Highlights
· This randomized controlled trial demonstrated that a brief online self-compassion intervention significantly reduced hemoglobin A1c in adolescents and young adults with type 1 diabetes. The improvement in glycemic control was sustained at 3 months, with a similar effect observed in the waitlist control group after receiving the intervention. This is the first trial in an Asian context to provide evidence supporting the integration of self-compassion practices into diabetes care.
Introduction
Despite advancements in managing type 1 diabetes (T1D), fewer than 20% of adolescents and young adults with T1D attain the recommended glycemic control target [1,2]. Psychosocial factors (e.g., depression, anxiety, family conflicts) are frequently linked to poor self-management [3,4]. Diabetes distress, also known as stress related to diabetes, is a frequently encountered condition, characterized by frustration and helplessness in managing diabetes. Notably, it is linked to suboptimal outcomes in individuals with diabetes, independent of depression effects [5-7]. Furthermore, it encompasses adverse mental reactions arising from the emotional burdens and specific concerns associated with managing the complex, round-the-clock care-demands of managing T1D [8], including the constant need for self-monitoring and self-care behaviors (e.g., dietary considerations, insulin administrations, blood glucose monitoring, regular physical exercise).
Among individuals with T1D, diabetes distress has been identified in a substantial percentage of the population (42%–54%), with minimal variations observed over time, underscoring its chronic nature [8]. It is notable that only a small fraction of individuals with diabetes communicate their distress to their diabetes care teams [9], implying that they internalize these concerns or keep them to themselves. Importantly, diabetes distress has been linked to higher levels of hemoglobin A1c (HbA1c) [6,7]. Adolescents and young adults with T1D often experience difficulties in self-care, commonly coupled with increased diabetes distress [10]. Glycemic control is known to be worse during this period [11]. Therefore, alleviating diabetes distress has a potential to enhance self-care behaviors and subsequently improve glycemic control and clinical outcomes [12].
Studies have explored interventions targeting diabetes distress, including self-compassion techniques, which emphasize kindness toward oneself [13]. Neff and Germer [14] defined self-compassion as being composed of 3 elements: self-kindness versus self-judgment, common humanity versus isolation, and mindfulness versus overidentification. Numerous studies have established self-compassion as an effective intervention for a wide variety of mental health conditions [15]. Neff and Germer [14] developed mindful self-compassion (MSC), a skills-building program focused on developing and practicing self-compassion skills with a focus on integrating the practices into daily life. In individuals with diabetes, several studies indicated that greater self-compassion was associated with improved mental health outcomes and played an important role in diabetes management [16]. However, among the 16 studies reviewed, only 2 were randomized controlled studies [13,16,17], and only one enrolled individuals with T1D [13]. In this study involving 63 individuals with either T1D or type 2 diabetes, conducted in New Zealand, the standard 8-week (9 sessions, 27 hours) MSC program led to reductions in diabetes distress and depressive symptoms along with improved HbA1c levels in the intervention group, as well as an increase in self-compassion, when compared to a waitlist control group [13].
Therefore, there is a pressing need to explore the potential benefits of self-compassion in individuals with T1D, particularly in diverse populations, as most of the published research has included participants from the Global North. This randomized controlled trial aimed to explore the effects of a self-compassion intervention in adolescents and young adults with T1D on the primary outcome of diabetes distress, along with secondary outcomes of glycemic control, diabetes self-efficacy, self-compassion, and other patient-reported outcomes. With the increased familiarity with telemedicine after the coronavirus disease 2019 outbreak, this study deployed a self-compassion intervention, derived from the MSC curriculum, adapted to the context of Thailand and delivered through an online platform. The results should provide further knowledge about the effects of kindness on diabetes distress and glycemic control in individuals with T1D.
Materials and methods
2. Participants
Adolescents and young adults (aged 18–30 years) with T1D who were receiving medical care from the Faculty of Medicine at Ramathibodi Hospital or the Faculty of Medicine at Chulalongkorn Hospital, Bangkok, Thailand, were recruited between October 2021 and January 2022. The inclusion criteria included the following: (1) a clinical diagnosis of T1D with key laboratory parameters documented—specifically, low C-peptide levels and positive autoantibodies (glutamic acid decarboxylase, tyrosine phosphatase–like protein IA-2, or zinc transporter 8) and (2) had access to a computer or tablet with internet connectivity for meeting participations. Exclusion criteria included (1) a history of self-harm or suicide attempt within 1 year, (2) severe complications of diabetes (e.g., blindness, dialysis), (3) active psychiatric disorders, (4) uncontrolled diabetes (HbA1c > 11% within the prior 6 months or a history of severe hypoglycemia or diabetic ketoacidosis [DKA] requiring medical interventions within the prior 3 months), and (5) pregnant or intending to become pregnant.
3. Treatment allocation
Participants were randomly assigned to the self-compassion intervention and waitlist control groups, at a 1:1 ratio, using a computer-generated randomization (performed by author SR), stratified by age (2 groups: ages 18–24 and 25–30 years). Due to the nature of a psychological intervention, the research team and participants were not blinded.
1) Interventions
The self-compassion intervention for this randomized controlled trial was derived from the standard MSC curriculum. Thailand's only bilingual, certified MSC teacher (author BW) collaborated with a bilingual team of Thai healthcare professionals with experience in caring for T1D patients. All team members underwent MSC training prior to the study start. The standard MSC curriculum consists of eight 3-hour sessions plus a retreat session, for a total of 27 hours [13]; however, given the significant time commitment and busy lifestyles of young adults with T1D, the MSC curriculum was adapted to be more concise for use in this study. This modification aimed to optimize participant engagement and adherence while ensuring the program's effectiveness within the constraints of time and resources available for delivery. This modification was done by extracting specific MSC practices and techniques deemed to be the most relevant and contextually appropriate for this population (Table 1). The MSC curriculum was previously abbreviated and adapted for use in different contexts and was shown to be effective [18].
The final self-compassion intervention for this study consisted of 6, 2-hour sessions, delivered biweekly via the Zoom conferencing application (Zoom Communications, USA) in a group session, over the course of 12 weeks. All sessions were conducted by the study psychologist, with assistance from team members. Each session taught specific self-compassion skills, emphasizing skill practice in small groups and home practice. The participants were required to attend the first 2 meetings in order to continue participating in the research. Therefore, the first 2 sessions were offered twice to accommodate participants' schedules (Table 1).
4. Measures
At baseline, demographic information (e.g., age, sex, education, insurance coverage), diabetes history (duration of diabetes, medical regimen), and frequencies of self-monitoring blood glucose (SMBG) and hypoglycemia were obtained using questionnaires. Weight (kg) and height (cm) were measured, and body mass index was calculated. A history of hypertension or hyperlipidemia, presence of diabetic microvascular complications (retinopathy, nephropathy, and neuropathy), a history of acute complications (severe hypoglycemia and DKA), and depression/antidepressant use were obtained from medical records.
1) Outcome measures
The following measures were collected at baseline (T1) and at the end of 12 weeks of self-compassion intervention (T2). The measures were repeated at 24 weeks (T3, i.e., 12 weeks after the end of T2). Measures at T3 aimed to assess the durability of the intervention in the intervention group, and the effects of self-compassion in the control group who elected to participate in self-compassion program after the end of randomization period.
(1) Primary outcome
Diabetes distress was a primary outcome in the study and was assessed using the Thai Diabetes Distress Scale.19] This scale has 17 items covering 4 domains: (1) emotional burden (5 items), (2) regimen-related distress (3 items), (3) physician- or nurse-related distress (4 items), and (4) diabetes-related interpersonal distress (3 items). The questions are based on a 6-point Likert scale (1=no problem, 6=serious problem). The overall and sub-component scores were each interpreted as the mean score of components, where <2 points indicate little or no distress, 2–2.9 points indicate moderate distress, and >3 points indicate a high level of distress. The Thai Diabetes Distress Scale has psychometric properties and was previously validated in 170 patients with diabetes, with Cronbach's alpha coefficients for the total score of 0.95 and 0.85–0.96 for the 3 subcomponent scores [19].
(2) Secondary outcomes
Secondary outcomes included glycemic control as assessed by HbA1c, diabetes self-efficacy, depressive symptoms, stress, self-compassion, and self-reported sleep quality. In addition, self-reported frequency of hypoglycemia and self-monitored blood glucose levels were collected. Satisfaction with the self-compassion program was also obtained.
HbA1c (%) was collected by blood sampling at baseline (T1) and the end of the randomization period (T2). HbA1c levels were measured at the clinical laboratory of the Faculty of Medicine at Ramathibodi Hospital or King Chulalongkorn Memorial Hospital. HbA1c levels at T3 were obtained from medical records reviews.
Diabetes self-efficacy was assessed using the Thai version of the Diabetes Management Self-efficacy Scale [20]. This scale contains 20 items covering 4 domains, including diet (9 items), blood sugar monitoring (4 items), physical exercise (4 items), and medication regimen (3 items). Responses are given using a 5-point Likert scale (1–5) points, with total scores ranging from 20–100 points and higher scores reflecting greater diabetes management self-efficacy. The scale has psychometric properties and was previously validated in 700 patients with type 2 diabetes, with Cronbach alpha coefficients for the total score of 0.98 and 0.71–0.92 for scores of the 4 subcomponents [20].
Depressive symptoms were assessed using the Patient Health Questionnaire–9 (PHQ-9) [21]. In the PHQ-9, a total of 9 items are scored on a scale of 0 points (no at all) to 3 points (almost every day), with total scores ranging from 0–27 points and higher scores reflecting stronger depressive symptoms [21]. A previous study reported a Cronbach alpha coefficient of 0.79 [21].
Stress was assessed using the Thai version of the Perceived Stress Scale [22]. This scale has 10 items scored using a 5-point Likert scale of 0 points (never) to 4 points (very often), with total scores ranging from 0–40 points and higher scores reflecting greater perceived stress. This scale has psychometric properties, with a Cronbach alpha coefficient of 0.85 [22].
Self-compassion was assessed using the Self-Compassion Scale [23,24]. There are 26 items consisting of 6 domains, which include both positive questions (self-kindness [5 items], common humanity [4 items], and mindfulness [4 items]) and negative questions (self-judgment [5 items], isolation [4 items], and overidentification [4 items]). The questions are based on a 5-point Likert scale of 1 point (never) to 5 points (very often), with higher scores reflecting greater self-compassion. The questionnaire was validated in a study of 327 Thai female adolescents, which revealed a Cronbach's alpha coefficient of 0.88 [24].
Sleep quality was assessed using the Thai version of the Pittsburgh Sleep Quality Index [25], whose global score ranges from 0–21 points, with higher scores reflecting poorer sleep quality. This index has psychometric properties, and a previous study revealed a Cronbach alpha coefficient of 0.84 [25].
The number of hypoglycemia episodes (weekly frequencies) was scored based on a 5-point Likert scale of 0 points (never), 1 point (<1 time/wk), 2 points (1–2 times/wk), 3 points (3–4 times/wk), and 4 points (>4 times/wk), and severe hypoglycemic episodes (needing assistance or medical intervention) and frequencies of daily SMBG were collected using questionnaires.
Finally, satisfaction with the program was assessed at the end of the intervention using a 7-point Likert scale questionnaire, where scores ranged from 1–7 points, with 7 points indicating most satisfied. Participants also commented on the overall impression of the program and identified the skills they used most often.
5. Sample size calculation
Sample size was calculated based on differences in diabetes distress among diabetes patients who underwent placement in the MSC or waitlist group [13]. The mean (standard deviation [SD]) change in diabetes distress was 0.83 (1.01). A sample size of 19 participants per group was required to achieve a power of 80% with a 2-sided alpha coefficient of 0.05.
6. Statistical analysis
Data are presented as mean±SD values for continuous data or as frequency (%) values for categorical data. The analysis was conducted in 2 parts, where the first part assessed the efficacy of the intervention by using an independent-samples t-test to compare the means of primary and secondary outcomes at T2 between the self-compassion and waitlist groups. This analysis was based on an intention-to-treat analysis.
The second part aimed to assess the durability of the intervention in the intervention group by comparing the outcomes between T2 and T3 in the intervention group using a paired t-test. In addition, the effects of the self-compassion program in the control group who elected to participate in the program after the end of the randomization period were estimated by comparing the outcomes this group between T2 and T3 using a paired t-test. The analyses were performed using IBM SPSS Statistics ver. 28.0 (IBM Co., USA). P<0.05 was used to indicate the level of significance across all analyses.
7. Ethical statement
The study protocol was approved by the Committee on Human Rights Related to Research involving Human Subjects of the Faculty of Medicine at Ramathibodi Hospital (COA. approval no. MURA2021/830, October 4, 2021), and the Faculty of Medicine of Chulalongkorn University (COA approval no. 1714/2021). All participants provided written informed consent. The study was registered at ClinicalTrials.gov (NCT05070949).
Results
The flow of the study is shown in Fig. 1. Participants were recruited from November 2021 to January 2022; among a total of 127 potential participants screened, 34 were eligible and willing to participate and were subsequently randomized to the self-compassion intervention (n=17) or the waitlist control group (n=17). One participant in the self-compassion group dropped out right after the enrollment, and their data were therefore excluded from the analysis. Baseline characteristics of the participants are shown in Table 2. The mean±SD age was 23.5±4.2 years in the intervention group and 23.6±3.2 years in the control group. There were no significant differences between groups in terms of their diabetes duration, body mass index, HbA1c, health insurance coverage, education, and history of acute diabetes complications requiring hospitalization or other comorbidities.
1. Diabetes distress, glycemic control, and other secondary outcomes after self-compassion intervention
Of the 16 participants in the intervention group, 1 (5.9%) attended 1 session and 15 (94.1%) attended at least 5 sessions. For the primary outcome of diabetes distress, there was no significant difference in the change in diabetes distress score after 12 weeks between the self-compassion and waitlist control groups (mean difference [MD], 0.02; 95% confidence interval [CI], -0.26 to 0.31; P=0.876) (Table 3). The analysis of each domain of the diabetes distress score yielded similar results.
For secondary outcomes, there was a significant reduction in A1c in the self-compassion intervention group between baseline and 12 weeks (7.71%±1.07% vs. 7.30%±0.97%), while there was no such change in the waitlist control group (8.24%±1.77% vs. 8.34%±1.96%). This resulted in an HbA1c MD of -0.51% (95% CI, -0.97 to -0.04; P=0.035) (Table 3).
Other secondary outcomes were not significantly different between the groups. The MD in the mean self-compassion score was 0.03 (95% CI, -0.31 to 0.37; P=0.855). Furthermore, self-compassion subscale scores (judgment, humanity, kindness, isolation, mindfulness, and overidentification) were also not significantly different between groups. Additionally, stress and depressive symptoms (PHQ-9), self-reported sleep quality, diabetes self-efficacy, and frequency of self-reported hypoglycemia and SMBG also showed no significant differences.
2. Post self-compassion intervention outcomes (T3 vs. T2)
1) Intervention group
At 24 weeks (T3), the diabetes distress score remained stable (Table 4). In addition, the participants maintained the reduction in HbA1c levels observed at T2 (7.33%±1.00% vs. T3: 7.49%±0.95%; P=0.118) (Table 4). Self-compassion score, stress and depressive symptoms, and self-reported sleep quality did not change. Although diabetes self-efficacy did not change, the frequency of self-reported SMBG decreased significantly (P=0.002).
2) Waitlist control group
Following the end of the randomization period (T2), all control participants opted to participate in the self-compassion program. Of the 17 participants, 2 (11.8%) attended 1 session, 1 (5.9%) each attended 2 and 3 sessions, 7 (41.2%) attended 4 sessions, and 3 (17.6%) attended at least 5 sessions, respectively. After completing the 12 weeks of the self-compassion program (T3) the diabetes distress did not change significantly in these participants (P=0.956); however, their A1c values were significantly decreased (8.34%±1.96% vs. 7.76%±1.4%, P=0.005; MD, -0.58; 95% CI, -0.95 to -0.20). There were no significant changes in self-compassion scores or other outcomes.
3. Satisfaction
Overall satisfaction with the program was found to be high, with a mean±SD score of 5.78±0.87 points among 32 respondents. Additionally, 97% of respondents scored their satisfaction as ≥5 points. Participants provided overwhelmingly positive feedback, such as that the activities "helped me to think about and be kind to myself, and made me relaxed and able to manage my emotions"; "helped me concentrate and solve problems"; "helped me understand different aspects of happiness and suffering and better understand myself and surrounding people"; and "helped me appreciate values in myself, and be able to accept and thank myself."
Participants reported frequent use of specific skills taught in the intervention. The most commonly used skills were STOP (Stop–Take a Breath–Observe–Proceed), followed by the self-compassion break (a core MSC practice) and giving gratitude to small things in daily life.
Discussion
This randomized controlled trial focused on assessing the effects of a contextually adapted, brief version of the self-compassion program in adolescents and young adults with T1D in Thailand. While there was no difference in the primary outcome of diabetes distress between the self-compassion intervention and waitlist control groups, the study demonstrated the effectiveness of self-compassion in improving glycemic control. First, the intervention group showed a significant reduction in HbA1c of 0.51% as compared to the control group, and this HbA1c reduction was maintained at 3 months after the end of the intervention, suggesting a durability of the effect. Second, a similar reduction in HbA1c of 0.58% was observed in the waitlist control participants who received the same self-compassion intervention after the treatment group. Importantly, the improvement in glycemic control was not associated with an increase in self-reported hypoglycemia or DKA. This magnitude of HbA1c reduction is clinically significant, comparable to that achieved with standard treatments for T1D such as carbohydrate counting [26] or automatic insulin delivery [27]. However, there were no changes in other secondary outcomes. Still, high participant satisfaction was uniformly observed. These data suggest that this brief self-compassion program is well-received, safe, and effective in improving glycemic control in adolescents and young adults with T1D.
Greater self-compassion in individuals with diabetes has been shown to be associated with decreased negative affect, positive well-being, better self-care behaviors, and improved self-regulation [28,29]. Moreover, parental empathy and compassion have been linked to greater adherence and positive emotions in children and adolescents with T1D [30,31]. In a cross-sectional study of 111 individuals with diabetes (including 67 with T1D), diabetes-specific distress was a better predictor of A1c than depression, and self-compassion was found to moderate the relationship between distress and A1c (greater distress predicted higher HbA1c levels at a lower level of compassion) [32]. Despite the recognized benefits of self-compassion, only a limited number of studies have explored the effects of self-compassion intervention, e.g., MSC, on glycemic control in individuals with diabetes, and only one study involved individuals with T1D [13]. Friis et al. [13] conducted a randomized controlled intervention study in 32 individuals with diabetes (26 with T1D) compared to 31 waitlist control participants with a mean age of 44.3 years. In their study, an 8-week MSC training led to a reduction in diabetes distress and depression, along with a reduction in HbA1c of ~1.0% [13]. Our results of glycemic improvement are in agreement with those of this previous study and further extend the demonstrated benefits to young adults with T1D, a group at risk for increased diabetes distress and poor glycemic control. More recently, a United States study in 276 adults revealed that a psychologist-led program specifically targeting diabetes distress resulted in reductions in diabetes distress and HbA1c levels [33]. Three additional studies have explored the benefits of MSC on glycemic control in individuals with T2D (n=20 [34], n=38 [17], and n=60 [35]). The results revealed that MSC led to a reduction in distress, better self-care, and improved glycemic control (based on mean glucose or HbA1c measurement) [17,34,35]. Collectively, the evidence suggests a positive influence of self-compassion on glycemic control, psychological well-being, and self-care behaviors in individuals with diabetes.
There are several possible mechanisms underlying the association between increased self-compassion and improved glycemic control, including better diabetes self-care. Diabetes distress and depressive symptoms were shown to be related to physiological changes such as alterations in cortisol secretion [36] and dysregulation of the hypothalamic-pituitary-adrenal axis, sympathetic nervous system activity, and systemic inflammation [37]; all of which can result in insulin resistance and adversely affect glucose metabolism. Furthermore, empathy and compassion were also shown to relate to inflammation and cortisol regulation [38,39]. In our study, however, we did not observe changes in diabetes distress, sleep, or depressive symptoms after the self-compassion intervention, suggesting that the mechanism of HbA1c improvement was mediated by other unmeasured factors, possibly including reductions in systemic inflammation or improved cortisol regulation. Detailed physiological assessments may help to elucidate mechanisms in future research. It is also possible that the questionnaires used in this study were not specific enough to capture changes in self-compassion in our participants with diabetes despite the overwhelmingly positive feedback received on the program.
Our study has strengths and adds significant breadth to the body of evidence about self-compassion in several ways. This is the first randomized controlled study to explore the effectiveness of a self-compassion program on glycemic control in adolescent and young adults with T1D. It is also the first study conducted in an Asian context, highlighting its novelty in this population. However, a limitation of the current study was the relatively small sample size, and we only reached 89.5% of targeted recruitment. While the modification of the self-compassion intervention from the original 8-week session, while maintaining the core components, was done to enhance adherence in this population, who is typically engaged in multiple life demands, it is possible that the intervention was not long enough to observe differences in the self-compassion score. Note that the self-compassion score indeed numerically increased after the intervention, but this was also observed in the waitlist group; thus, the differences were not statistically significant. The lack of attention on the control group does not exclude the possibility that an A1c reduction resulted from simply being in the study, but our study design was similar to that of the previous study [13]. While standardized questionnaires were used, a diabetes-specific self-compassion assessment could possibly better capture this variable [40]. Finally, measurements of detailed physiological measures (i.e., stress hormones), except for A1c, were not obtained, and their inclusion could suggest mechanistic pathways linking the benefits of self-compassion on glycemic control.
In conclusion, a brief self-compassion intervention resulted in clinically significant improvements in glycemic control among adolescents and young adults with T1D. This study strengthens the evidence that healthcare providers should recognize the importance of self-compassion and consider integrating self-compassion practices into diabetes care to improve metabolic outcomes.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Print
