Social support and diabetes distress: Does the messenger matter as much as the message?

Introduction

In Canada, about 8–30% of patients with type 2 diabetes (T2D) report feeling distressed and overwhelmed by the relentless demands of diabetes self-management.^1–4 Not surprisingly, rates are even higher in specialty care, reaching 39%. ⁵ Diabetes distress has consistently been associated with elevated HbA1c, worse health outcomes, and poorer self-care behaviors. ⁶ Fortunately, social support can serve as a protective factor; having a large support network and feeling satisfied with support received from healthcare professionals and family have been associated with lower distress.^7–9 However, these studies predominantly focus on primary care and examine overall distress without analyzing the specific diabetes distress subscales: emotional burden and regimen, physician, and interpersonal distress. Therefore, our objectives are: (1) to investigate the impact different sourcesof support (friends/family vs. health care team) has on each diabetes distress dimension among adults with T2D presenting to a specialty care setting and (2) to examine the sociodemographic correlates of the diabetes distress subscales.

Methods

This cross-sectional study was approved by UBC's clinical research ethics board and utilizes baseline data from a larger study called Clinic to Community (C2C). Interested adults with T2D presenting to specialty care in Vancouver (n = 196) provided informed consent and completed a survey examining diabetes distress, perceived social support, and socio-demographic background. For the complete protocol, please refer to the C2C study. ¹⁰

Diabetes distress was assessed using the Diabetes Distress Scale (DDS), a 17-item scale with four subscales: physician-related distress, diabetes-related interpersonal distress, regimen-related distress, and emotional burden. ¹¹ Higher DDS scores indicate greater levels of diabetes distress (Cronbach's α = 0.93, r = 0.56). Social support was measured by four items assessing level of and satisfaction with received support from either health care professionals (HCP) or friends/family. ¹²

Demographic variables, shown in Table 1, were grouped due to low counts in individual levels. Items were averaged for each diabetes distress subscales and mean subscale scores <3.0 were classified as normal distress and ≥3.0 as high levels of distress. ¹¹ The two social support items (family and friends support, HCP support) were also averaged and mean scores ≥4.0 were classified as “good” and <4.0 as “inadequate”.

OPEN IN VIEWER

Table 1.

Demographic characteristics of participants (n = 196).

	N
Age (years), mean ± SD	195	59.5 ± 11.2
Age of diagnosis (years), mean ± SD	193	47.8 ± 11.5
Female, n (%)	196	98 (50%)
Born in Canada	196	103 (53%)
Marital status, n (%)
Single	195	48 (25%)
Married/co-living	195	98 (50%)
Separated/divorced/widowed	195	49 (25%)
Ethnicity, n (%)
Caucasian	196	101 (52%)
East Asian	196	30 (15%)
Indigenous	196	9 (5%)
South Asian	196	21 (11%)
Southeast Asian	196	7 (4%)
Others	196	28 (14%)
Completed post-secondary education	194	112 (58%)
Household income <$70,000	175	125 (71%)
Employed	195	87 (45%)
Experience Distressa, n (%)
Emotional	196	72 (37%)
Physician-related	196	27 (14%)
Regimen	196	81 (41%)
Interpersonal	196	46 (23%)
HbA1c, mean ± SD	196	8.1 ± 1.7
Elevated HbA1cb, n (%)	196	148 (75%)
Received Adequate Supportc, n (%)
Friends/Family	196	90 (46%)
Healthcare	196	116 (59%)

Note. ^aDDS score ≥ 3. ^bHbA1c ≥7%. ^csupport score ≥ 4.

Descriptive statistics were calculated as means and standard deviations for continuous variables and counts and percentages for categorical variables. Multiple logistic regression models were performed to examine the relationship between each distress outcome and the social support and sociodemographic predictors. All analyses were conducted with SAS software, version 9.4 (Cary, NC, USA).

Results

The mean age of our sample was 59.5 years; Half (n = 98) were women (Table 1). On average, individuals were diagnosed with T2D around 47.8 years of age and had a mean HbA1c of 8.1 ± 1.7. About 53% were born in Canada (n = 103), 52% were Caucasian (n = 101); 58% (n = 112) had completed some form of post-secondary education; 71% (n = 125) reported an annual household income of <$70,000 CAD. Some participants reported moderate to high levels of emotional distress (37%, n = 72), physician distress (14%, n = 27), regimen distress (41%, n = 81), and/or interpersonal distress (23%, n = 46). About half felt that they received adequate support from their friends/family (46%, n = 90) and healthcare providers (59%, n = 116) respectively.

Social support and diabetes distress

Having greater friends/family support emerged as the only predictor of interpersonal distress (71% lower odds), while greater health care support emerged as the predictor for physician distress (84% lower odds) and regimen-related distress (62% lower odds). After controlling for the demographic variables, these relationships remained significant (Table 2).

OPEN IN VIEWER

Table 2.

Support-related and sociodemographic predictors of four diabetes distress subscales (n = 173).

Variable	Emotional Distress		Physician Distress		Regimen Distress		Interpersonal Distress
	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Friends/Family	0.743	0.331, 1.668	1.156	0.277, 4.817	1.180	0.514, 2.709	0.295*	0.103, 0.846
Healthcare	0.675	0.315, 1.445	0.160**	0.043, 0.593	0.382*	0.172, 0.848	0.975	0.387, 2.458
Gender	1.054	0.528, 2.106	1.057	0.352, 3.179	0.747	0.364, 1.535	0.757	0.325, 1.762
Birth country	1.368	0.603, 3.100	3.086	0.830, 11.476	1.734	0.750, 4.009	1.625	0.603, 4.379
Marital status (a)	0.866	0.329, 2.282	1.950	0.422, 9.011	0.344*	0.124, 0.955	0.910	0.285, 2.900
Marital status (b)	1.569	0.607, 4.055	2.499	0.562, 11.115	1.128	0.424, 3.001	2.190	0.730, 6.575
Ethnicity	0.910	0.407, 2.033	0.919	0.250, 3.375	0.839	0.366, 1.927	0.676	0.257, 1.782
Education	0.550	0.280, 1.081	0.718	0.250, 2.064	0.966	0.477, 1.956	0.455	0.201, 1.026
Income	1.017	0.424, 2.436	0.093*	0.010, 0.852	1.244	0.500, 3.093	0.465	0.141, 1.538
Employment	0.947	0.447, 2.007	2.123	0.673, 6.697	1.450	0.666, 3.155	1.629	0.658, 4.036
Age of diagnosis	0.981	0.948, 1.016	1.041	0.978, 1.108	1.003	0.967, 1.040	0.997	0.954, 1.043
Age	0.981	0.943, 1.021	0.962	0.900, 1.028	0.943**	0.903, 0.985	0.962	0.915, 1.011

Note. Independent variables include friends/family and healthcare support (0 = poor, 1 = good), gender (0 = female, 1 = male), birth country (0 = Canada, 1 = outside Canada), marital status (a) 0 = single, 1 = married/co-living and (b) 0 = single, 1 = separated/divorced/widowed, ethnicity (0 = other, 1 = white), education (0 = college-, 1 = bachelor+), income (0 = 70000-, 1 = 70000+), employment (0 = unemployed, 1 = employed). High distress considered as DDS score ≥3, good support considered as ≥4. Odds Ratio OR.

*p < 0.05, **p < 0.01, ***p < 0.001.

Discussion

Previous research has focused, primarily, on the impact of social support on overall distress levels.^7,8 This study elucidates this relationship further by examining the specific source of support and nature of distress. Specifically, we found that higher levels of social support delivered by the health care providers (HCP) were associated with lower physician- and regimen-related distress, whereas higher levels of friends/family support were associated with lower interpersonal distress.

Our finding linking HCP support to lower distress around self-management demands is consistent with Cummings et al.'s study of a combined nurse and behavioral provider intervention. ¹³ Among 139 adults with T2D and high distress levels, the authors found that participation in the healthcare team intervention led to reductions in regimen-related distress. Similarly, patients randomly assigned to a family-focused intervention involving patient-family member pairs reported greater reductions in interpersonal distress compared to a wait-list control condition. ¹⁴ Clearly, precision support models matching the type of intervention(ist) to patients experiencing a specific aspect of distress warrant further investigation.

That neither family/friends or HCP support were related to emotional distress is not surprising. The worries, exhaustion, anger, and hopelessness carried by patients is a unique experience that friends, family, and providers cannot fully identify with. In fact, the only individuals who have a visceral understanding of this emotional burden are other adults living with diabetes, i.e., peers. In fact, peer support interventions have been found to reduce diabetes distress and improve glycemic control. ¹⁵ Future studies should examine whether peer support significantly improves emotional burden compared to other subscales.

This study was not without limitations. We used a cross-sectional design, which prevented us from establishing causal relationships. Our social support measure consisted of four questions and did not capture different support types or “messages” (e.g. tangible, informational, emotional). Future research should include more comprehensive support measures and examine the interaction between support sources and types of distress.

Diabetes distress is reported by many adults with T2D and interferes with self-management. Our results show that while social support can protect against distress, the type of support provider, or “messenger”, and the nature of distress patients experience can play a critical role in optimizing psychosocial outcomes. Ultimately, more personalized interventions targeting types of diabetes distress will better meet the unique needs of each patient.