1. Background

Diabetes mellitus is a common lifestyle-related disease in the old adult population in Korea, with a prevalence of approximately 16.7% in the population aged 30 years and older and over 30.1% in the population aged 65 years and older [1]. Diabetes, if not properly managed, can cause various complications; therefore, active lifestyle changes and appropriate drug treatment should be maintained from the early stages of the diagnosis [2]. In particular, to maintain an appropriate blood sugar level, self-management, including diet, is required throughout one's life. In this process, patients with diabetes experience a lot of stress related to the disease and its management [3]. In the case of old adult patients with diabetes, diabetes self-management is more difficult because they have complex diseases and geriatric syndromes such as sleep disturbance, depression, cognitive impairment, and physical dysfunction [4, 5]. Additionally, in a previous study that studied the intensive glucose lowering in type 2 diabetes, the intensive therapy group (targeting a glycated hemoglobin level below 6.0%) had more unexplained mortality than the standard group (targeting a level from 7.0 to 7.9%) [6]. Strict control of blood sugar has been reported to increase not only the risk of hypoglycemia, but also increase the risk of death [6], making the management of diabetes in old adult patients more difficult. When diabetes was strictly managed, the drug or insulin dose was changed several times [6]. Adaptation to the changing drug therapy, including a tight diet, may be more difficult for older diabetes patients than for younger diabetes patients. Therefore, older diabetes patients are expected to have a high level of diabetes distress. Diabetes distress must be managed because it is a factor that reduces their quality of life [7].

People diagnosed with diabetes face great disappointment as they have to manage themselves for the rest of their lives, and they also have adjustment difficulties [8]. Additionally, it is difficult to inform others about having diabetes as they are aware of the social stigma around diabetes that it is caused by a poor lifestyle, and they feel ashamed of having diabetes [8]. In particular, they do not understand the need to completely change their lifestyle until they are diagnosed with diabetes and have a diabetes management diet [9]. Additionally, there are cases where people think that diabetes or its complications are caused by their own mistakes or karma, and they perceive it negatively [9]. As such, people diagnosed with diabetes feel stressed out due to various factors, and when these negative emotions are internalized, self-stigma occurs.

In diabetes self-stigma, people with diabetes devalue themselves and have negative emotions regarding their disease [10]. Diabetes self-stigma reduces self-efficacy related to diabetes management, and causes a decrease in self-management behavior [11]. Additionally, since self-stigma is a factor that lowers diabetes patients' quality of life due to depression or social isolation [10], it is necessary to manage self-stigma in diabetes patients so that they do not develop mental health conditions.

In fact, in a qualitative study of diabetes patients' life experiences, those diagnosed with diabetes for the first time considered whether they should accept that they had diabetes in the early stages and were depressed [12]. They also denied management methods to treat diabetes while worrying about an unhealthy future [12]. However, once they accepted the diagnosis, they revealed their diabetes diagnosis to those around them, found treatment methods that suited them, and created a new self-identity, turning their weakness into opportunity [8]. This process is called acceptance of the disease; acceptance means actively accepting thoughts, emotions, and physical sensations with non-judgmental self-perception [13]. Previous studies have shown that acceptance and commitment therapy reduces emotional distress in older adults with type 2 diabetes [14]. Based on this, it is expected that a patient's acceptance of diabetes will play a mediating role in diabetes distress among the old adults, thereby reducing self-stigma due to disease or disease management. However, studies on the factors affecting self-stigma in diabetes patients and the role of acceptance behavior in managing the stress affecting self-stigma are insufficient.

This study identifies the role of acceptance action in the relationship between diabetes stress and self-stigma and provides fundamental data for establishing strategies for reducing self-stigma in diabetes patients by examining the relationship between stress, self-stigma, and acceptance action in older diabetes.

2. Purpose

This study aimed to explore the mediating effects of acceptance action between distress and self-stigma among old adults with diabetes, and it had the following goals. First, the degree of acceptance diabetes distress, self-stigma, and acceptance action in older adults with diabetes were identified. Second, the differences in diabetes distress, self-stigma, and acceptance action according to general characteristics were identified. Third, the correlation between diabetes distress, self-stigma, and acceptance action was investigated. Fourth, the mediating effect of acceptance action between diabetes distress and self-stigma was investigated.

1. Study Design

This secondary analysis study used the data collected during the "Development of a tool for measuring the health equilibrium of diabetes patients" (JIRB-2020101201-02-201101), and explored the mediating effects of acceptance actions on diabetes distress and self-stigma.

2. Participants

The participants of this study were analyzed using data from 187 diabetes patients aged 55 years or older among the data of 4,800 panel. The inclusion criteria were as follows: 1) adults aged 55 years or older diagnosed with diabetes by a doctor, 2) those who could understand and respond to the questionnaire, and 3) those who voluntarily agreed to the participate and subsequently completed the questionnaire.

3. Measures

4. Data Collection

The data for this study were collected from 26 to 31 March 2020, through a survey institution. A self-report questionnaire consisting of general characteristics, diabetes distress, self-stigma, and acceptance action questionnaires was prepared as an online questionnaire and distributed to 4,800 panelists of specialized survey institutions. The questionnaire was administered to those who answered that they had diabetes in the questions related to current health problems and diseases diagnosed by a doctor among 15 chronic diseases. The questionnaire was set to automatically end for participants who did not select diabetes for either of the two questions. Next, five questions about diabetes-related characteristics were asked, and after the responses to diabetes-related questions were completed, the remaining questions were presented.

The analysis in this study was conducted using the above raw data after obtaining approval from the Institution Review Board of Joongbu University (JIRB-2021050302-01).

5. Statistical Analysis

The collected data were analyzed using IBM SPSS/WIN 24.0. Participants' general characteristics, diabetes distress, the degree of self-stigma, and acceptance action were analyzed using descriptive statistics. The differences in diabetes distress, self-stigma, and acceptance action according to the participants' general characteristics were analyzed using independent t-test and one-way analysis of variance (ANOVA); the Scheffé test was used as a post-hoc test. The correlation between participants' diabetes distress, self-stigma, and acceptance action was analyzed using Pearson's correlation coefficients. The mediating effect of acceptance action on the relationship between diabetes distress and self-stigma of participants was analyzed using a 3-step regression analysis as presented by Baron and Kenny [20]. In steps 1 and 2, regression analyses are performed on the influence of the independent variable on the parameter, and on the influence of the independent variable on the dependent variable to check whether each has a significant effect. In step 3, the influence on the dependent variable is checked by inputting the independent variable and the parameter simultaneously. In step 3, the influence of the independent variable is reduced as compared to step 2 to check if the parameter has a significant influence on the dependent variable to have a mediating effect; if the influence of the independent variable on the dependent variable is significant, there is a partial mediating effect, and if it is not significant, it is judged that there is a full mediating effect. The significance of the mediating effect was evaluated by bootstrapping in the PROCESS macro program. The number of bootstrap samples was set at 1,000, with a 95% confidence interval. If the confidence interval does not include 0, the mediating effect is considered significant.

1. Participants' Characteristics

The general characteristics of the participants are listed in Table 1. The participants of this study were 21.9% female and 78.1% male. Their mean age was 62.53±5.20 years, and the age range was 55~79 years. Regarding their educational level, 58.8% had the university level or higher education, and 86.6% had a spouse. Those who had a job accounted for 63.6%, and 48.7% answered that they practiced a religion. Regarding social activities, 46.5% answered that they engaged in social activities less than once a month, and only 19.3% answered that their perceived health was good. The average duration of diabetes was 9.81±8.03 years, and the prevalence ranged from 1~41 years. As for the type of hospital that they were being treated for diabetes, clinics accounted for the majority (56.1%). Of all, 73.3% were prescribed oral medications. Those with comorbidities accounted for 80.2%. The mean of diabetes distress was 2.98±0.64, and that of self-stigma was 2.54±0.74. The mean of acceptance action was 4.16±0.35.

Table 1 Participants' General Characteristics (N=187) Click for larger image

2. Difference of Independent Variables according to General Characteristics

Table 2 presents the results of the analysis of the differences in diabetes distress, self-stigma, and acceptance action according to the general characteristics of the participants. The differences in diabetes distress according to the general characteristics of participants were gender (t=-3.18, p=.002), perceived health status (F=7.44, p=.001), duration of diabetes (F=3.79, p=.024), type of hospital being treated at (F=4.44, p=.005), and the treatment method (F=6.91, p<.001).

Table 2 Differences in Independent Variables according to General Characteristics (N=187) Click for larger image

The differences in self-stigma according to the general characteristics of participants were perceived health status (F=9.40, p<.001), type of hospital being treated at (F=3.72, p=.013), and the treatment method (F=10.56, p<.001).

The differences in acceptance action according to the general characteristics of participants were perceived health status (F=8.97, p<.001) and the treatment method (F=4.21, p=.007).

3. Correlations among Study Variables

As shown in Table 3, diabetes distress was positively correlated with self-stigma (r=.71, p<.001) and negatively correlated with acceptance action (r=-.48, p<.001). That is, the higher the diabetes distress, the higher the self-stigma and the lower the acceptance action. Self-stigma also had a negative correlation with acceptance action (r=-.43, p<.001), indicating that higher the self-stigma, the lower the acceptance action.

Table 3 Correlation among Study Variables (N=187) Click for larger image

4. Mediating Effect of Acceptance Action in the Relationship between Diabetes Distress and Self-stigma

To confirm the mediating effect of acceptance action on the relationship between diabetes distress and self-stigma in patients with diabetes, a 3-step regression analysis was performed according to the procedure suggested by Baron & Kenny [17] (Table 4). The Durbin-Watson value was checked to verify the independence of the residuals before verifying the mediating effect, and was found to be 1.81 to 2.04, which is close to 2, indicating that the dependent variable had no autocorrelation. In addition, as a result of confirming multicollinearity, the tolerance limit was 0.774, all below 1.0, and the variance expansion coefficient for all variables was 1.00 to 1.29, not greater than 10, so there was no multicollinearity, confirming that the model was suitable for regression analysis.

Table 4 Mediating Effects of Acceptance Action on the Relationship between Diabetes Distress and Self-stigma (N=187) Click for larger image

The result of the regression analysis in step 1 confirmed that diabetes distress, an independent variable, had a statistically significant effect on acceptance action as a parameter (β=-.48, p<.001). In the 2nd step of regression analysis, the effect of diabetes distress, an independent variable, on self-stigma, a dependent variable, was confirmed to be statistically significant (β=.71, p<.001). In the last step, regression analysis was performed by inputting the independent variable diabetes distress and the parameter acceptance action as independent variables to analyze the effect of the parameter acceptance action on the dependent variable self-stigma. Both diabetes distress (β=.65, p<.001) and acceptance action (β=-.12, p=.044) were found to affect self-stigma. In step 3, both the independent variables, diabetes distress and the parameter acceptance action, were significant, and the coefficient of diabetes distress, the independent variable, was smaller in step 3 than in step 2, so the acceptance action partially mediated the relationship between diabetes distress and self-stigma. That is, diabetes distress had a direct effect on self-stigma, and diabetes distress had an indirect effect on self-stigma by changing acceptance action (Figure 1). The explanatory power of the regression analysis in this study was confirmed to be 51.0%. As a result of the significance of the indirect mediating effect using bootstrapping, the effect size was 0.06, the lower limit confidence interval was -0.01, and the upper limit confidence interval was -0.13, which was significant as it did not include 0 (Table 5).

Figure 1 Mediating effects of acceptance action on the relationship between diabetes distress and self-stigma. Click for larger image

Table 5 Statistical Significance of Indirect Mediation Effects (N=187) Click for larger image