Prevalence of diabetes distress and its relationship with self-management in patients with type 2 diabetes mellitus.

CONTEXT: Diabetes distress (DD) in India has been studied mainly in the context of depression. Little is known about DD, its determinants, distribution, and its impact. AIMS: This study aims to estimate the prevalence of DD and identify its socio-demographic and clinical determinants in type 2 diabetes mellitus patients. To assess the relationship of DD with self-management in nondepressed type 2 diabetes mellitus. SETTINGS AND
DESIGN: A cross-sectional descriptive study in noncommunicable disease clinic of a tertiary care medical center. SUBJECTS AND METHODS: DD Scale was used to assess DD in 200 patients with type 2 diabetes mellitus. Diabetes Self-Management Questionnaire was used to evaluate self-management behavior. Patient Health Questionnaire 9 was used to exclude depression. STATISTICAL ANALYSIS USED: Sample size calculation was done as per prevailing prevalence estimates. SPSS 20.0 was used for statistical analysis. ANOVA and Independent t-tests were done to compare between groups means. Hierarchical multiple regression analysis was done, keeping self-management as a dependent variable and socio-demographic, clinical variables, and DD as independent variables.
RESULTS: The prevalence of DD was 42% in our sample. The duration of diabetes showed that a significant association with DD. DD was significantly and negatively associated with all four domains of self-management, while statistical significance was reached in three domains: dietary control (β = -0.378, P < 0.01); glucose management (β = -0.181, P < 0.01); and healthcare use (β = -0.244, P < 0.01).
CONCLUSIONS: DD is widespread in our patients and harms self-management. There is a need to address DD to improve disease-specific outcomes. Copyright:

The concept of diabetes distress (DD) has origins in research on stress and coping and is relatively recent.[1] Diabetes is a challenging and chronic disease with the potential for irreversible complications, requiring optimum blood glucose levels. Emotional distress associated with a need for self-management in diet control, regular physical exercise, monitoring blood glucose levels often takes a toll on the affected individual. Managing self often overwhelms the individual coping capacity and manifests as anger, irritability, sadness, and a sense of doom. This diabetes-specific emotional stress is termed DD and is understood as a continuous variable.[2] It is independent of clinical depression yet often overlaps both subsyndromal and syndromal depression.

Depression, on the other hand, is considered a dichotomous variable, as present or not. Co-occurrence of DD often makes the diagnosis of depression difficult as a clinical manifestation of DD mimics depression. DD often results in high false positivity on depression screening instruments because emotional distress is similar to the affective experience of depression.[3] Frequently, screen-positive depression turns out to be DD and not clinical depression when assessed on gold standard clinical assessment instruments.[4]

DD in type 2 diabetes mellitus harms self-management more than reported in depression. Poor self-management is often associated with poor glycemic control. DD is known to have a negative association with physical activity and medication adherence.[5] Furthermore, the presence of DD in depression is understood to have a mediating role in glycemic control. While treatment of depression has conflicting results on glycemic control and self-management, adequate DD management has often positively impacted self-management and glycemic control.[67] The identification of DD, hence has therapeutic potential owing to its impact on patient outcomes.[3]

Existing literature is replete with determinants, distribution, and impact of depression on diabetes outcomes, DD; however, it has not been addressed adequately. DD is more common than depression, with prevalence rates as high as 50% across the continents.[1] Only a few studies across various Indian centers have addressed DD with varying prevalence estimates. Its relation with self-management has not been reported to date.

Here, we discuss the prevalence, distribution, and determinants of DD in nondepressed patients with type 2 diabetes mellitus. We also study the relationship of DD with self-management behaviors using validated self-report instruments.

SUBJECTS AND METHODS

Study site and participants

Cross-sectional study design with convenience sampling was used. Patients meeting American Diabetes Association criteria for type 2 diabetes mellitus seeking treatment at the Noncommunicable disease clinic of a tertiary care center were enrolled in the study. Adults with type 2 diabetes mellitus and taking oral hypoglycemic agents were included. Patients with comorbid severe psychiatric illness such as depression, psychosis, or substance use disorders were excluded. Tobacco dependence was not excluded. Furthermore, patients having chronic medical illness unrelated to diabetes-like cancers, interstitial lung disease, and coronary artery disease were excluded. Patients having experienced significant life events such as divorce, financial loss, losing a job, and a close family member's death within the last 6 months were not included in the study.

Study instruments

Semi-structured socio-demographic and clinical questionnaire, PHQ-9, DD Scale (DDS), Diabetes Self-Management Questionnaire.

Semi-structured pro forma included socio-demographic details, duration of diabetes, anthropometric measures like body mass index (BMI) and biochemical profile including fasting blood sugar and glycated hemoglobin (HbA1C).

Patient Health Questionnaire

We used the Odiya version of Patient Health Questionnaire-9 (PHQ9) screening depression. PHQ9 is a validated screening instrument of depression in primary care and general medical settings with robust psychometric properties. A score of >5 indicates mild depression, >10 means moderate depression, and >10 indicates severe depression.[8]

Diabetes Distress scale

To specifically evaluate diabetes-specific distress, problem area in diabetes was developed. However, it did not address distress associated with the physician or regimen. DDS was designed to incorporate these additional elements and measures in four domains. It is more reflective of physician-related distress and problems concerning diabetes self-management. DDS has 17 items in four domains: Physician associated distress, emotional burden, interpersonal stress, and regimen-related stress.[9] The items are scored on a six-point scale: From “not a problem” to “very serious problem.” A score above three on DDS indicates a high degree of stress.[10]

Diabetes self-management questionnaire

This is a 16-item questionnaire with five subscales, “Glucose Management,” Dietary Control, Physical Activity, Health Care Use and a “Sum Scale” to assess self-care activities associated with glycemic control.[11] It has high overall internal consistency (Cronbach's alpha = 0.84) and subscale consistency (Glucose Management': 0.77; Dietary Control: 0.77; Physical Activity: 0.76; Health Care Use: 0.60) and correlation with HbA1C levels.

Sample size calculation for the study was done with the formula: Keeping the prevalence of DD at 50% and admissible error at 15%, the sample size came to 176.[1] Expecting 10% data loss due to incomplete or erroneous entry, we increased the total sample size to 194 and then rounded it to 200 participants.

Ethical considerations

Written informed consent was be taken from participants, Institutional Ethics approval was taken.

Statistical analysis

We double-checked the data and entered it into MS excel. We used SPSS version 20.0(Chicago, SPSS Inc.) for analysis. The normality of data was tested using Kolmogorov–Smirnov test. Means and standard deviations were calculated for continuous variables. Bivariate association between continuous variables was tested using independent t-test and ANOVA, keeping statistical significance at 0.05. Stepwise hierarchical regression analysis was carried out, maintaining diabetes self-management as a dependent variable and socio-demographic variables, clinical characteristics, and DD as independent variables. Standardized β coefficients and P values were calculated to evaluate the relationship between independent and dependent variables.

RESULTS

A total of 200 participants were included in the study; the mean age was 51.34 ± 11.16 years (27–79), 63.5% were male. A detailed socio-demographic and clinical profile is described in Table 1. Based on total DDS scores, 35.5% experienced moderate distress, and 41.5% reported high distress levels. Out of 200 patients, 83 were positive on DDS; hence, the prevalence of DD in our population was 42%. Regimen-related distress was present in 70%, interpersonal distress in 63%, emotional distress in 45%, and physician distress in 27%. The DD scores in different domains are illustrated in Table 2. The highest mean distress score was seen in regimen-related distress.

Table 1

Sociodemographic and clinical parameters of the study population

Characteristics	Frequency (%)
Gender
Male	127 (63.5)
Female	73 (36.5)
Age (years)
<45	56 (28)
45-60	98 (49)
>60	46 (23)
Marital status
Ever married	180 (90)
Unmarried	20 (10)
Education
Primary	54 (27)
Secondary	87 (43.5)
Tertiary	59 (29.5)
Income
Below poverty line	131 (65.5)
Above poverty line	69 (34.5)
Family type
Nuclear	123 (61.5)
Joint/extended	77 (38.5)
Residence
Urban	57 (28.5)
Rural	143 (71.5)
Lifestyle
Sedentary	63 (31.5)
Nonsedentary	137 (68.5)
Duration of diabetes (years)
<5	101 (50.5)
5-10	69 (34.5)
>10	30 (15)
BMI
<25	138 (69)
25-30	54 (27)
>30	8 (4)
DDS severity
No distress	46 (23)
Moderate distress (DDS 2-2.9)	71 (35.5)
High distress (DDS ≥3)	83 (41.5)

BMI – Body mass index; DDS – Diabetes Distress Scale

Table 2

Diabetes distress domain scores of the study participants

Domains	Minimum score	Maximum score	Mean±SD
Emotional burden	1	4.4	2.62±0.97
Physician-related distress	1	4.25	1.91±0.98
Regimen-related distress	1	4.6	3.19±0.93
Interpersonal distress	1	5	3.09±1.59
Total score	1	4.29	2.70±0.83

SD – Standard deviation

Association DDS in the four domains and total DDS scores with socio-demographic variables and clinical variables were carried out using ANOVA and independent t-tests. DDS was high in <5 years of duration of diabetes and also in patients with more than 10 years duration of diabetes. DD was high in all four domains; however, statistical significance was reached in the interpersonal burden domain [Table 3].

Table 3

Bivariate association between sociodemographic and clinical variables and domains of Diabetes Distress Scale

Characteristic	Mean±SD
	Emotional burden	Physician related burden	Regimen related burden	Inter-personal burden	Total distress
Gender
Male	2.61±0.999	2±1.001	3.26±0.835	3.19±1.577	2.76±0.83
Female	2.61±0.927	1.75±0.913	3.07±1.077	2.91±1.622	2.598±0.83
P	0.98	0.16	0.16	0.244	0.18
Age
<45	2.78±0.93	2.04±0.99	3.26±0.89	3.04±1.64	2.79±0.83
45-60	2.62±0.98	1.85±0.97	3.13±0.96	3.09±1.56	2.67±0.84
>60	2.43±0.99	1.87±0.96	3.22±0.93	3.14±1.64	2.66±0.84
P	0.185	0.46	0.71	0.95	0.63
Marital status
Ever married	2.58±0.97	1.89±0.96	3.18±0.918	3.057±1.604	2.68±0.827
Unmarried	2.94±0.95	2.01±1.09	3.27±1.076	3.40±1.516	2.90±0.896
P	0.12	0.08	0.69	0.36	0.27
Education
Primary	2.77±0.93	2.04±0.97	3.3±0.88	3.16±1.63	2.83±0.84
Secondary	2.50±0.89	1.81±0.94	3.12±0.97	2.86±1.59	2.59±0.77
Tertiary	2.64±1.09	1.92±1.04	3.18±0.93	3.37±1.54	2.76±0.89
P	0.19	0.40	0.55	0.15	0.289
Income
Above poverty line	2.50±1.04	1.77±0.99	3.09±1.05	3.09±1.59	2.61±0.908
Below poverty line	2.68±0.93	1.98±0.97	3.24±0.861	3.08±1.60	2.75±0.79
P	0.22	0.15	0.31	0.97	0.25
Family type
Nuclear	2.65±0.95	1.71±0.93	3.11±0.97	3.17±1.57	2.73±0.86
Joint/extended	2.57±1.01	2.03±0.99	3.31±0.87	2.97±1.63	2.66±0.80
P	0.565	0.168	0.13	0.211	0.58
Residence
Urban	2.70±0.94	1.95±0.97	3.26±0.85	3.11±1.58	2.76±0.81
Rural	2.41±1.01	1.81±0.98	3.00±1.09	3.04±1.64	2.55±0.89
P	0.448	0.85	0.08	0.394	0.475
Lifestyle
Sedentary	2.62±0.93	1.88±0.96	3.14±0.98	2.99±1.59	2.67±0.84
Nonsedentary	2.60±1.06	1.96±1.01	3.31±0.81	3.30±1.59	2.79±0.83
P	0.096	1.222	0.23	0.494	0.504
Duration of diabetes (years)
<5	2.66±0.97	1.90±0.88	3.17±0.96	2.81±1.62	2.66±0.83
5-10	2.50±0.96	1.89±0.99	3.18±0.89	3.31±1.55	2.70±0.81
>10	2.73±0.97	1.96±0.94	3.26±0.90	3.49±1.48	2.84±0.88
P	0.44	0.93	0.89	0.04*	0.56
BMI
<25	2.59±0.94	1.91±0.98	3.17±0.96	3.04±1.58	2.68±0.82
25-30	2.75±1.03	1.99±0.99	3.23±0.89	3.18±1.63	2.79±0.87
>30	2.24±1.10	1.41±0.64	3.29±0.90	3.22±1.77	2.53±0.82
P	0.30	0.26	0.86	0.85	0.6

*Statistically significant. SD – Standard deviation; BMI – Body mass index

We carried out a correlation analysis between DDS and DSMQ scores. Pearson correlation coefficient was modest in 0.4 between regimen-related distress and dietary control and regimen-related distress and total scores on DSMQ, both of which reached statistical significance [Table 4].

Table 4

Correlation between domains of diabetes self-management and domains of diabetes distress

DSMQ	DDS, r (P)
	Emotional burden	Physician related distress	Physician related distress	Interpersonal distress	Total scores
Dietary control	−0.211** (0.003)	−0.318** (0.000)	−0.418** (0.000)	−0.288** (0.000)	−0.394** (0.000)
Glucose management	−0.131 (0.06)	−0.185** (0.009)	−0.262** (0.000)	−0.38 (0.594)	−0.195** (0.006)
Physical activity	−0.06 (0.4)	0.06 (0.40)	−0.158* (0.026)	−0.046 (0.517)	−0.040 (0.572)
Health care	−0.136 (0.056)	−0.173* (0.014)	−0.294** (0.000)	−0.205** (0.004)	−0.260** (0.000)
Total scores	−0.209** (0.003)	−0.259** (0.000)	−0.404** (0.000)	−0.183** (0.009)	0.338** (0.000)

* (P < 0.05, Statistically significant), **(P < 0.01, statistically highly significant). DDS – Diabetes Distress Scale; DSMQ – Diabetes self-management questionnaire

Multiple hierarchical regression analysis was carried out using SPSS 20.0, keeping self-management behaviors of dietary control, glucose management, physical activity, and health care use as independent variables. Age, gender, and education were entered as independent variables in Step 1, BMI category, and diabetes category in Step 2, and total DDS scores were entered in Step 3. We found a statistically significant positive relationship between age with dietary control and glucose management. There was no evidence of multicollinearity among independent variables in any analysis. The relationship of gender with dietary control, glucose management, and health use were also positive and reached statistical significance. Total DDS scores were negatively associated with all four self-management domains and reached statistical significance in dietary control, glucose management, and health care [Table 5].

Table 5

Standardized regression coefficients for diabetes distress on self- management on hierarchical multiple regression analysis

	Dietary control	Glucose management	Physical activity	Health care use
Step 1
Age	0.228**	0.300**	−0.011	0.022
Gender	0.246**	0.165*	−0.133	0.163*
Education	−0.011	0.141	0.281**	0.120
Step 2
BMI category	−0.026	0.070	0.047	−0.028
Diabetes duration category	0.070	0.135	0.042	−0.035
Step 3
DDS total	−0.378**	−0.181**	−0.049	−0.244**

*P<0.05, **P<0.01. DDS – Diabetes Distress Scale; BMI – Body mass index

DISCUSSION

The diminished psychological well-being which DD brings in is often due to psychosocial problems. DD ushers in a state of fatalism wherein the affected individual think and feel he is doomed and suffers from an incurable disease and refrains from diet-control, exercise, and medication adherence.

Literature on DD is sparse from India; only a handful of studies have addressed this critical disease determining variable. Gahlan et al., in their cross-sectional study, have reported a prevalence rate of 18% from a north Indian tertiary care specialty set up. The highest prevalence was emotional distress, and DD was associated with low education and microvascular complications.[12] Roy et al. had studied DD in the context of depression in east India, wherein 25% had DD.[13] Kumar et al. from south India reported prevalence rates of 42% in nondepressed patients with diabetes in a specialty endocrinology setting.

Our population had a 42% prevalence of DD, similar to reported rates from south India.[14] Regimen related distress was the most common among all domains of DD, contrary to existing Indian reports. Regimen-related distress also had the highest mean score among all four domains of DDS.[1214] High regimen distress is associated with poor medication adherence, and interventions reducing regimen distress have documented self-management and glycemic control improvement. While studies have been done to assess the relationship of DD with medication adherence, none of the studies has systematically addressed self-management using a validated questionnaire. Diabetes management is much more than medication adherence and includes dietary control, physical exercise, glucose regulation, and regular healthcare use.[1415]

Low education and complication of diabetes mellitus (microvascular and macrovascular) were associated with DD.[13] In our sample, DD was associated with the duration of diabetes. High DD was seen in recent diagnosis (within 5 years) and >10 years' duration of diabetes. High stress levels in the early years of diagnosis could be because of an emotional reaction to an unanticipated diagnosis. In contrast, high distress in patients of long-standing diabetes could be because of complications setting in and the realization of irreversible nature and the need for continuous self-management.[16]

After controlling for socio-demographic and clinical variables, total DDS were the highest self-management predictors in all four domains. Our data highlight the intricate relationship of DD with self-management behavior. The data being cross-sectional causal association cannot be ascertained. However, it can be safely assumed that high DD levels are associated with poor self-management.

Emotional distress is known to be associated with poor diabetes self-management and glycemic control in type 1 diabetes.[17] In type 2 diabetes, both depression and DD are known to be related to poor self-management. Interventions have been suggested addressing DD independently and as part of clinical depression in improving diabetes management.[18]

Existing literature recommends addressing DD through various psychological interventions. Varying from simple monitoring to specialized psychological interventions, these are very effective in improving DD. Improvement in DD is known to enhance self-management.[18] Hence, it is imperative to screen for DD for optimizing self-management.

Strengths and limitation

We did not have HbA1C levels of the entire study sample, which limited our understanding of DD on glycemic control. However, all domains of self-management were adequately addressed using a validated questionnaire.

CONCLUSIONS

DD rates are very high in our nondepressed population with type 2 diabetes mellitus. DD is associated with poor self-management. Diagnosis of DD should be made in routine diabetes care, and the risk factors for DD need to be identified. DD is amenable to psychological interventions. Hence, there is a need to address DD to improve self-management of diabetes. Reducing DD has the potential to improve overall clinical outcomes in patients with diabetes mellitus.

Financial support and sponsorship

ICMR.

Conflicts of interest

There are no conflicts of interest.