Abstract

The Current Study

The conceptual framework for this study was based on a stress-adaptation model (Pollock, 1986) that describes the process of adaptation specific to youth with type 1 diabetes (Whittemore, Jaser, Guo, & Grey, 2010) (Figure 1). How youth adapt to the stress of diabetes is impacted, or mediated, by the ways in which they cope with the stress of diabetes (Jaser et al., 2012), their self-efficacy for diabetes management (Johnston-Brooks, Lewis, & Garg, 2002), their social competence (Edgar & Skinner, 2003), and family functioning (Anderson et al., 2002; Cohen, Lumley, Naar-King, Partridge, & Cakan, 2004; Wysocki et al., 2009). Thus, improvements in QOL and metabolic control are proposed to occur through changes in coping, self-efficacy, social competence, self-management, and family conflict. Our research team designed two Internet -based psycho-educational interventions—TEENCOPE^TM and Managing Diabetes—to impact these mediating factors. We demonstrated that both the education (Managing Diabetes) and coping skills training (TEENCOPE^TM) interventions improved QOL in adolescents with type 1 diabetes over 12 months, but metabolic control did not improve in either intervention group (Grey et al., 2013). The purpose of the current study was to examine the proposed mediators (coping, self-efficacy, social competence, self-management, and family conflict) of the effects of two Internet -based interventions on the primary outcomes of metabolic control and QOL.

Open in a separate window

Figure 1.

Conceptual framework used to determine theoretical mediators of the interventions (individual and family responses).

Method

The current study was a multisite clinical trial aimed at comparing the efficacy of two Internet psycho-educational interventions for youth with type 1 diabetes transitioning to adolescence (TEENCOPE™ and Managing Diabetes) (Whittemore et al., 2012). The coping skills training intervention, TEENCOPE^TM, developed a cast of ethnically diverse adolescent characters who had type 1 diabetes to model coping skills. A graphic novel format demonstrated each of the following skills: Assertive communication, social problem-solving, stress management, positive self-talk, and conflict resolution (Grey et al., 2000). Further, a monitored online discussion board allowed TEENCOPE™ participants to communicate with other youth enrolled in the study. The educational intervention, Managing Diabetes, used a visually appealing and highly interactive interface to present diabetes education (American Diabetes Association, 2013; Silverstein et al., 2005). This program included tailored problem-solving exercises on healthy eating, physical activity, glucose control, sick days, and diabetes technology. The programs were developed with feedback from adolescents with type 1 diabetes and input from diabetes educators (Whittemore, Grey, Lindemann, Ambrosino, & Jaser, 2010). Each program had five sessions that were released weekly, and each session took approximately 30 min to complete. Participation was high, with 77% of TEENCOPE^TM and 79% of Managing Diabetes participants completing at least 4 of 5 sessions, and 52% of TEENCOPE^TM participants contributing to the online discussion board. Satisfaction was also high (mean score = 3.9 of 5).

Procedure

Institutional review boards at all clinical sites approved the study. Youth and parents were recruited from outpatient diabetes clinics, and trained research personnel obtained informed consent/assent. Parents provided demographic data at the time of enrollment, and youth were provided instructions for secure, online collection of psychosocial data at baseline and at 3, 6, 12, and 18 months. Only data up to 12 months were used in the current analysis, as participants were invited to crossover to the other intervention after 12 months. If online data collection was not completed within 3 months of enrollment, youth were considered as passively refusing study participation. Routine clinic-acquired HbA1c values were obtained by chart review. See Figure 2 for CONSORT (Schulz, Altman, & Moher, 2010) flow diagram.

Open in a separate window

Figure 2.

CONSORT flow diagram.

Measures

All measures have been used previously in studies of adolescents with type 1 diabetes. Psychometric information on each measure is provided in Table I.

Table I.

Psychometric Properties of the Major Study Variables

Variable	N	M	SD	Α	Range
					Potential	Actual
HbA1c	314	8.3	1.5			5.4–14.4
PedsQL	320	80.8	11.8	0.87	0–100	43.5–100.0
Primary control coping	320	.19	.04	0.77
Secondary control coping	320	.26	.06	0.80
Disengagement coping	320	.15	.03	0.77
Stress reactivity	320	.19	.04	0.87
Diabetes self-efficacy	320	45.1	11.7	0.88	24–120	24–81
Self-management of diabetes
Collaboration	320	21.8	6.7	0.68	0–39	4–38
Activities	320	32.3	5.7	0.74	0–45	12–44
Problem solving	320	14.5	3.9	0.80	0–21	2–21
Social competence	319	16.4	3.2	0.75	5–20	5–20
Diabetes family conflict	320	26.3	5.9	0.87	19–57	19–57

Open in a separate window

Note. The Primary Control Coping, Secondary Control Coping, Disengagement Coping, and Stress Reactivity variables do not have a range, as they are proportion scores, calculated by dividing the total score for each variable by the total score for the entire measure.

Metabolic control was measured by HbA1c, an indicator of the metabolic control over the past 8–12 weeks, using the DCA2000 (Bayer, Tarrytown, NY) in the majority of participants. Very few (3%) results were done by outside laboratories, and these results were not significantly different from those from the DCA2000.

QOL was evaluated with the Pediatric Quality of Life Inventory (PedsQL) (teen version)–Core (Varni, Seid, & Kurtin, 1999). This instrument consists of 23 items that measure health-related QOL, with higher scores reflective of better QOL.

Coping was assessed with the Responses to Stress Questionnaire (Connor-Smith, Compas, Wadsworth, Thomsen, & Saltzman, 2000), in which adolescents reported on how they responded to diabetes-specific stressors with 57 items. The four factors included in the current study were primary control coping (problem solving, emotional modulation, emotional expression), secondary control coping (positive thinking, cognitive restructuring, acceptance, distraction), disengagement coping (avoidance, denial, wishful thinking), and stress reactivity (physiological arousal, rumination). Proportion scores were used to control for response bias, and higher scores indicate that adolescents reported greater use of that type of coping in relation to their total coping. Cronbach’s alpha ranged from 0.77 to 0.88 in our sample.

The Self-Efficacy for Diabetes Scale measures youth perceptions about their personal competence and resourcefulness for successfully managing their diabetes (Grossman, Brink, & Hauser, 1987). The diabetes-specific self-efficacy subscale (24 items) was used in this study, with lower scores indicative of higher self-efficacy.

The Self-Management of Diabetes–Adolescents (Schilling et al., 2009), a 52-item self-report measure, was used to evaluate adolescents’ self-management. The subscales Collaboration with Parents, Diabetes Care Activities, and Diabetes Problem Solving were used in this analysis. Higher scores indicate better self-management.

Social competence was measured with the social acceptance subscale (five items) of the Self-Perception Profile for Adolescents scale (Harter, 1988). Higher scores reflect greater perceived social competence, such as satisfaction with number of friends.

The Diabetes Family Conflict Scale was used to evaluate diabetes-related treatment conflict (Hood, Anderson, Butler, & Laffel, 2007). This 19-item scale measures conflict between family members on diabetes management activities, and higher scores on this scale are indicative of greater diabetes-related conflict.

Results

Preliminary Analyses

The two groups were comparable at baseline, with the exception of years of parental education, with those in Managing Diabetes having 0.7 years more education (Table II). There were, however, significant differences across the four sites in race/ethnicity, income, therapy type (pump or injections), parent education, and HbA1c, and these were controlled in the analyses.

Table II.

Comparison of Demographic and Clinical Characteristics by Group

	TEENCOPETM n = 167	Managing Diabetes n = 153
	n (%)	n (%)	p-value
Gender
Male	74 (44.3)	69 (45.1)	.89
Female	93 (55.7)	84 (54.9)
Race
White, non-Hispanic/ non-Latino	109 (65.7)	95 (62.5)	.56
Non-White	57 (34.3)	57 (37.5)
Annual income
<$40,000	35 (22.2)	30 (19.9)	.89
$40,000–$79,999	44 (27.8)	43 (28.5)
$80,000+	79 (50.0)	78 (51.6)
Therapy type
Pump	99 (59.3)	91 (59.5)	.97
Injection	68 (40.7)	62 (40.5)
	Mean (SD)	Mean (SD)
Age	12.3 (1.1)	12.3 (1.1)	.89
Parent guardian education (years)	14.3 (2.7)	15.0 (2.9)	.03
Duration (months)	365 (230)	342 (208)	.35
HbA1c at baseline (%)	8.4 (1.6)	8.2 (1.3)	.14

Open in a separate window

Effects of Intervention on Mediators

In the TEENCOPE^TM group, there were significant changes in self-efficacy, primary control coping, secondary control coping, stress reactivity, and collaboration with parents, after controlling for demographic variables (Table III). In the Managing Diabetes group, there were significant changes in self-efficacy, collaboration with parents, problem solving, and social acceptance (Table II).

Table III.

Mediation Effects on HbA1c Level and Quality of Life

Variable (standardized)	Changed mediator at 12 months	Mediation effect on HbA1c level	Mediation effect on quality of life
	Δ ± StdErr (p-value)	Coeff ± StdErr (p-value)	Coeff ± StdErr (p-value)
TEENCOPE
Self-efficacy	−0.651 ± 0.120 (<.0001)	0.123 ± 0.085 (.1494)	−2.100 ± 0.676 (.0022)
Primary control coping	0.272 ± 0.117 (.0219)	−0.105 ± 0.081 (.1992)	3.032 ± 0.625 (<.0001)
Secondary control coping	0.284 ± 0.104 (.0074)	−0.096 ± 0.095 (.3178)	3.226 ± 0.737 (<.0001)
Disengagement coping	−0.233 ± 0.124 (.0627)	0.081 ± 0.076 (.2934)	−0.942 ± 0.617 (.1286)
Stress reactivity	−0.324 ± 0.114 (.0054)	0.034 ± 0.084 (.6820)	−2.959 ± 0.655(<.0001)
Collaboration with parents	−0.360 ± 0.090 (.0001)	−0.040 ± 0.088 (.6491)	0.785 ± 0.716 (.2750)
Diabetes care activity	−0.090 ± 0.089 (.3098)	−0.125 ± 0.086 (.1500)	2.536 ± 0.695 (.0003)
Diabetes problem solving	0.189 ± 0.104 (.0711)	−0.036 ± 0.084 (.6689)	1.136 ± 0.664 (.0886)
Social acceptance	0.080 ± 0.094 (.4008)	−0.041 ± 0.093 (.6582)	2.648 ± 0.778 (.0008)
Family conflict	0.109 ± 0.102 (.2856)	0.186 ± 0.084 (.0290)	−2.790 ± 0.617(<.0001)
Managing Diabetes
Self-efficacy	−0.242 ± 0.114 (.0352)	−0.022 ± 0.065 (.7299)	−2.565 ± 0.490 (<.0001)
Primary control coping	0.087 ± 0.099 (.3833)	0.012 ± 0.080 (.8820)	1.112 ± 0.602 (.0662)
Secondary control coping	0.026 ± 0.103 (.8007)	0.029 ± 0.071 (.6835)	1.415 ± 0.567 (.0135)
Disengagement coping	0.013 ± 0.115 (.9118)	−0.031 ± 0.064 (.6320)	−0.774 ± 0.518 (.1368)
Stress reactivity	−0.146 ± 0.086 (.0940)	0.061 ± 0.082 (.4571)	−0.953 ± 0.662 (.1514)
Collaboration with parents	−0.375 ± 0.085 (<.0001)	−0.021 ± 0.076 (.7851)	−0.162 ± 0.639 (.8002)
Diabetes care activity	−0.097 ± 0.081 (.2350)	−0.014 ± 0.078 (.8599)	1.181 ± 0.638 (.0656)
Diabetes problem solving	0.278 ± 0.086 (.0017)	0.039 ± 0.071 (.5855)	−0.458 ± 0.620 (.4606)
Social acceptance	0.216 ± 0.095 (.0244)	−0.024 ± 0.080 (.7655)	2.210 ± 0.609 (.0004)
Family conflict	−0.205 ± 0.114 (.0754)	0.032 ± 0.058 (.5782)	−1.703 ± 0.492 (.0007)

Open in a separate window

Note. The change in mediators at 12 months was estimated after controlling for demographic variables (gender, age, race/ethnicity, duration of diabetes, household income, therapy type, and study site). The mediation effect was estimated considering the variability of the cross-sectional correlation between mediator and outcome over time, controlling for demographic variables.

Discussion

This study examined the theoretically proposed mediators of two Internet psycho-educational programs for young adolescents with type 1 diabetes, both of which had a positive effect on QOL. Youth who received the TEENCOPE^TM intervention reported increased use of adaptive coping strategies (i.e., greater primary control and secondary control coping) and a decrease in stress reactivity, which mediated the effects of the intervention on QOL. Youth who received an educational intervention reported increased social competence, which mediated the positive effects of the intervention on QOL. In addition, self-efficacy improved in both groups, which also mediated the positive effects of the interventions on QOL. Results of the mediation analysis support several hypothesized relationships of the theoretical framework that guided the study (Whittemore, Jaser, et al., 2010), and suggest that components of both diabetes education and coping skills may help to achieve better QOL outcomes in youth with type 1 diabetes.

Self-efficacy is recognized as a central factor contributing to behavioral change and adherence to treatment regimens in youth with type 1 diabetes (Johnston-Brooks et al., 2002). Many psycho-educational interventions for youth with type 1 diabetes are based on Bandura’s social learning theory, which posits that behavior change occurs when self-efficacy for the behavior increases (Bandura, 1986). As such, the proposed mechanism of change for these interventions is increased self-efficacy; however, few researchers have explicitly tested this. By examining self-efficacy as a mediator of treatment effects, we were able to determine whether the interventions worked as intended. In fact, in both the TEENCOPE^TM and Managing Diabetes groups, improved self-efficacy mediated intervention effects on QOL. Self-efficacy is enhanced through role modeling and skill mastery, which were objectives of both Internet interventions. The TEENCOPE^TM intervention focused on modeling coping skills using role models (e.g., how to effectively communicate with a teacher about diabetes), whereas the Managing Diabetes intervention included practice in specific diabetes management skills (e.g., calculating insulin adjustments). Thus, the two interventions may have improved self-efficacy in different aspects of diabetes management, both of which resulted in improved QOL.

As expected, the TEENCOPE™ intervention improved coping skills by encouraging adolescents to use more primary control coping (e.g., problem solving) and secondary control coping (e.g., positive thinking) strategies. TEENCOPE™ also reduced adolescents’ reports of stress reactivity, which may reflect the use of strategies learned in the stress reduction module. Results of the mediator analyses of the Internet TEENCOPE™ program replicates outcomes from previous in-person group coping skills training groups, which also resulted in improved QOL (Grey et al., 2000), and provide further support for the proposed model of adaptation. In an intervention for children of depressed parents, increased use of secondary control coping skills also mediated the effect of the intervention on psychosocial outcomes (e.g., depressive symptoms) (Compas et al., 2010). Taken together, these results support the importance of teaching adaptive coping skills to youth with type 1 diabetes (Jaser et al., 2012).

For adolescents who received the Managing Diabetes intervention, increased social competence mediated the effects of the intervention on QOL. This was somewhat surprising, given that the Managing Diabetes intervention did not target social skills. Perhaps by teaching diabetes-related skills in the context of healthy, positive images of teens with diabetes, youth who received this program had improved self-concept. This result, however, is in line with previous findings that social competence and positive self-concept were associated with better emotional well-being in youth with type 1 diabetes (Edgar & Skinner, 2003).

Although diabetes-specific family conflict has a strong association with QOL (Laffel et al., 2003) and poorer metabolic control (Ingerski et al., 2010), we did not find a decrease in family conflict to be a significant mediator of the effects of the interventions on either of these primary outcomes. There may have not been enough variation in the level of family conflict reported in our sample of early adolescents, who possibly had greater parental engagement in diabetes management than would be evident in an older group of adolescents.

Unlike in the original in-person Coping Skills Training groups (Grey et al., 2000), we did not find a significant effect of the Internet intervention on metabolic control. There are several possible reasons for this discrepancy. First, the original intervention was conducted with older adolescents (ages 12–20, mean age = 14), whereas the current study focused on early adolescents (ages 11–14, mean age = 12). The decision to target a younger age-group was made to prevent the deterioration in adherence and metabolic control often seen in older adolescents before it occurs (Hegelson, Siminerio, Escobar, & Becker, 2009). Thus, results from the current study are more in line with the Coping Skills Training groups conducted in-person with school-age children (ages 8–12, mean age = 10), in which we found no effect on metabolic control, but did show improvements in QOL, self-efficacy, and coping over 12 months for both the coping skills training and education interventions (Grey et al., 2009). Low rates of participation in the school-age groups (24% of those consented could not be scheduled for groups) were a driving factor for moving to online interventions. Although our effects may be smaller than in the original study, our potential reach is greater. Another difference was that the original coping skills training intervention included 6 weekly sessions, followed by monthly visits for 12 months. This is a significantly larger “dose” than what was included in the TEENCOPE^TM and Managing Diabetes interventions, which consisted of 5 weekly sessions. It is possible that a greater number of sessions or more booster sessions are needed to have an effect on metabolic control. Thus, although we did not find an improvement in metabolic control, the minimal increase in HbA1c over 12 months during this critical stage of adolescence may be seen as a positive outcome in this age-group. Contrary to other studies demonstrating that self-efficacy enhanced glucose control (e.g., Johnston-Brooks et al., 2002), self-efficacy did not mediate intervention effects on HbA1c in our sample. This may be due to the relatively small changes in HbA1c over time.

Funding

This work was supported by the National Institute of Nursing Research (grant number R01-NR04009); the Jonas Centers for Nursing Excellence and National Institute of Nursing Research award to A.C. (grant number T32-NR-008346-10); and the National Institute of Diabetes and Digestive and Kidney Diseases award to S.J. (grant number K23-DK088454;). Clinical Trials Registry Number: NCT00684658.

Conflicts of interest: None declared.

References