Do caregivers’ involvement in Type 2 diabetes education affect patients’ health outcomes?: A systematic review and meta-analysis

Introduction: The prevalence of Type 2 diabetes mellitus (T2DM) is rising worldwide. Patients frequently struggle with controlling their diabetes and need the assistance of caregivers for effective self-management because managing diabetes requires a variety of strategies, including diet, glucose monitoring, and exercise. This study aimed to examine the effect of caregiver involvement in T2DM education within a community on patients’ diabetes care outcomes. Methods: Based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, a systematic review of all published studies from the earliest record to May 2022 that reported adult caregivers of T2DM patients who participated in educational interventions concerning diabetes management and that reported one or more outcomes of the interventions were conducted. Four databases were used, including PubMed, Cochrane Library, EMBASE, and CINAHL. The meta-analysis focused on glycated hemoglobin (HbA1c) levels among randomized controlled trials (RCTs), with additional attention to lipid levels. Review Manager 5.4 was used to perform this meta-analysis. Results: A total of 17 out of 683 studies were synthesized. Involvement of caregivers in T2DM education is shown to reduce body mass index and HbA1c. This involvement also improves patients’ knowledge, physical activity, and self-efficacy, but the effect on medication adherence varies. A meta-analysis of six RCT studies shows that caregiver involvement in T2DM education reduced pooled HbA1c levels by 0.83 (95% Confidence interval: −1.27–−0.38) compared to involvement ( p = 0.0003). Meta-analysis of three types of lipids (low-density lipoprotein, total cholesterol, and high-density lipoprotein) showed no strong evidence that caregiver participation in diabetes education improved lipid levels. Conclusions: Caregivers play key roles in diabetes management and can contribute to improving patient HbA1c levels. Future research should focus on enhancing caregiver participation in T2DM education.


INTRODUCTION
Type 2 diabetes mellitus (T2DM) is a major public health concern (1,2), and the International Diabetes Federation predicts that the number of T2DM patients will continue to rise (3).Although blood glucose levels can be controlled through diet, exercise, and/or medication adherence, glucose management remains challenging for those with T2DM.Community-based diabetes management programs have been implemented in response to these challenges.The previous studies show that patients who participated in community-based diabetes programs lowered their fasting blood glucose levels or improved other health indicators related assistance from their caregivers to effectively manage T2DM in the community.
Caregivers often share responsibility in managing disease.They can provide significant support to the patient in following medical treatment recommendations or in self-management activities (such as through making grocery purchases, refilling prescriptions, or transporting patients to appointments), and may also provide psychological support (12).Due to the importance of the caregiver's role, patient-caregiver involvement and education in diabetes management have recently been highlighted.Supportive caregivers (e.g., spouses or family members) can improve the quality of life and self-management of diabetes patients (13).On the other hand, a lack of support from family members could negatively impact medication compliance and blood glucose levels (12).The previous research indicates that caregiver involvement in diabetes management is a critical component of successfully managing T2DM at home and in the community.Especially given the distinctive environmental characteristics of communities as compared to hospital settings, the presence of caregivers, the degree of care participation, and caregiver education can impact the self-management of diabetes patients.
A recent meta-analysis by Kodama et al. (14) examined the effectiveness of family-oriented diabetes programs on glycemic control among both Type 1 and Type 2 diabetes patients of all ages and found that participation in the analyzed intervention programs decreased mean HbA1c (14).Despite this study providing important insights into the effects of family involvement in diabetes care, stratifying analysis and review by caregiver dynamic, diabetes type, and age group can provide a more precise picture of the relationships among various factors.Moreover, this study was conducted in 2019 with the data collection performed in 2017, and only experimental studies with HbA1c as an outcome were included in the study.According to the Cochrane Handbook, reviews are recommended to be updated approximately every 5.5 years (15).Other recent studies only looked at certain areas of diabetes management, such as foot ulcers (16).Our study not only investigates experimental studies but also observational studies with behavioral outcomes, thus providing a more comprehensive understanding of the subject area with more recent data.
The purpose of this study is to find out the effect of caregiver participation in diabetes education on patient health.To this end, we conducted a systematic literature review to examine the association between caregiver involvement in adult T2DM education within a community setting and patients' diabetes care outcomes.The outcomes focused on are biological results (e.g., HbA1c level) and self-management outcomes (e.g., diabetes knowledge).Further, a meta-analysis was conducted to determine the effectiveness of caregiver involvement in T2DM education on HbA1c levels and lipid levels.The results from this study will help diabetes educators and policymakers make decisions about whether to include caregivers in providing more effective patient diabetes education.

METHODS
A systematic literature search was conducted using four electronic databases (i.e., PubMed, Cochrane Library, EMBASE, and the Cumulative Index to Nursing and Allied Health Literature) on research published from the earliest record to May 2022.Medical Subject Headings (MeSH) terms and keywords were utilized, such as "diabetes mellitus, Type 2," "diabetes self-management education," "caregivers," "home care," and "community health."A full list of search terms used and the total number of studies identified from each database are presented in Table 1.
This study used a PICO-SD (Participants, Intervention, Comparison, Outcomes, Study Design [SD]) tool as follows: (1) Population (P), Type 2 diabetes patients and caregivers; (2) Intervention (I), patient diabetes education with caregivers (with interventions including self-management education, health education); (3) Comparison (C), usual care (without caregivers) or no control group; (4) Outcome (O), biological results (e.g., HbA1c level), self-management outcomes (e.g., diabetes knowledge); and (5) SD, interventional studies including RCTs.Furthermore, a manual search was conducted using the reference lists of all the included studies.
According to the Centers for Disease Control and Prevention, a caregiver is a person who provides care to someone who needs some ongoing assistance with everyday tasks on a regular or daily basis (17).Caregivers are referred to as either "formal" or "informal."Informal caregivers, also called family caregivers, provide care to friends or family, typically without payment (18).In this study, "caregiver" refers to informal caregivers who are not professional healthcare providers, personal care workers, or home health aides.In other words, caregivers in this study include those who live with the patient and care for the patient (e.g., spouses, adult children, friends, and cohabitants).We reviewed studies that included caregivers in the education of adults with diabetes in the home and community, but not in hospitals.The specific inclusion criteria were as follows.
All identified studies were imported into a citation management tool (EndNote), then duplicate studies were removed using the "find duplicates" function and manually (19).Three authors (JK, JS, and AT) independently conducted title/abstract screening of studies using Covidence systematic review software.Then, a full-text review was conducted for the final selection.Any discrepancies were resolved through weekly discussion between all authors until a final consensus was achieved.The details of the selection process are presented in Figure 1.
All authors independently conducted data extraction and quality assessments of the included studies.The following variables were extracted and used to synthesize the findings: SD/setting, country of study origin, study aim, sample size, and eligibility for study participation, characteristics of 115,667 S2 "Independent Living" OR "independent living" OR "community-dwelling" OR "community based multicenter" OR "community health" OR "Home Care Services" OR "Home Care Agencies" OR "Home Nursing" OR "Home Care Agencies" OR "Community Health Services" OR "home care" OR "community setting" OR "community" OR "outpatient" 451,784 S3 ((MH "Patient Education") OR (MH "Education") OR (MH "Adult Education") OR (MH "Education, Non-Traditional")) OR ("diabetes education" OR ("education in diabetes" OR "nursing education" OR "patient education")) 163,502 S4 (MH "Dependent Families") OR (MH "Family") OR "family caregiver" OR "supporter" OR "spouse" OR "family caregiver" OR "informal caregiver" OR "caretaker" OR "caregiver" OR "patient's family" OR "family-based" OR "couple-based" OR "care-partner" OR "in-home supporter" OR "supporter" OR "care giver" OR "spouse" OR "family caregiver" OR "informal caregiver" OR "care taker" "diabetes mellitus" OR "non insulin dependent diabetes mellitus" OR "diabetes" OR "T2DM" OR "diabetic" OR "type 2 diabetes" 1,465,363 #2 "independent living" OR "community-dwelling" OR "community dwelling person" OR "community based multicenter" OR "community health" OR "community care" OR "home care services" OR "home care" OR "community setting" OR "outpatient care" 383,209 #3 "patient education" OR "patient engagement" OR "self-care" OR "self management support" OR "diabetes education" OR "diabetes self-management education" OR "diabetes patient education" OR "diabetes education program" OR "education program" OR "health education" OR "educational intervention" OR "supportive educational intervention" OR "medical education" OR "early intervention" 807,962 #4 "caregiver" OR "informal caregiver" OR "informal caregiving" OR "family" OR "patient's family" OR "family-based" OR "supporter" OR "care partner" OR "couple-based" OR "in-home supporter" OR "spouse" 1,671,672 #5 "child" OR "pediatrics" OR "insulin dependent diabetes mellitus" OR "t1dm" OR "type 1 diabetes" This study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (22).The included literature was qualitatively synthesized and the results of randomized controlled trials (RCTs) that studied the effect of caregiver involvement in diabetes education on HbA1c and lipids were included for meta-analysis.Qualitative synthesis was performed to meet the synthesis without meta-analysis guidelines (23).A descriptive analysis was performed on the characteristics of all the participants in the included studies (n = 17) if made available by those studies, using the mean and standard deviation (SD).However, for the following demographics, the mean and SD values were calculated and presented only for the studies that provided the relevant information.Those characteristics are diabetes duration (n = 8), comorbidity (n = 10), mean age of caregivers (n = 6), and gender of caregivers (n = 6).
The results of the qualitative synthesis were primarily divided into two groups: Biological indicators and behavioral indicators (or self-management indicators).This is because the characteristics of clinical results and behavioral outcome indicators are distinct.Thus, we as clinical experts, including a certified diabetes educator, determined that it would be more appropriate to analyze the biological and behavioral indicators separately in examining the impact of the intervention.In addition, many studies examined behavioral outcomes for caregivers' participation in diabetes education, thus excluding these studies would lead to a loss of significant scientific evidence.Therefore, we determined that presenting behavioral results rather than just the clinical results would make the research findings more comprehensive.Since HbA1c is directly related to the level of diabetes, it is considered the primary outcome among clinical indicators in our study.The results of the qualitative synthesis were presented as the mean (standard deviation), median, and range of each result value.The summary of the effect estimates pre-and post-intervention is also supplied, together with the p-value.
Of all the included studies, only RCTs that examined HbA1c and lipids as an outcome were included for the meta-analysis.RCT studies were excluded from the synthesis if they met the following criteria: A study that received a low-quality rating in the quality assessment (e.g., one in which a high bias was identified in most domains according to the RoB) (24); a study in which pre-and post-outcome values were not sufficiently presented to perform meta-analysis ( 25); a study difficult to compare with other studies due to the presence of numerous intervention groups (26).The size of the effect was calculated using Review Manager 5.4 (RevMan) provided by The Cochrane Collaboration (27) and heterogeneity was estimated using the χ 2 test and the standard I 2 test.The Cochrane Handbook defines moderate heterogeneity as 50-75% between studies (28,29).Because of some heterogeneity between studies, a random-effects meta-analysis model was used to calculate the effect size for each study and the pooled HbA1c effects across the studies.
The effect size of the result value was presented as mean difference.To identify potential causes of heterogeneity, subgroup analyses were performed.Meta-analysis was considered for other indicators, but only HbA1c and lipids were found to be eligible.

RESULTS
The study selection process is depicted in Figure 1.A total of 683 studies were initially identified: 678 studies from the database search and five studies from the manual search.After duplicates were removed (n = 247), additional studies were excluded after title and abstract screening (n = 356).A total of 80 studies were included in full-text reviews, and 63 studies were excluded for failure to meet inclusion criteria.In total, 17 studies were included for qualitative analysis in this systematic review, and six RCTs were included for quantitative analysis in the meta-analysis (30)(31)(32)(33)(34)(35).HbA1c levels in all six studies were analyzed, but only some of the six were used in the analysis of lipids, depending on the type(s) examined in a given study.More specifically, low-density lipoprotein (LDL) and High-density lipoprotein (HDL) were analyzed using two articles (30,34), while total cholesterol (TC) was analyzed using three articles (30,34,35).
The major aim of the included studies was to investigate the effectiveness/impact of caregiver-involved educational intervention on T2DM patient outcomes.A full list of data extraction is shown in Table 2.
The risk of bias for the nine RCT studies is shown in Figure 2. Overall, the risk of bias in most domains was low for all of the studies except one (24).The most frequently reported "low risk of bias" domain was selective reporting, and the most commonly reported "high or unknown risk of bias" domain was blinding of participants and personnel.Figure 3 shows the risk of bias for eight non-RCT studies.Among non-RCT studies, selective outcome reporting bias was the most frequently reported "low risk" area.In most studies, the risk of bias was low in the participant selection and measurement exposure domains, and it was "unclear" in blinding of outcome assessment domain.
All 17 studies included in our qualitative analysis reported the age, gender, and HbA1c levels of the patients.Overall, the mean age of the included patients (n = 1816) was 55.15 years (SD 4.59).More than half of patients were female (62.74%).In general, the diabetes of patients was poorly controlled, with an average HbA1c of 8.67% (SD 1.17; range 6.0-10.46%).The duration of diabetes was reported in eight studies and was 7.66 years on average (SD 2.37; range 3.4-10.6).In addition, ten studies reported patient comorbidity, with over 60% of patients having at least one comorbidity (e.g., hypertension, hyperlipidemia,  cardiovascular disease).Among the studies that provided caregiver demographic information (n = 6), the mean age of caregivers was 51.1 years (SD 7.17), and 63% were female.
All 17 studies presented the caregiver as a family member, but the type of family member designated varied from a spouse, relative, or significant other, including friends.In     addition, nine studies required that family members live in the same residence as patients in their SDs, whereas the remaining eight studies did not require cohabitation or did not present residency status as inclusion criteria.Of these different types of family caregivers involved, the majority were spouses (63%) followed by adult children (13.5%).
On the other hand, three studies included friends as caregivers along with family members (33,37,40).
The educational program in another study concentrated on skill-building and social support (40).
HbA1c and body mass index (BMI) decreased as a result of caregiver involvement in T2DM education; however, the effects on lipid profiles were inconsistent.Involving a caregiver in T2DM education could also significantly increase a patient's diabetes knowledge, level of physical activity, and self-efficacy, but the impact on medication adherence varied.
All studies measured changes in HbA1c as a metric of the effectiveness of caregiver education.Overall, HbA1c significantly decreased by 1.21% after a given educational intervention (range 0.41-4.3%).However, three studies reported that there was no statistically significant improvement between intervention versus control groups, although HbA1c levels decreased over time (26,30,38).
Nine studies examined changes in patients' lipid profiles.Some studies found that after caregiver participation in T2DM education, patients exhibited significant improvement in LDL, TC, and triglycerides (26,34,35).However, most studies showed no statistically significant change between pre-and post-tests or between intervention and control groups.LDL was used as an outcome indicator in five studies, and in one study, LDL decreased by 0.36 mmol/L (p = 0.041) following intervention (34).In four studies, LDL was lowered, although the reductions were not significant (p > 0.05) (30,38,41,43).Five studies treated HDL as an outcome indicator, and HDL value increased or decreased depending on the study, but none of the findings were statistically significant (p > 0.05) (30,34,36,41,43).In two of the five studies that used TC as an outcome indicator, it was significantly decreased by 18.24 mg/dL (range: 12.50-23.97mg/dL) (p < 0.05) (26,35).In the rest of the three studies, TC dropped, but it was not significant (p > 0.05) (30,34,40).Triglycerides were treated in four studies, and only one study found significant improvement compared to the control group (p = 0.003) (26).In the other three studies, there was no statistically significant change between the intervention and control groups (34,40,41).
Nine studies reported weight-related patient outcomes, and the results varied.Two studies reported significant reductions in BMI by 1% after caregiver involvement in T2DM education (range 1.12-1.28%)(41,42).Three studies reported a reduction in BMI, but it was not significant (30,35,38).Two studies reported no difference in BMI in the group with caregiver involvement in T2DM education (31,34).Two studies reported a slight but not statistically significant increase in BMI (36,43).
Six studies that reported on diabetes knowledge revealed variations in patient knowledge before and after interventions.The caregiver involvement in T2DM education group had significantly higher diabetes knowledge scores in two studies that examined diabetes knowledge using the Diabetes Knowledge Questionnaire (33,34).Other studies that assessed diabetes knowledge using the Spoken Knowledge of Diabetes in Low Literacy Patients with Diabetes tool also revealed gains in diabetes knowledge over time (38,39).In addition, Cai and Hu found that between pre-and post-intervention, the mean diabetes knowledge score increased by 161% (41).Likewise, where Kang et al. utilized knowledge and attitude toward the diabetes questionnaire, authors reported an increase in the mean difference score for patients in the family partnership intervention care group compared to the control group (5.32 vs. 2.32, respectively) (30).
Physical activity was examined in four studies.In two studies (26,31), physical activity significantly improved by a minimum of 1.55 times to a maximum of 2.18 times after family participation intervention (p < 0.01), while the other two studies found no significant difference (24,38).Physical activity was measured using various questionnaire items, such as the number of physical activity days, the number of walking days per week, and the degree of participation in moderate-intensity activities or sedentary activities.Physical activity was queried as one in a set of health behavior questions, or with questionnaires such as the Summary of Diabetes Self-Care Activities (SDSCA) and the International Physical Activity Questionnaire.
Adherence to diabetes medication was measured in two studies (26,34).One study used the Morisky Medication Adherence Scale and found that the intervention group participants showed greater score increases than those in the control group (34).However, the other study, which observed differences in medication adherence using SDSCA measures, did not report any improvement in medication adherence (26).
In three studies (31,33,39), the intervention group showed a significant improvement in diabetes self-efficacy compared to the control group.In a study conducted by Cai and Hu, compared to the control group, the intervention group demonstrated a statistically significant improvement in diabetes management self-efficacy when observed from baseline to 3-month follow-up (41).Similarly, Withidpanyawong et al. (34) investigated diabetes self-efficacy and reported that at 9 months follow-up, self-efficacy scores were significantly higher in the intervention group with a group difference of 0.98 (0.68 vs. 1.67).The study by Hu et al. (39) reported that self-efficacy in the intervention group increased by 33%.
As a result of HbA1c meta-analysis, the heterogeneity between studies was moderately heterogeneous with an I 2 value of 60%.The meta-analysis for six RCT studies showed that the group with caregiver involvement in T2DM education was associated with pooled HbA1c levels 0.83 (95% Confidence interval [CI]: −1.27-−0.38)lower than the control group (p = 0.0003) (Figure 4).Subgroup analyses were performed to identify the potential source of heterogeneity.We checked subgroups of patients' gender, study duration (3 vs. 6 vs. ≥8 months), duration of diabetes (≤6 years vs. >6 years), and types of caregivers (family vs. family and friend).We found no statistically significant differences except types of caregivers.However, a far smaller number of trials and participants contributed data to the family and friend subgroup (One trial, 70 participants) than to the family subgroup (Four trials, 227 participants), meaning that the analysis is unlikely to produce useful findings.The results of the meta-analysis of three types of lipids (LDL, TC, and HDL) showed no strong evidence that caregiver participation in diabetes education improved lipid levels.The p-value for LDL was 0.15 (95%: −13.68-2.07), the p-value for TC was 0.18 (95%: −12.39-2.36),and the p-value for HDL was 0.82 (95%: −1.47-1.85).Heterogeneity (I 2 ) was 0% in all three cases.A meta-analysis was not possible for triglycerides given that of the four studies that reported triglyceride as an indicator, there were only two RCT studies, and one of the RCT studies had multiple intervention groups, making them unsuitable for comparison with other studies.

DISCUSSION
This study aimed to explore the association between caregiver involvement in T2DM education interventions and patient care outcomes.Through this review, we have identified that caregiver involvement in T2DM education reduced HbA1c by 0.83 (95%: −1.27-−0.38)(p = 0.0003).
Our findings regarding the impact of caregiver intervention on the reduction of HbA1c are also supported by a previous meta-analysis study that examined the impact of peer support interventions (44).Thus, although not a causal relationship, this suggests that diabetes education in which caregivers participated improved the patient's physical activity and medication intake and indirectly led to improvement in HbA1c.There was no strong evidence that caregiver participation in diabetes education improved lipid levels.Since only two or three limited studies were used in the lipid meta-analysis, additional research on the effect of the intervention is needed.
Caregivers play key roles in the management of diabetes not only by enhancing patient knowledge of diabetes but also by reinforcing the patients' self-care skills (45,46).The majority of included studies focused on imparting knowledge on T2DM, such as introduction to diabetes or the importance of both healthy dietary options and increasing physical activity.Only one included study, conducted by García-Huidobro et al., focused on skill-building or social support in addition to knowledge (24).However, knowledge is not enough to manage T2DM; rather, self-care skills are more critical to improving glucose control.Therefore, cultivating patient and caregiver skills should be taken into consideration in caregiver involvement in T2DM education.Such skills may include carb counting ability for a diabetes meal plan or monitoring glucose levels using a glucometer.In addition, only two studies focused on counseling or social support (24,40).Given that caregiver burden can affect patient outcomes (47-49), individual counseling or social support should also be included in caregiver education programs for successful T2DM management.
In the included studies, several patient characteristics and performance measurements were inconsistent, thereby possibly contributing to discrepancies in the study results.For example, the care patients' duration of T2DM varied from 3.4 to 10.6 years.As the duration can influence the effectiveness of DM self-management, caregiver education on managing glucose and modifying patient behavior may be less effective for patients with long-standing T2DM (50).Therefore, educational interventions for the caregiver should take into consideration patients' duration of disease and should be individually tailored to be regular, more intense, and reinforced with sustained encouragement.
In addition, family participation has been emphasized in diabetes management, but previous meta-analysis research did not examine in detail the characteristics of the family (14).The present study comprehensively explored different types of caregivers with a scope not limited to the family, and their demographic characteristics.We also investigated whether a tailored intervention takes these caregivers' factors into account during the intervention.The type of caregiver varied from family members, including those living in the same residence, to friends.A family member who mostly stays with the patient in the same place has more opportunities to encourage beneficial patient behavior based on the knowledge they learned from caregiver education (51).Moreover, the majority of caregivers were spouses or adult children.Considering the included studies were conducted in different countries, the expectations and role of caregivers may differ across cultures.For example, adult children of Asian ethnicity are socialized to have a greater sense of filial obligation and caregiving burden where such is a strong cultural norm (52,53).Hence, the participation of caregivers in patients' diabetes education should be tailored to each country's situation and role expectations to improve outcomes.Caring for a loved one with a chronic disease such as T2DM involves a number of challenges, both physical and mental.Therefore, to reduce caregiver burden while achieving optimal outcomes in patient care, caregiver characteristics including demographics, cognitive function, physical condition, and psychological profiles should be considered when developing educational programs.
There are several limitations to this systematic review that are worth noting.First, as we only have six RCT studies in the meta-analysis, we did not analyze the funnel plot.If there are under ten studies eligible for analysis, the Cochrane Handbook advises against using the funnel plot to observe publication bias.The authors conducted a formal search of the gray literature, which may have otherwise helped overcome publication bias.Second, as is the nature of the caregiver-involved intervention, all included studies have a potential performance bias due to incompletely blinded research.Third, there was significant variability in the study outcome measurements, impairing the confidence that can be drawn from the results.Fourth, this study provided only an overview of the effectiveness of various caregiver-involved interventions, with the interventions very briefly summarized.Therefore, clinicians will need to refer directly to cited articles to gain a sufficiently detailed understanding of the interventions studied for potential application.Fifth, as both a pilot study and its associated RCT study targeting the same population were included in the study, the results of the family intervention effect may have been somewhat inflated (35,43).Finally, only six articles with small sample sizes were included in this meta-analysis; therefore, homogeneity power and effect size were limited in this review.Despite the limitations of this study, the results imply that caregivers are not simply supporters and observers but important subjects to be considered along with patients for diabetes education.Therefore, community health-care providers and policymakers should consider involving caregivers at an early stage when educating people with diabetes.In addition, when organizing an education program, it is vital to consider the content for improving caregiver skills so that the caregiver can effectively support the patient's self-management and in turn improve health outcomes.

CONCLUSIONS
This review found that the included studies generally report improvement in most biological and self-management outcomes.In addition, the intervention group in which the caregiver involvement in T2DM education significantly decreased HbA1c compared to the control group from the meta-analysis.The findings of this study suggest that caregiver involvement in education leads to improvement in glycemic control, diabetes knowledge, self-efficacy, and physical activity.Caregiver participation in diabetes self-management education depends on a variety of factors, including caregiver characteristics, lifestyle, and education needs.Therefore, future research should focus on enhancing caregiver participation and incorporating caregiver involvement in T2DM education efficiently and effectively.
Furthermore, no studies were found that investigated the effect of caregiving education on complications and/or hospitalizations.Since many studies focus on limited health indicators, we recommend that the impact of caregiver involvement in T2DM education be investigated in relation to a broader range of health markers with a longitudinal study.

FIGURE 1 .
FIGURE 1. Flow diagram of the literature search according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.

FIGURE 2 .
FIGURE 2. Methodological quality assessment based on the Cochrane Risk of Bias (RoB) tool for randomized controlled trials.Note: +: low risk; -: high risk; ?: unclear risk.

FIGURE 3 .
FIGURE 3. Methodological quality assessment based on the Risk of Bias Assessment tool for Non-randomized Studies (RoBANS) for non-randomized studies.Note: +: low risk; -: high risk; ?: unclear risk.

FIGURE 4 .
FIGURE 4. Meta-analyses on the effect of caregiver involvement in T2DM education on HbA1c.

TABLE 1 .
Search terms Mesh] OR "independent living" OR "community-dwelling" OR "community based multicenter" OR "community health" OR "Home Care Services" [Mesh] OR "Home Care Services, Hospital-Based" [Mesh] OR "Home Care Agenes" [Mesh] OR "Home Nursing" [Mesh] OR "Home Care Agenes" [Mesh] OR "Community Health Services" [Mesh] OR "home care" OR

TABLE 2 .
Characteristics of included studies