If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Adherence in Diabetes Questionnaire (ADQ) score as predictor of 11-year HbA1c trajectories in children and adolescents with type 1 diabetes: A population-based longitudinal study
Department of Clinical Medicine, Aarhus University, Aarhus, DenmarkSteno Diabetes Center Aarhus, Aarhus University Hospital, Aarhus, DenmarkDepartment of Paediatrics and Adolescent Medicine, Aarhus University Hospital, Aarhus, Denmark
Department of Psychology, University of Southern Denmark, Odense, DenmarkSteno Diabetes Center Odense, Odense, DenmarkDepartment of Medical Psychology, Amsterdam UMC, Amsterdam, The Netherlands
The Danish Clinical Quality Program – National Clinical Registries (RKKP), Copenhagen, Denmark.Steno Diabetes Center Copenhagen, The Capital Region, Denmark
Four distinct 11-year HbA1c trajectories were identified in youth with T1D.
•
Lower youth adherence scores predicted the two highest risk HbA1c trajectories.
•
Lower caregiver adherence scores predicted three higher risk HbA1c trajectories.
•
Low caregiver education level predicted the two highest risk HbA1c trajectories.
•
Single parent family status predicted the highest risk HbA1c trajectory.
Abstract
Aims
To identify 11-year HbA1c trajectories in children/adolescents with type 1 diabetes and determine whether baseline caregiver- and/or child/adolescent-reported Adherence in Diabetes Questionnaire (ADQ) scores and multiple covariates predict HbA1c trajectory membership.
Methods
For a 2009 population-based cohort of children/adolescents with type 1 diabetes, we analyzed HbA1c follow-up (2010−2020) data from Danish diabetes registries. HbA1c trajectories were identified with group-based trajectory modeling. Using multinomial logistic regression, we tested whether ADQ scores predicted trajectory membership when adjusting for sex, age at diabetes diagnosis, diabetes duration, family structure, and caregiver education.
Results
For 671 children/adolescents (10–17 years at baseline) with 5644 HbA1c observations over 11 years, four trajectories/groups were identified: 1) “on target, gradual decrease” (27%), 2) “above target, mild increase then decrease” (39%), 3) “above target, moderate increase then decrease” (25%), and 4) “well above target, large increase then decrease” (9%). Using group one as the reference, lower caregiver-reported ADQ scores predicted group 2, 3, and 4 membership. Lower child/adolescent-reported ADQ scores predicted group 3 and 4 membership. Low caregiver education predicted group 3 and 4 membership. Single-parent status predicted group 4 membership.
Conclusions
ADQ scores and socio-demographics may serve as tools to predict glycemic control in youth with type 1 diabetes.
Follow up data from the Diabetes Control and Complication Trial show that improvement in glycemic control lowers the risk of chronic complications and mortality [
Writing Team for the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group
Sustained effect of intensive treatment of type 1 diabetes mellitus on development and progression of diabetic nephropathy: the epidemiology of diabetes interventions and complications (EDIC) study.
]. Despite this knowledge and advancement in technologies, most children and adolescents with type 1 diabetes mellitus remain well above the appropriate target of a HbA1c of 7.0% (53 mmol/mol) and the majority are above the less stringent target of 7.5% (58 mmol/mol) [
Exploring variation in glycemic control across and within eight high-income countries: a cross-sectional analysis of 64,666 children and adolescents with type 1 diabetes.
]. This indicates that most children and adolescent with type 1 diabetes mellitus are challenged by the adherence to general diabetes treatment recommendations [
ISPAD clinical practice consensus guidelines 2018: Glycemic control targets and glucose monitoring for children, adolescents, and young adults with diabetes.
]. Adherence comprises all child/adolescent's or family's efforts that are required to fulfill an array of diabetes-specific recommendations in collaboration with health care professionals to optimize glycemic control [
Cross-sectional and longitudinal population studies in children and adolescents show an increase in average HbA1c levels during puberty up to 16–18 years [
]. However, all children with type 1 diabetes mellitus do not follow the same HbA1c trajectories during adolescence and young adulthood. Some have optimal and others have suboptimal glycemic control [
Longitudinal trajectories of metabolic control across adolescence: associations with parental involvement, adolescents’ psychosocial maturity, and health care utilization.
], it has been possible to differentiate the longitudinal population-based HbA1c average into 3–5 distinct groups with large differences in average HbA1c levels over time [
Continuity and change in glycemic control trajectories from adolescence to emerging adulthood: relationships with family climate and self-concept in type 1 diabetes.
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. Several factors, including daily blood glucose monitoring frequency, daily insulin dose, and socio-demographic factors have shown associations with the distinct HbA1c trajectories [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. For example, Helgeson et al. reported that answers from 132 adolescents on a child/adolescent-reported Self-Care Inventory questionnaire, including topics on adherence to diabetes-specific duties, could distinguish the least favorable HbA1c trajectory from the “stable on target” group over 11 years [
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. However, caregivers' perspectives about adherence were not investigated, and there are indications that they may be more strongly correlated with HbA1c than child/adolescent-reported adherence [
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
], caregiver- and/or child/adolescent-reported questionnaires offer a more collaborative approach to measuring adherence to a wider array of diabetes management activities [
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
] was developed to better measure and address self-care behavior while accounting for different informants' unique observations. In a Danish cross-sectional setting, lower ADQ scores were significantly and negatively correlated with glycemic control [
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
], our study's aims were: 1) to identify distinct 11-year HbA1c trajectories in children/adolescents with type 1 diabetes mellitus and 2) to determine whether caregiver- and/or child/adolescent-reported adherence predict HbA1c trajectory membership while adjusting for sex, diabetes-specific and socio-demographic factors. Our hypotheses were that: 1) adherence/self-care scores, as measured by the caregiver- or child/adolescent-report ADQ, would be predictive for HbA1c trajectories over an 11-year period, and 2) covariates such as sex, age at diabetes diagnosis, diabetes duration, highest caregiver education level, and family structure may predict HbA1c trajectories.
2. Methods
2.1 Study design and participants
The data for the present longitudinal study are from a population-based cohort, including all eligible Danish children diagnosed with type 1 diabetes mellitus in 2009 in the age group of 2–17 years [
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
]. In short, a nationwide web-based survey was conducted in 2009 with a focus on psychosocial conditions, including adherence as measured by the ADQ. The survey return rate was 60% and participants provided a blood sample for HbA1c [
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
]. The present study comprises the subpopulation of children and adolescents aged 10–17 years, who had a diabetes duration of more than one year, a baseline HbA1c value, socio-demographic data, and completed the ADQ, and whose parents had also completed the ADQ. In total, 671 patients from the 2009 cohort were included (Fig. 1). HbA1c data from 2010 to 2020 were obtained from the Danish Diabetes Databases (see below).
Fig. 1Flow chart for the inclusion of the patients.
Adherence to a diabetes treatment plan (e.g., dietary considerations, carbohydrate counting, insulin use, blood glucose monitoring, prevention of hypoglycemia or hyperglycemia, physical exercise) was the main predictor variable, and it was assessed using the ADQ, which has good psychometric and readability properties [
]. The caregiver who was primarily involved in the child's daily diabetes treatment and the children/adolescents completed 17 items (insulin pen treatment group) or 19 items (insulin pump treatment group) [
]. One example of an ADQ question included: “How did your child (possibly with your help) within the preceding month handle his/her diabetes care in relation to… taking his/her insulin every time he/she eats?” Item responses were rated on a 5-point Likert scale from 1 to 5 and scored by calculating the mean of all items. Higher total or mean scores indicate higher/better adherence. For the present study, the ADQ demonstrated good internal consistency, with Cronbach's alpha ranging from 0.85 (insulin pump ADQ form) to 0.90 (insulin pen ADQ form) using caregivers' responses, and 0.83 (insulin pump ADQ form) to 0.87 (insulin pen ADQ form) using child/adolescents' responses.
2.3 Predictor variables – clinical and demographic variables
Sex, age at diabetes diagnosis, and diabetes duration were obtained at the time of answering the ADQ from caregivers/participants. Socio-demographic variables were obtained from registers held by Statistics Denmark and were measured in 2009, the same year the ADQ was completed. Family/household structure comprises all persons living at the same address, regardless of their mutual relations but was categorized into single versus two-parent families. Caregivers' educational status was grouped into low, medium and high categories by using the International Standard Classification of Education system [
]. HbA1c data were tracked in the DanDiabKids database to the age of eighteen, and thereafter in the DADD database. HbA1c values under illogical thresholds of 4.9% (30 mmol/mol) or over 20.4% (200 mmol/mol) were dropped. All reported HbA1c values were measured in accordance with the International Federation of Clinical Chemistry (IFCC) in mmol/mol [
IFCC reference system for measurement of hemoglobin A1c in human blood and the national standardization schemes in the United States, Japan, and Sweden: a method-comparison study.
]. Corresponding HbA1c in National Glycohemoglobin Standardization Program (NGSP) units as a % is reported.
2.5 Statistical analysis
Participant characteristics, ADQ results, and HbA1c were examined using descriptive statistics. Baseline and outcome variable comparisons to non-participants were examined using independent t-tests. A sample of 671 patients with 5644 HbA1c observations was analyzed using GBTM. Due to smaller longitudinal sample sizes at ages 10, 11, and 28 years, our GBTM analysis focuses on measurements made at ages 12–27 years. As originally described by Nagin [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
], the Stata command “traj”, a GBTM procedure, was used to identify subgroups within our longitudinal dataset that followed distinct trajectories for HbA1c. GBTM is a semi-parametric technique applied to longitudinal data to define groups that follow similar patterns for a specified outcome variable over time [
]. Each group's trajectory was examined with cubic and quadratic trends. When the cubic parameter lacked significance, the quadratic parameter was examined instead. If the quadratic parameter lacked significance, a linear model was used. In line with previous studies [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
], the optimal number of groups with homogenous clusters of trajectories was based on: 1) model fit assessed by the Bayes Information Criterion (BIC) whereby a larger percent decrease in BIC with each change in the model signifies an improvement in model fit, 2) visual inspection of the data for all participants and the meaningfulness (i.e., clinical context) of the distinctions in the trajectories, and 3) a minimum percent membership in each group of ≥5%.
Regarding labeling of each group, the first part of the label represents the trajectory's HbA1c starting point. The second part describes the trajectory over time. After the optimal number of groups was selected, we explored the relationship between probabilistic group membership and the ADQ, adjusted for sex, age at diabetes diagnosis, diabetes duration, highest caregiver educational level, and family structure. Two models were used in our analyses because our goal was to determine the predictive value of either the caregiver or child/adolescent-reported ADQ in relation to HbA1c trajectories. Multinomial logistic regression models were used to assess which variables predicted HbA1c trajectory membership [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. The reported estimates are changes in log odds ratio of being in one trajectory versus a reference trajectory, given a one-unit increase in the risk factor i.e., similar to regression coefficients [
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. Coefficients <0 indicate a decreased probability of membership in the trajectory group under investigation compared to the reference group as the risk factor increases, and vice-versa for coefficients >0. P- values <0.05 were considered statistically significant. Statistical analysis was performed with Stata 17 software.
2.6 Ethics
The study was approved by the Danish Society of Childhood and Adolescent Diabetes, and by the Regional and Danish Ethical Committees under paragraph 10. The study was registered under Aarhus University's journal reference number 2016–051-000001.
3. Results
3.1 Baseline participant characteristics
We analyzed data from 671 children/adolescents (53% females) (Fig. 1) with type 1 diabetes mellitus who were followed 11 years from childhood/adolescence into young adulthood, and from whom both the caregiver- and child/adolescent-reported ADQ were completed in 2009. Baseline characteristics (Table 1) were as follows: mean (SD) age was 14.4 (2.2) years with a range of 10–17 years; the mean age at diabetes diagnosis was 8.5 (3.5) years; the mean diabetes duration was 5.9 (3.3) years. Regarding caregivers' characteristics, 11.8% (n = 79) had a low, 46.9% (n = 315) had a medium, and 42.3% (n = 277) had a high education level. Regarding family structure, 15% (n = 101) of participants came from a single-parent family (Table 1). The mean (SD) caregiver reported ADQ score was 4.00 (0.59) and mean child/adolescent reported ADQ score was 4.00 (0.56) on a Likert scale from 1 to 5. Baseline correlations included: r = 0.612 for overall mean caregiver-reported ADQ and mean child/adolescent-reported ADQ (p < 0.0001); r = −0.40 for mean caregiver-reported ADQ and HbA1c (p < 0.0001); r = −0.37 for mean child/adolescent-reported ADQ and HbA1c (p < 0.0001).
Table 1Baseline study and longitudinal patient characteristics (means and standard deviations (SD) or number (N), percent) for the whole sample and for probabilistic membership to HbA1c trajectory groups. ADQ = Adherence in Diabetes Questionnaire. *The data under the columns for trajectory groups 1, 2, 3, and 4 represent latent strata (i.e., they are groups of individuals who followed the same course over time). † The ADQ score was rated on a five-point Likert scale.
Over 99.4% of participants contributed three or more HbA1c values during the 11-year period. The mean number of HbA1c observations per participant was 8.4 with a median of 9. The longitudinal number of HbA1c values was weakly but significantly correlated with ADQ scores (r = 0.10, p < 0.0001 for caregiver-reported and r = 0.15, p < 0.001 for child/adolescent-reported). In total, the 671 participants had 5644 HbA1c measurements with a mean HbA1c of 8.3 (1.4)% (67 [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
] mmol/mol) (Table 1). The HbA1c values were fairly evenly distributed across time with a maximum and minimum number of observations of 528 (in 2011) and 396 (in 2010), and an average of 470 observations per calendar year.
The improvement in the BIC values was large (11%) going from a 1-trajectory solution to a 2-trajectory solution and a further 3% reduction was obtained for a 3-trajectory solution. However, the improvement going from 4-trajectory solution to a 5-trajectory solution was less than 1% and smaller for even more elaborate models. The four- and five-group models had the most optimal combination of a low BIC and a minimum percent membership in each group of ≥5%. Both the four- and five-group models were clinically meaningful, but the four-group model (Fig. 2) was chosen because of its model fit, its simplicity, and a statistically more acceptable number of individuals in each age-interval for all groups. The mean and SD for the number of HbA1c values in participants over time from 2009 to 2020 was similar between group 1 (8.74 [1.37]), group 2, (8.45 [1.61]), group 3 (8.13 [1.88]), and group 4 (7.98 [1.80]).
Fig. 2HbA1c versus time trajectories spanning over childhood, adolescence and young adulthood. Group 1 “on target, gradual decrease” (blue), Group 2 “above target, mild increase then decrease” (red), Group 3 “above target, moderate increase then decrease” (green), and Group 4 “well above target, large increase then decrease” (orange). Number of HbA1c observations = 5644 for 671 participants. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
] mmol/mol) and a significant linear trajectory, which remained fairly stable while gradually decreasing below the general target HbA1c of 7.0% (53 mmol/mol) by 24 years. Therefore, group 1 was made our reference group. HbA1c group 2 “above target, mild increase then decrease” was the largest group (n = 269), had a mean longitudinal HbA1c of 8.1 (0.9)% (65 [
Longitudinal trajectories of metabolic control across adolescence: associations with parental involvement, adolescents’ psychosocial maturity, and health care utilization.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
] mmol/mol) at 27 years. Group 2's HbA1c starting points were the same as group 3 at age 12, but then diverged in their trajectories. Group 2 remained above target throughout adolescence but was nearly on target in the mid-twenties while group 3 was well above target throughout most of adolescence and young adulthood (Fig. 2).
3.3 Predictors of membership to the HbA1c trajectories
As reported in Table 2, lower caregiver-reported ADQ scores significantly predicted membership to groups 2, 3, and 4, compared to group 1 (comparison group) (p = 0.04, p < 0.0001, p < 0.0001, respectively). Male sex and age at diabetes diagnosis did not distinguish membership to any group. Longer diabetes duration significantly predicted membership to group 2 or 3 (p = 0.039 and p = 0.034, respectively). Single-parent family structure significantly predicted membership to group 4 (p = 0.005). Low caregiver education level (as compared to high) predicted membership to group 3 or 4 (p = 0.0002 and p = 0.002, respectively).
Table 2Adherence in Diabetes Questionnaire (ADQ), sex, and age at diabetes diagnosis, diabetes duration, family structure, and caregiver education level of HbA1c trajectory group membership. Numbers presented in this table indicate the changes in log odds ratio of being in the HbA1c trajectory group versus the reference or comparison group 1, given a one-unit increase in the risk factor.
Prediction models
Predictive measures at baseline
Coefficients / maximum likelihood estimates from group-based trajectory modeling (standard error), P value
In the model for children/adolescents (Table 2), lower child/adolescent-reported ADQ scores predicted membership to groups 3 and 4 (p = 0.027 and p = 0.0004, respectively). Male sex (as compared to female sex) did predict membership to group 2. Age at diabetes diagnosis did not distinguish membership to any group. Longer diabetes duration predicted group 2 or 3 membership (p = 0.028 and p = 0.039, respectively). Single-parent family status predicted group 4 membership (p = 0.004). Low caregiver education level (as compared to high) predicted group 3 or 4 membership (p = 0.0003 and p = 0.002, respectively).
3.4 Selection bias: comparisons to non-participants
The mean age (14.4 years) of participants (n = 671) when completing the ADQ was significantly lower than the mean age (15.0 years) of non-participants 10 years or older (n = 771), p < 0.0001. The percentage of male participants (47%) was somewhat lower than the percentage of male non-participants (54%), p = 0.01. At baseline, the percentage of participants with single-parent families and caregivers with a low education level were lower compared to non-participants (15% vs 29%, p < 0.0001; 12% vs 22%, p < 0.0001, respectively). The mean (SD) baseline HbA1c of 8.1 (1.2)% (65 [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
] mmol/mol) for 7618 HbA1c values in non-participants (p < 0.0001).
4. Discussion
Using the National Danish cohort of children and adolescents with type 1 diabetes mellitus from 2009 and registry-based follow up data from 2010 to 2020, we identified four distinct HbA1c trajectories: 1) “on target, gradual decrease”, 2) “above target, mild increase then decrease”, 3) “above target, moderate increase then decrease”, and 4) “well above target, large increase then decrease”. We found that higher caregiver-reported ADQ scores predicted the most favorable HbA1c trajectory, and correspondingly, lower caregiver-reported ADQ scores predicted less favorable HbA1c trajectories. Lower scores on the child/adolescent-reported ADQ were predictive of membership to the two higher risk HbA1c trajectories. Low caregiver education level predicted group 3 and 4 membership while single-parent family status predicted group 4 membership.
4.1 HbA1c trajectories
In 2009, the ages of our population-based/national cohort sample were 10–17 years and the participants were followed prospectively until 20–27 years. Comparable studies include two multi-country registry-based studies, which identified five distinct trajectories in 6433 children/adolescents with type 1 diabetes followed for 10 years from age 8 to 19 years [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Five heterogeneous HbA1c trajectories from childhood to adulthood in youth with type 1 diabetes from three different continents: a group-based modeling approach.
]. Both of these two studies followed the participants to the age of 18 years. Helgeson et al. also identified five HbA1c trajectories over an 11-year period in a smaller sample of 132 younger U.S. adolescents (11–13 years at baseline) with type 1 diabetes into early adulthood [
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. When comparing our HbA1c trajectories with those of the other longitudinal studies with 10 or more years of follow up, the forms of the trajectories vary depending on the country, the population being investigated, the baseline age range, and, particularly, the ending age range. Common features were found among these studies. Just like Schwandt et al.'s findings about the “intermediate stable” and “intermediate increase” groups [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
], we also found that individuals with membership to two different groups (our groups 2 and 3) began with similar HbA1c levels but then, their glycemic control diverged over time with group 2 increasing to a maximal HbA1c level of 8.5% (69 mmol/mol), but group 3 increasing to a maximal HbA1c level of 9.9% (84 mmol/mol). Our results, along with those of the other longitudinal studies with longer follow up periods, collectively illustrate that moderate- and high-risk groups had increasing HbA1c levels throughout adolescence (up to 19 years) [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
Five heterogeneous HbA1c trajectories from childhood to adulthood in youth with type 1 diabetes from three different continents: a group-based modeling approach.
]. Thereafter, our study and Helgeson et al.'s study (which extended HbA1c trajectories into the mid-twenties) showed decreasing HbA1c trajectories after age 22 [
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. Our study is in accordance with previous longtime follow up studies showing collectively that there is a large increase in mean HbA1c levels throughout adolescence in a type 1 diabetes population [
]. However, this elevation in mean HbA1c levels should not be assumed to be the case for all adolescents. Instead, adolescents can be differentiated into optimal (low-risk) and less optimal (including moderate-risk and high-risk) HbA1c.trajectory groups.
4.2 Predictors of membership to HbA1c trajectories
No previous longitudinal study has used both a caregiver- and/or child/adolescent-reported adherence/self-care questionnaire as a predictor for glycemic control during adolescence and young adulthood. For the caregiver-reported model, we found that lower ADQ scores predicted membership to the moderate and high-risk HbA1c trajectories (groups 2, 3, and 4) over 11 years. For the child/adolescent-reported model, lower ADQ scores predicted membership to the two high-risk HbA1c trajectories (groups 3 and 4). In accordance with our study, Helgeson et al. showed that lower scores on a fourteen-item Self-Care Inventory completed by children/adolescents at ages 11–13 years, significantly distinguished membership to the most unfavorable HbA1c trajectory compared to their reference “stabile on target” trajectory. The inability of lower child/adolescent-reported diabetes self-care scores to distinguish membership to three other unfavorable HbA1c trajectories may be due to the smaller sample size of the population studied.
In both the caregiver- and child/adolescent-reported model, socio-demographic risk factors significantly predicted membership to high-risk HbA1c trajectories (i.e., low caregiver education level predicted membership to groups 3 and 4; single-parent status predicted membership to group 4). Helgeson et al. reported that “social status” and “household structure” distinguished HbA1c trajectories; however, the authors did not report any statistical associations between baseline socio-demographic variables and likely membership to specific HbA1c trajectories [
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
]. In a study with HbA1c trajectories spanning over two years, Hilliard et al. found that unmarried caregiver status predicted membership to the HbA1c trajectory subgroup with the least optimal glycemic control; however, in this study of 150 participants, caregiver education status did not significantly distinguish group membership [
]. Our national, population-based data suggests growing up in a home with a single-parent or less educated parent complicates adherence to diabetes management and elevates the risk of suboptimal glycemic control.
Sex seems to have less value in regard to predicting HbA1c trajectory membership in adolescents. In accordance, three studies, including two larger multi-country studies, which did not control for ADQ scores, found sex did not distinguish between HbA1c trajectories [
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
Five heterogeneous HbA1c trajectories from childhood to adulthood in youth with type 1 diabetes from three different continents: a group-based modeling approach.
We found age at diabetes onset did not distinguish HbA1c trajectory membership in a sample that was 10–17 years old at baseline; however, we observed that longer diabetes duration significantly predicted group 2 and 3 membership. This last observation was in accordance with Hilliard et al. (24), but in contrast to the study of Rohan et al. [
Our study has several strengths. First, our cohort comprised a large national sample with complete data concerning all of the baseline variables, including the caregiver- and child/adolescent-reported ADQ [
]. Second, the cohort's HbA1c levels were tracked over an 11-year period, and this included the critically important periods of adolescence and young adulthood [
]. Fourth, we have investigated associations using GBTM, which has the advantage of presenting an outcome over time in an easily understood graphical summary [
]. Finally, we adjusted our models for sex, diabetes-specific variables, and socio-demographic factors.
Our study also has limitations. We were able to compare survey participants with non-participants and identified selection bias. The HbA1c levels of survey non-participants were significantly higher than those of survey participants at baseline [
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
] and longitudinally. This finding suggests that non-respondents may have had less optimal adherence to an agreed upon plan of care than our study's participants. Despite this limitation, we found robust associations between caregiver-reported ADQ scores and the HbA1c trajectories while adjusting for socio-demographic and diabetes-specific clinical covariates. Another consideration is that, during the period of 2010 to 2020, the generally recommended target for HbA1c in Denmark changed from 58 mmol/mol (7.5%) [
ISPAD clinical practice consensus guidelines 2018: Glycemic control targets and glucose monitoring for children, adolescents, and young adults with diabetes.
]. Over that same time period, the recommended use of diabetes technology – insulin pumps and continuous blood glucose monitoring systems – increased markedly [
]. Changing recommendations and treatment modalities over the 11-year period could have influenced the form of the HbA1c trajectories.
4.4 Conclusions
We identified four HbA1c trajectories in a national cohort of Danish children and adolescents with type 1 diabetes mellitus over the course of a decade. The majority of participants (73%) had membership to one of three less favorable HbA1c trajectories. Both lower caregiver- and child/adolescent-reported ADQ scores, and low caregiver education and single-parent status, significantly predicted unfavorable HbA1c trajectories across adolescence and young adulthood. Thus, ADQ scores and socio-demographic information may serve as tools for diabetes care providers for the identification of children and adolescents, who may need more support from their multidisciplinary diabetes team. Future research is needed to investigate whether discussion of ADQ scores, and additional support from their multidisciplinary diabetes care team, if needed, contributes to more optimal HbA1c trajectories.
Funding
KPM's PhD scholarship and work was supported by a research grant from the Danish Diabetes and Endocrine Academy (DDEA), which is funded by the Novo Nordisk Foundation, grant number NNF17SA0031406, with additional support from Steno Diabetes Center Aarhus. NHB received a research grant from Poul and Erna Sehested's Fond.
Contribution statement
All authors contributed to the design of the study and collection of data. KPM was responsible for registry-based data management. MBJ was responsible for the statistical analysis. KPM, MT and NHB wrote the manuscript. All authors critically revised the manuscript and approved the final version.
Data availability
MT, MBJ and NHB are the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
The authors extend their gratitude to all the health care providers at the pediatric diabetes units throughout Denmark who supported this research and are most grateful to all the participants and their families who took the time to participate in our survey. A different abstract related to this same study has been presented as an oral presentation at the European Association for the Study of Diabetes (EASD) 58th Annual Meeting in Stockholm, September 20-23, 2022 [
Writing Team for the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group
Sustained effect of intensive treatment of type 1 diabetes mellitus on development and progression of diabetic nephropathy: the epidemiology of diabetes interventions and complications (EDIC) study.
Exploring variation in glycemic control across and within eight high-income countries: a cross-sectional analysis of 64,666 children and adolescents with type 1 diabetes.
ISPAD clinical practice consensus guidelines 2018: Glycemic control targets and glucose monitoring for children, adolescents, and young adults with diabetes.
Longitudinal trajectories of metabolic control across adolescence: associations with parental involvement, adolescents’ psychosocial maturity, and health care utilization.
Continuity and change in glycemic control trajectories from adolescence to emerging adulthood: relationships with family climate and self-concept in type 1 diabetes.
Longitudinal trajectories of metabolic control from childhood to young adulthood in type 1 diabetes from a large German/Austrian registry: a group-based modeling approach.
Featured article: trajectories of glycemic control over adolescence and emerging adulthood: an 11-year longitudinal study of youth with type 1 diabetes.
Multi-informant path models of the influence of psychosocial and treatment-related variables on adherence and metabolic control in adolescents with type 1 diabetes mellitus.
IFCC reference system for measurement of hemoglobin A1c in human blood and the national standardization schemes in the United States, Japan, and Sweden: a method-comparison study.
Five heterogeneous HbA1c trajectories from childhood to adulthood in youth with type 1 diabetes from three different continents: a group-based modeling approach.
00033-5&id=gr1.jpg){kind=link}
00033-5&id=gr2.jpg){kind=link}