|Home | About | Journals | Submit | Contact Us | Français|
To examine age and time trends in responsibility for diabetes management tasks and diabetes-specific family conflict and their relationship to blood glucose monitoring (BGM) frequency and blood glucose control (HbA1c).
A sample of 147 adolescents (mean=15.5±.4 years) with type 1 diabetes and their caregivers completed measures of diabetes-specific responsibility and family conflict at baseline and six months. BGM frequency and HbA1c were measured during outpatient clinic appointments.
Responsibility for diabetes management tasks shifted from caregivers to adolescents with increasing age by adolescent and caregiver report. Diabetes-specific conflict was stable. Similar trends in responsibility and conflict were seen over the 6 month follow-up period. Less frequent BGM and higher HbA1c were also observed with increasing adolescent age. Multivariate analyses demonstrated adolescents taking greater responsibility for management tasks and experiencing greater family conflict at baseline reported lower BGM at six months. Family, demographic, psychosocial, and disease-specific variables accounted for 26% of the variance in BGM frequency by both adolescent and caregiver report. Adolescents reporting greater diabetes-specific family conflict at baseline experienced higher HbA1c values at six months. Variables accounted for 23 and 28% of the variance in HBA1c by adolescent and caregiver report respectively.
Diabetes-specific responsibility and conflict have important implications for improving disease outcomes. Interventions targeting responsibility and conflict (i.e., reducing conflict while keeping caregivers involved in diabetes management) may help prevent the deterioration in BGM and HbA1c frequently seen during adolescence.
Adolescence is a critical period of growth accompanied by changes in interpersonal roles, responsibilities, and identity development . Not surprising, these years are more complex for adolescents diagnosed with type 1 diabetes. A common chronic illness of childhood, data suggest current rates of type 1 diabetes are increasing and affect approximately every 1 in 500 youth ages 10 to 19 years in the United States [2, 3]. In addition to experiencing the same challenges as their peers, adolescents with diabetes must contend with intensive medical regimens, regular clinic appointments, complicated carbohydrate calculations, and frequent monitoring of blood glucose levels .
Although tools are currently available to facilitate more optimal glycemic control  and decrease the risk of long-term, negative health consequences , many adolescents with type 1 diabetes achieve suboptimal glycemic control. The literature documents decreased blood glucose monitoring (BGM) frequency  and deterioration in blood glucose control or hemoglobin A1c (HbA1c) [8, 9], with few adolescents achieving optimal HbA1c values (< 7.5%) .
These poor health outcomes have prompted investigation into possible explanations for this deterioration in glycemic control. Research has demonstrated that several demographic (e.g., family composition, ethnicity), disease (e.g., mode of insulin delivery), individual, and family factors contribute to poor control . In addition to individual psychosocial factors, such as child depression  and anxiety , a number of family factors  affect the frequency of BGM frequency, a commonly cited indicator of adherence  and blood glucose control. Of note, diabetes-specific supervision and/or responsibility (e.g., the extent to which the parent and/or the adolescent are responsible for managing diabetes management tasks) and family conflict have emerged as important correlates of both BGM and blood glucose control. Less parental involvement, greater adolescent responsibility [13-17], and greater family conflict surrounding diabetes management [18-22] are frequently related to less frequent BGM and decreased blood glucose control. Furthermore, responsibility for daily diabetes care often shifts from the parent to the adolescent [23, 24] and family conflict increases .
While it seems that responsibility and conflict have important implications for BGM frequency and blood glucose control, previous research has not investigated how these important family variables are related to BGM frequency and blood glucose control across time. Most of the earlier data [13, 15, 16, 18-21] derive from cross-sectional studies. In fact, only three of the above studies (i.e., [14, 17, 22]) completed longitudinal data collection across multiple time points. Additionally, even these studies [14, 17, 22] utilized a wide age-range of children and adolescents and/or used varying methodologies (e.g., parent versus child-report of functioning, measured only one aspect of family functioning). No research currently exists that examines both age and time trends in family responsibility and family conflict and how they relate to BGM frequency and blood glucose control in an exclusively adolescent sample. Given the well-documented deterioration in blood glucose control that occurs in this age group [8, 9], such new information on this age cohort may have important implications for providers working to identify treatment targets and to plan targeted interventions to assist adolescents in achieving more optimal control.
The current study examined trends in responsibility for diabetes management tasks and diabetes-specific family conflict across adolescent age groups and changes in these variables across time from baseline to six months. In addition, the study examined the relationship of these variables at baseline to BGM frequency and blood glucose control at six months. We hypothesized that (1) responsibility would shift from parent to adolescent and family conflict would increase across age groups and across time from baseline to six months and (2) greater family conflict and greater adolescent responsibility for diabetes management at baseline would be related to less frequent BGM and worse blood glucose control at six months.
Data for this report are part of a larger longitudinal study examining the psychosocial functioning of adolescents with type 1 diabetes. A total of 147 adolescents (aged 13 to 18 years) with type 1 diabetes and their primary caregivers participated in the current study. All adolescents had a diagnosis of type 1 diabetes according to the practice guidelines of the American Diabetes Association  and were receiving care from a multidisciplinary team at a pediatric diabetes center in the Midwest. Exclusion criteria included: (1) the presence of a major psychiatric or neurocognitive disorder that would inhibit their ability to participate; (2) a significant medical disease other than type 1 diabetes, treated thyroid disorders, or celiac disease; or (3) the inability to read or understand English. Participants completed measures at baseline (time 1) and again six months later (time 2). A total of 150 adolescents participated from the 166 eligible adolescents approached (participation rate of 90%). Of those 150 adolescents participating at time 1, 147 adolescents participated at time 2 (98% retention rate). Only participants with data at both time points were included in analyses. Data for eligible families that declined participation were removed from the database, precluding additional analyses to determine if participants differed significantly from those who declined. All study procedures were approved by the local Institutional Review Board. After obtaining written informed consent from caregivers and assent from adolescents (consent if aged 18 years), a research assistant administered questionnaires. All questionnaires were completed in the pediatric diabetes clinic before or after the adolescent's scheduled clinic visit.
Diabetes-specific family conflict was evaluated with the revised Diabetes Family Conflict Scale (DFCS) . This version of the original DFCS  contains updated language and additional items related to contemporary diabetes management (e.g., new technologies). Caregivers and youth completed this scale independently. The level of diabetes-specific family conflict was rated on a 3-point scale (1 = never argue, 2 = sometimes argue, 3 = always argue) across 19 diabetes management tasks such as insulin administration, checking blood glucose values, and telling others about diabetes. A total score, ranging from 19 to 57, was calculated for each youth and caregiver with higher scores indicating more conflict. Internal consistency for the DFCS in this study sample was high for both youth and caregivers (adolescent coefficient α = 0.86; caregiver coefficient α = 0.87).
The 17-item Diabetes Family Responsibility Questionnaire (DFRQ) was used to assess caregiver-child teamwork in daily diabetes management tasks. This measure has strong psychometric properties  and both caregivers and youth completed equivalent forms and assigned a value ranging from 1-3 to each item (1 = child taking or initiating responsibility for this task almost all the time; 2 = caregiver and child sharing responsibility for this task almost equally, and 3 = caregiver taking or initiating responsibility for this task almost all of the time). Scores could range from 17 (child has complete responsibility) to 51 (caregiver has complete responsibility). In this sample, internal consistency was high for both youth and caregivers (adolescent coefficient α = 0.72; caregiver coefficient α = 0.85).
The 27-item Children's Depression Inventory (CDI)  assessed adolescent depressive symptoms at baseline. Items are rated from 0 (no symptom) to 2 (distinct symptom). Possible scores range from 0-54 with a clinical cutoff score of 13 or higher indicative of elevated depressive symptoms and suggestive of further evaluation . Youth responses on the CDI demonstrated a high degree of internal consistency (coefficient α = 0.90).
The 40-item State-Trait Anxiety Inventory (STAI)  measured adolescent anxiety symptoms at baseline. The STAI includes items related to present feelings (state scale) and feelings in general (trait scale). On both scales, there was a high degree of internal consistency (state coefficient α = .87; trait coefficient α = .87).
Adherence was measured as the frequency of blood glucose monitoring (BGM frequency). BGM frequency was calculated based on available data from clinic visits. Participants' meters were downloaded and daily frequency calculated across the past two weeks. If meters were not available, the value reported in the medical chart (i.e., patient-report) or adolescent self-report was used. If self-report was used, we adjusted this value. For example, at time 2, self-report was used for 21 adolescents and 126 adolescents had both meter and self-report values. Meter values were 84% of the self-report values, so this served as the adjustment. The self-report value was multiplied by 0.84 to obtain the value used in analyses. At time 1, self-report was used for 32 adolescents and the self-report value was multiplied by 0.89. We have reported previously on using this methodology to calculate BGM frequency . Adolescent blood glucose control (HbA1c values) was measured by the DCA 2000+ (reference range 4.3-5.7%, Bayer Inc., Tarrytown, NY). Duration of diabetes and mode of insulin administration were obtained from chart review. Family demographic data were obtained from a self-report questionnaire completed by the adolescent's primary caregiver during the study visit.
Prior to analysis, data were double entered and cross-checked for accuracy. Descriptive statistics and frequencies were calculated across variables of interest as appropriate. To examine baseline age trends in responsibility and conflict, Analysis of Variance (ANOVA) models tested for differences between age groups. Given the small number of adolescents aged 18 years (n = 11), these youth were added to the 17-year-old age group. ANOVA models also examined trends in responsibility and conflict from baseline to six-months. Hierarchical regression analyses examined the relationship between responsibility and conflict at baseline to BGM and HbA1c at six months. Demographic and disease-specific variables [age, disease duration, gender (male/female), minority status (white, minority), insulin mode (injection/CSII) health insurance (private/public) caregiver marital status (married/other), and caregiver education (college degree/other)] were entered into the first block. Psychosocial covariates [child depression, state anxiety, trait anxiety] were entered into the second block and responsibility and conflict entered into the third block. This method allowed researchers to examine the unique contribution of family factors beyond that accounted for by demographic/disease-specific and individual psychosocial variables. Interaction effects between responsibility and conflict with time were also entered into regression models. Analyses were conducted in SAS 9.1 for Windows (SAS Institute Inc., 2002-2003).
At baseline, the majority of adolescents (mean = 15.5 ± 1.4 years old, 52% female) were white, not of Hispanic origin (86%), used subcutaneous insulin infusion (CSII) pumps (63%), and had private insurance (85%). Duration of type 1 diabetes ranged from 6 months to 16.75 years (mean = 6.0 ± 3.8 years; n=10 duration less than 1 year). Caregivers of adolescents were mostly mothers (86%), married (76%), and had a college degree (47%; Table 1).
No significant differences across gender (male vs. female), ethnicity (white vs. minority), health insurance (private vs. public) caregiver marital status (married vs. other), or caregiver education (college degree vs. other) were found between age groups.
Across adolescent age groups, diabetes-specific responsibility shifted from parent to adolescent with increasing age according to adolescents (F (4,146) = 6.13, p < .001) and caregivers (F (4,146) = 4.84, p < .01). Age was not associated with diabetes-specific family conflict; conflict remained stable according to both adolescents (F (4,146) = 0.78, p = .54) and caregivers (F (4,146) = 0.71, p = .59). BGM frequency decreased (F (14,146) = 7.78, p < .001) while HbA1c trended higher (F (4,146) = 2.31, p = .06) across age groups.
Diabetes-specific responsibility shifted from parent to adolescent by both adolescent (t (146) = 3.04, p < .01) and caregiver report (t (144) = 2.88, p < .01) from baseline to six months. Diabetes-specific family conflict remained stable by both adolescent (t (146) = 1.92, p = .06) and caregiver report (t (144) = 1.96, p = .05) over six months. Additional analyses revealed that BGM frequency decreased (t (146) = 4.43, p < .001) and HbA1c increased (t (146) = -2.89, p < .01) over six months (Table 2).
After controlling for demographic variables and psychosocial covariates, hierarchical regression analyses revealed significant main effects for adolescent report of family responsibility (β = .18, p < .05) and conflict (β = -.19, p < .05) in the model predicting BGM frequency at six months (F(13,146) = 3.55, p < .001, R2 = .26). Analyses also revealed a significant main effect for adolescent report of family conflict (β = .19, p < .05), but not responsibility (β = -.02, p = .78), in the model predicting HbA1c at six months, (F(13,146) = 3.13 p < .001, R2 = .23; Table 3). Interaction effects with time were not significant and excluded from the model.
Hierarchical regression analyses revealed significant main effects for caregiver report of family responsibility (β = .21, p < .05) and conflict (β = -.21, p < .05) in the model predicting BGM frequency at six months (F(13,146) = 3.60, p < .001, R2 = .26) after controlling for demographic variables and psychosocial covariates. Analyses also revealed a significant main effect for caregiver report of family conflict (β = .31, p < .001), but not responsibility (β = -.05, p = .61), in the model predicting HbA1c at six months (F(13,146) = 4.05, p < .001, R2 = .28; Table 4). Interaction effects with time were not significant and were excluded from the model.
This study demonstrated that diabetes management tasks shifted from caregivers to adolescents with increasing adolescent age, and family conflict related to diabetes management remained stable across time. During this shift in responsibility for care, BGM frequency declined and the concentration of HbA1c increased. Younger adolescents taking greater responsibility for care and experiencing greater family conflict at baseline reported less frequent BGM six months later. In addition, for adolescents and caregivers reporting greater diabetes-specific family conflict at baseline, adolescents obtained higher HbA1c values at six months. These findings suggest that improving the transition of responsibility from caregivers to adolescents in daily diabetes care and decreasing family conflict may reduce the deterioration in diabetes outcomes for adolescents with type 1 diabetes.
Current findings support previous studies that have also described a similar shift in responsibility for diabetes management tasks from parents to adolescents with increasing age [23, 24]. While parents of 13-year-olds in the current sample were sharing fairly equal responsibility for management tasks, responsibility for disease management tasks had clearly shifted to greater responsibility on the part of the adolescent by age 17. Some of this shift to greater responsibility by adolescents is developmentally appropriate as they attain greater independence and prepare to transition away from the home and their pediatric healthcare providers . However, this shift in responsibility for disease management tasks adversely affects the frequency of BGM measurement. In contrast, families sharing greater responsibility for management tasks reported more frequent BGM. This finding supports the need for continued involvement of parents and caregivers in their adolescent's care. Given that some adolescents were already taking greater responsibility for disease management as early as age 13, the promotion of continued involvement by caregivers in disease management needs to begin in the pre-adolescent years.
A growing body of literature now exists documenting the significant difficulties associated with the transition from pediatric to adult healthcare settings [32-35] and a number of interventions and programs to help ease the transition process to adult healthcare have been described or are currently underway [36-38]. While it may seem that more extensive interventions regarding responsibility and self-management are needed during this period, current data suggest that difficulties related to self-management may begin long before this transition takes place. It may be that prevention efforts targeting both family conflict and division of responsibility in the adolescent age group may be especially beneficial earlier in adolescence. Discussions surrounding disease management, sustained parental involvement, and family conflict should then be an integral component of prevention efforts throughout the adolescent years. For example, Anderson and colleagues  have previously described a four-session, clinic-based, parent-adolescent teamwork intervention that includes a focus on both responsibility and conflict reduction within the family. Demonstrating positive effects on both HbA1c and family conflict, this intervention provides promising evidence regarding possible strategies to decrease the difficulties documented in our current data. Future, longitudinal research will determine if prevention efforts during this time may subsequently serve as a foundation for the family as they finally transition to adult healthcare providers.
We also found that greater diabetes-specific family conflict at baseline was related to less frequent BGM and higher HbA1c values at six months. In contrast to previous literature , family conflict remained relatively stable across both age and time through adolescence. While conflict related to management tasks remains an important factor in understanding adolescents' overall disease functioning, it may be that patterns of conflict are fairly established by the adolescent years. Family conflict is perhaps one of the more widely studied family variables known to impact adherence and glycemic control in pediatric diabetes samples [18-22], and current data certainly support its continued consideration in explaining the health outcomes of adolescents in future studies. A number of interventions targeting family conflict in type 1 diabetes populations are described in the literature (e.g., ) and have had some success in improving disease-related outcomes. Given our results regarding the stability of conflict over time, it seems such intervention may be especially beneficial early in adolescence to address those difficulties related to conflict found to impact BGM and HbA1c.
Similar to other studies , we found that child anxiety was related to BGM frequency in the model using caregiver report of responsibility and conflict, supporting continued consideration of child anxiety symptoms in understanding disease-related care. That family characteristics remained significant even after individual child psychosocial variables were added to the model, speaks to the importance of family factors to diabetes outcomes. Also consistent with existing literature, we found deterioration in BGM frequency and glycemic control with advancing age [7-9]. Specifically, 17-year-olds were monitoring their blood glucose levels almost two times less frequently each day than 13-year-olds. This finding was associated with a concomitant increase in HbA1c. Interestingly, these changes were observed in a sample where the majority of adolescent participants were white, enrolled in private insurance, living with a married caregiver, and where two-thirds of the adolescents used CSII. This decline in glycemic control over only six months further justifies the need to understand those factors affecting this important health outcome.
This study has several limitations. First, despite efforts to document family conflict and responsibility by both adolescent and caregiver report, families may not have accurately disclosed current conflict and/or patterns of responsibility due to social desirability. Second, the data were collected from a single diabetes center where adolescents were predominantly white and from relatively high functioning families limiting general applicability. However, current findings regarding BGM frequency and HbA1c are similar to those published elsewhere [8, 9] and provide support for its relevance to other adolescent samples. Third, only two diabetes-specific family factors were examined in the current study. While demographic, disease-specific, and child psychological factors known to influence BGM and A1c in previous literature were considered in analyses, there are likely other family or individual psychosocial variables influencing BGM and HbA1c in this population.
In summary, this study suggests that targeting responsibility and family conflict may benefit both BGM and HbA1c. Certainly, the current data have important implications for adolescents followed in multidisciplinary specialty clinics and for healthcare professionals providing care to this high-risk age group. Differences in models by reporter are consistent with the current literature (e.g., ), suggesting important differences in how parents and adolescents view the family environment and confirming the needed for multiple informants. Psychosocial assessment is needed to identify those family factors that may be impacting type 1 diabetes management. In addition to strategies related to transitional care (i.e., sustained parental involvement with increasing adolescent age) these data also provide additional support for the use of family-based interventions in working with adolescents to address challenges related to care. Furthermore, a focus on modifiable family factors, such as responsibility and conflict, may help to improve overall outcomes. Finally, interventions that focus on reducing conflict within the family around diabetes-related tasks while promoting sustained parental involvement in diabetes management tasks may prove to be the most effective in preventing or reducing the deterioration in glycemic control frequently seen during adolescence.
Research supported by a grant from the National Institute for Diabetes and Digestive and Kidney Diseases (K23 DK073340) awarded to Dr. Korey K. Hood.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.