of the key findings was a significantly higher caregiver burden and strain in the intervention group. This highlights the importance of recognizing the potential impacts on caregivers when engaging them in the remote monitoring of patients with T2DM. A review performed by Doherty et al [49] found that caregivers also have their own unique set of needs, which center around psychosocial support to maintain normalcy, maintaining their social lives, and support with navigating health care systems. Likewise, caregivers may also suffer from underlying medical conditions, and studies have shown that poor social support and depressive symptoms are associated with poor outcomes in caregiving [50]. Familial strife may also ensue from the overload of caregiving tasks, limited financial resources, and underappreciation of the work rendered by caregivers [51]. With the looming shortage of caregivers and decreasing caregiver-to-patient ratio globally due to declining fertility rates [52], developing strategies to ensure efficient and safe enlistment of caregivers in remote monitoring of patients with T2DM will be crucial. Some interventions suggested before the initiation of remote monitoring strategies include having caregivers attend clinics to allow physicians to formally assess them for their suitability for remote monitoring programs [53].

In this review, only 2 studies evaluated the potential of mHealth or web-based apps with caregivers. Nudge strategies, which encompass reminders, gamification, and social modeling approaches to drive positive health-related behaviors, have been increasingly recognized as a potential solution to improving diabetes control [54]. With advances in information and communication-related technology, there is a growing number of multifunctional social media and mobile apps that can deliver scalable behavior nudges for patients with T2DM [55]. For example, mHealth applications on a widely used Chinese social media platform, WeChat, allow for the sharing of blood pressure readings with patients’ caregivers, which allows them to aid in nudging patients [56]. Future studies should look into exploring the role of mHealth apps and the involvement of caregivers in diabetes care.

The results of this systematic review should be interpreted in the context of the following limitations: due to significant heterogeneity in the training of caregivers and outcomes assessed across the studies, meta-analyses could not be performed. Future reviews should consider performing meta-analyses of the developing literature on remote monitoring technologies used in patients with T2DM and the role of caregivers. Another limitation was the lack of standardized outcomes assessed across studies. Future studies should look into developing a core set of outcomes for studies examining remote monitoring technologies for diabetes patients. Currently, generic instruments such as the CONSORT-eHealth checklist have been developed, but disease-specific instruments are required for the assessment of patients with T2DM [57]. This will in turn aid in facilitating meaningful cross-comparisons of different caregivers and their impact on remote monitoring of patients with T2DM. Lastly, the follow-up period of the studies included in this review is generally short. One of the main challenges with the use of remote monitoring technologies is their increasing attrition rates with time, which can be as high as 50% [58,59]. More longitudinal studies are required to assess the role of caregivers in reducing patient attrition rates and their long-term impact on patients with T2DM who are remotely monitored.

This review has shown that caregivers may play a role in improving clinical outcomes among patients with T2DM under remote monitoring. More studies are required to understand the impact of caregivers on less studied populations, such as Asian populations, and long-term patient outcomes.