Detangling Associations Between Maternal Depressive Symptoms and Diabetes Relationship Distress With Adolescents’ HbA1c

Abstract

Objective 

Previous research in families of children with type 1 diabetes demonstrates that maternal depressive symptoms are a known risk factor for poor diabetes outcomes. We sought to examine whether maternal diabetes relationship distress or maternal depressive symptoms were more strongly associated with adolescent glycemic outcomes.

Methods 

Analyses were conducted using data from mothers who consented to screen for a behavioral intervention. The screener included the Patient Health Questionnaire and the Parent Diabetes Distress Scale, Parent/Teen Relationship Distress subscale. Hemoglobin A1c (HbA1c) was extracted from adolescents’ medical records.

Results 

Our sample consisted of 390 maternal caregivers of adolescents with type 1 diabetes aged 11–17. Screening data revealed that 35% of mothers reported clinically significant diabetes distress related to their relationship with their adolescents, and 14% of mothers reported clinically significant depressive symptoms. The adolescents of mothers who reported diabetes relationship distress had significantly higher mean HbA1c levels (9.7 ± 2.2%) compared to those whose mothers were not distressed (8.2 ± 1.8%, d = .72). Similarly, adolescents whose mothers reported clinically significant depressive symptoms had higher mean HbA1c levels (9.6 ± 2.4%) than those whose mothers were not depressed (8.6 ± 2.0%, d = .48). After adjusting for clinical and demographic factors, mothers’ reports of diabetes relationship distress were more strongly associated with adolescents’ HbA1c than maternal depressive symptoms.

Conclusions 

Our findings suggest that screening for maternal distress—particularly distress related to the caregiver–adolescent relationship—could match families with psychosocial support or other resources to improve both psychosocial and glycemic outcomes.

Introduction

Type 1 diabetes (T1D) is one of the most common chronic health conditions in youth, and T1D prevalence continues to increase among people 19 years or younger (Lawrence et al., 2021). The recommended treatment plan for T1D requires consistently monitoring blood glucose, carefully tracking carbohydrate intake, administrating insulin for every meal, and treating low and high blood glucose levels (ElSayed et al., 2022). In one of the first studies investigating the role of parental involvement in T1D management, Anderson et al. (2002) demonstrated that parents are primarily responsible for their children’s diabetes care, and around ages 13–17, management responsibilities begin transitioning from caregiver to adolescent. Positive parental involvement in diabetes management during adolescence predicts better glycemic outcomes and better adolescent quality of life (Cameron et al., 2007; Wysocki et al., 2009; Williams et al., 2020). However, research suggests that parental distress is a common hindrance to effective communication between caregivers and their adolescent children (Jaser & Grey, 2010; Whittemore et al., 2012). While earlier studies of parental distress focused on symptoms of anxiety and depression (Whittemore et al., 2012), attention has shifted more recently to the experience of diabetes distress or the emotional burden of caring for a child with diabetes (Hessler et al., 2016). High levels of diabetes distress and depressive symptoms in caregivers of adolescents with T1D may interfere with positive caregiver–adolescent relationships, putting adolescents at risk for problems with diabetes management (Hessler et al., 2016; Whittemore et al., 2012).

Symptoms of general life stress, depressive symptoms, and diabetes distress often overlap, but there is empirical support that these symptoms are conceptually distinct and should be measured and targeted separately (Fisher et al., 2014; Hessler et al., 2016). Caregivers often experience stress when relinquishing control and giving their adolescent more responsibility for their diabetes care, and this stress can be exacerbated when their child does not meet glycemic targets (HbA1c <7.0%). Caregivers who have difficulty coping with the stress of diabetes management may feel overwhelmed and helpless about achieving glycemic targets and building their adolescents’ diabetes management skills, but these feelings may not be captured by measures of depression or anxiety (Cameron et al., 2007; Streisand et al., 2005; Whittemore et al., 2012). Hessler and colleagues (2016) created the Parent Diabetes Distress Scale to assess the emotional burden experienced by caregivers of children with T1D. Their validation study, with 332 parents of youth with T1D (ages 11–21), supported four factors of parental diabetes distress: Parent/Teen Relationship concerns around conflicts and disagreements with their child; Personal Distress about the impact of diabetes on their quality of life; Teen Management Distress about concerns around their adolescents’ ability to manage; and Healthcare Team Distress related to the adequacy of their adolescents’ diabetes care. The total score and all four individual factors were significantly associated with parents’ general life stress and depressive symptoms, supporting the validity of using the Parent/Teen Relationship subscale to measure diabetes distress as a distinct measure.

Differences Between Caregiver Depressive Symptoms and Diabetes Distress

Extensive literature highlights parental psychological symptoms as a risk factor for problems with diabetes management (Whittemore et al., 2012). In families of children with T1D, maternal depression is associated with higher family conflict, reduced family cohesion, less consistency in encouraging discipline, and lack of warmth and adaptability in their parent–child relationship (Eckshtain et al., 2010; Mackey et al., 2014; Whittemore et al., 2012). Further, research identified links between diabetes distress in parents and adolescents’ glycemic outcomes (Whittemore et al., 2012). In contrast to parental depression, characterized by general disinterest and low mood, diabetes distress captures parents’ worries about their child’s ability to self-manage diabetes, the impact on the parent’s personal quality of life, disagreements with their adolescent on diabetes-related responsibilities (diabetes relationship distress), and the quality of healthcare available to their adolescent (Hessler et al., 2016; Shapiro et al., 2018). The transition of responsibility for diabetes management from parents to adolescents is often challenging, which can contribute to diabetes relationship distress in this population (Young et al., 2014). Few studies have assessed both depressive symptoms and diabetes distress in parents of adolescents with T1D, and findings are mixed: a large study of parents of youth ages 11–21 years (n = 332) found that the correlation between adolescents’ HbA1c and parental diabetes distress (r = .42) was higher than the correlation between HbA1c and general life stress (r = .17) or depressive symptoms (r = .13), but this study was limited by using parental report of HbA1c (Hessler et al., 2016). In a smaller study of mothers of children ages 10–16 (n = 81), mothers’ depressive symptoms were more strongly correlated with adolescents’ HbA1c (extracted from the medical record) (r = .35) than mothers’ diabetes distress (r = .16) (Rumburg et al., 2015).

Current Study

We sought to examine whether maternal diabetes relationship distress or maternal depressive symptoms were more strongly associated with adolescents’ HbA1c, to inform future screening initiatives and identify families who could benefit from additional support. The current study focused on maternal caregivers because several studies confirm that mothers are more likely to be primarily involved in diabetes care, and mothers typically experience higher rates of distress than fathers (Landolt et al., 2005; Whittemore et al., 2012).

Methods

The current study is a secondary analysis of data from mothers who consented to screen for an ongoing randomized controlled trial (NCT03818711) of a novel behavioral intervention to improve diabetes-related communication and coping in mothers and their adolescent children with T1D (Jaser et al., 2019). Mothers were recruited from a pediatric diabetes clinic in an academic medical center. Research staff approached families either at their diabetes clinic visit or over the phone within 2 weeks of their scheduled clinic appointment. Female caregivers were approached if they met the following criteria, based on the child’s medical record: (1) both child and parent were English speaking, (2) the child was age 11–17, and (3) the child had been diagnosed with T1D for at least 12 months. Of the mothers approached, 43% (478 out of 1118) consented to complete a screening survey. Mothers completed the screening survey prior to enrollment in the behavioral intervention.

The current analysis of the screening data includes a sample of 390 female caregivers who completed the screener and whose child had an HbA1c value within 4 weeks of the screening survey completion date. If a mother screened multiple times for the RCT, the most recent screening record was used in the analyses. The adolescent’s demographic and glycemic data were extracted from the electronic medical records. Screening data for this behavioral intervention were collected from 2019 to 2022, and 27% were collected prior to March 11, 2020, when the WHO declared COVID-19 a pandemic (Cucinotta & Vanelli, 2020).

Measures

The Patient Health Questionnaire-9 (PHQ-9) is used to measure depressive symptoms concisely and reliably (Kroenke et al., 2001). The PHQ-9 measures the frequency of nine depressive symptoms over the past 2 weeks on a scale of 0 (Not at All), 1 (Several Days), 2 (More than Half the Days), and 3 (Nearly Every Day). Total scores between 5 and 9 on the PHQ-9 indicate mild depressive symptoms and a score of 10 or higher indicates risk for major depressive disorder (Kroenke et al., 2001). In the current study, Cronbach’s alpha was 0.87.

The Parent Diabetes Distress Scale (DDS-P) is a validated measure used to assess diabetes distress in parents of adolescents and young adults and includes four subscales: (1) Personal, (2) Teen Management, (3) Parent/Teen Relationship, and (4) Healthcare Team Distress (Hessler et al., 2016). The DDS-P Parent/Teen Relationship subscale was selected for this screening process, given that the intervention was aimed at improving parent–child communication around diabetes management. Eight items assessed how often mothers’ diabetes-related disagreements obstruct the parent–child relationship on a scale of 0 (Not at all) to 4 (A great deal) (e.g., Worrying that my nagging about diabetes is hurting my relationship with my teen and Feeling that trying to help my teen with his/her diabetes is always a battle). A mean score of 2 or higher indicates clinically significant distress (Hessler et al., 2016). In the current study, Cronbach’s alpha for the Parent/Teen Relationship scale was 0.95.

HbA1c values were primarily extracted from the adolescents’ medical records using the point-of-care value from the clinic visit closest to the screening, if available. If a family consented to screen and enrolled in the study, missing HbA1c values were replaced with either a complete at-home HbA1c kit (n = 16, or 4% of the sample) or an estimated Glucose Management Indicator (GMI) from the child’s diabetes device (n = 7, or 2% of the sample). These two alternative HbA1c collection methods are comparable to the point-of-care visit HbA1c measure (Bergenstal et al., 2018; CoreMedica, 2020).

Analysis Plan

SPSS version 28.0 was used to analyze data. First, descriptive analyses were conducted to identify whether demographic (adolescent age, sex, race, and ethnicity) or clinical factors (duration of diabetes) were associated with maternal depressive symptoms and diabetes relationship distress. Next, bivariate correlations were conducted to evaluate the associations between maternal depressive symptoms, diabetes relationship distress, and HbA1c. To determine the relative strength of these associations, we used an interactive calculator to obtain a z-score (Lee & Preacher, 2013). In addition, we conducted an analysis of variance to assess whether there was a cumulative effect of clinically significant scores on adolescents’ HbA1c. Finally, multivariable linear regression analyses were conducted to determine the relative contribution of depressive symptoms and diabetes relationship distress to HbA1c, adjusting for child age, sex, race and ethnicity (dichotomized as non-Hispanic White/other race or ethnicity), and duration of diabetes. Based on findings from an earlier study indicating age-related differences (i.e., the association between diabetes relationship distress and HbA1c was stronger for caregivers of younger adolescents than for caregivers of older adolescents) (Rumburg et al., 2015), we also tested for a moderating effect of age on the relationship between diabetes relationship distress and maternal depressive symptoms with HbA1c.

Results

Participants included 390 maternal caregivers of adolescents with T1D (51% female, mean age = 14 ± 2.1 years, 2% Hispanic or Latinx, 76% White, 12% Black/African American, 2% Other, 10% race or ethnicity not indicated). The mean duration of diabetes was 5.6 ± 3.5 years, and the mean HbA1c was 8.7 ± 2.1%. Only 18% of adolescents were meeting the target HbA1c (<7%).

Demographic and Clinical Differences in PHQ-9 and Diabetes Relationship Distress Scores

As reported in Table 1, child race and ethnicity and duration of diagnosis were not significantly related to diabetes relationship distress or maternal depressive symptoms. Adolescent age was negatively correlated with PHQ-9 scores, indicating that mothers of younger children reported higher depressive symptoms, but there was no association between child age and diabetes relationship distress. There were no significant differences in PHQ-9 or DDS-P relationship subscale scores in mothers who completed the screener before or after the start of the COVID-19 pandemic (March 11, 2020).

Table I.

Descriptive Statistics and Correlations for Study Variables

Variable	M	SD	1	2	3	4	5	6
1. Child race/ethnicity	n/a	n/a	–
2. Child age	14.1	2.1	−.09	–
3. Diabetes duration	66.6	42.6	.02	.32***	−
4. PHQ-9	4.7	4.6	.08	−.11*	−.00	–
5. P-DDS	1.5	1.1	.06	−.01	.07	.42***	–
6. HbA1c	8.7	2.1	.27***	.07	.09	.16**	.36***	–

Note. Child race/ethnicity is dichotomized as non-Hispanic White or other race or ethnicity. Diabetes duration is in months. PQH-9 = Patient Health Questionnaire; PDD-S = Parent Diabetes Distress Scale, Parent–Teen Relationship subscale.

^* p < 0.05.

^** p < 0.01.

^*** p < .001.

In our sample, 61% (n = 237) of mothers did not score above the clinical cutoff on either measure; 35% (n = 137) of mothers reported clinically significant diabetes relationship distress; 39% (n = 151) of mothers reported at least mild PHQ-9 depressive symptoms (PHQ-9 score ≥5); and 14% (n = 56) of mothers reported clinically significant depressive symptoms (PHQ-9 score ≥10). Of the mothers who reported clinically significant relationship distress, 29% (40/137) also reported clinically significant depressive symptoms (see Figure 1).

Figure 1.

Maternal caregivers’ reports of distress. DDS-P = Parent Diabetes Distress Scale, Parent–Teen Relationship Subscale; PHQ-9 = Patient Health Questionnaire; Clin. Sig. = Clinically Significant.

Maternal Distress and Adolescents’ HbA1c

The adolescents of mothers who reported clinically significant diabetes relationship distress scores had significantly higher mean HbA1c values (9.7 ± 2.2%) compared to adolescents of mothers with lower distress scores (8.2 ± 1.8%, t = -6.34, p < .001, d = .72). We also observed a significant difference in mean HbA1c levels between the adolescents of mothers who reported clinically significant depressive symptoms (9.6 ± 2.4%) and minimal symptoms of depression (8.6 ± 2.0%, t = −3.23, p < .001; d = .48). In contrast, there was no significant difference in HbA1c levels between adolescents of mothers with mild depressive symptoms (scores 5–9) and those with minimal depressive symptoms (scores 0–4), (t = −.33, p = .74). In addition, the mean HbA1c for youth of mothers who scored in the clinically significant range for both measures (9.5 ± 2.1%) was not significantly different from the youth whose mothers reported clinically significant symptoms on only one of the measures (HbA1c = 9.7 ± 2.9%, p = .629).

As seen in Table I, adolescent HbA1c was significantly associated with both PHQ-9 scores (r = .16) and DDS-P Parent/Teen Relationship subscale scores (r = .36). Using the interactive tool to calculate the difference between two dependent correlations with one variable in common (Lee & Preacher, 2013), we found a z-score of −3.97 (p < .001). This indicates that the correlation between maternal diabetes relationship distress and HbA1c is significantly stronger than the correlation between maternal depressive symptoms and adolescents’ HbA1c. No significant difference was found in maternal diabetes relationship distress between mothers who learned their child’s HbA1c value prior to screening (mean score = 1.4 ± 1.1) and mothers who did not know their child’s HbA1c at the time of screening (mean score = 1.6 ± 1.1, t(387) = 1.78, p =.075). There was also no significant difference in mothers’ PHQ-9 scores related to knowledge of HbA1c values (mean score = 4.6 ± 4.5 for those who knew HbA1c vs. 4.6 ± 4.9 for those who did not know, t(387) = 0.26, p =.796).

Regression Analysis

As seen in Table II, a multivariable regression analysis revealed that the model predicting HbA1c was significant (F = 11.47, p < .001), explaining 19% of the variance in HbA1c levels. After adjusting for adolescent race and ethnicity, sex, age, and duration of diagnosis, maternal diabetes relationship distress was a significant predictor of HbA1c (β = .33, p < .001), but maternal depressive symptoms were not significant (β = .02, p = .689). Finally, the interaction term for age × PHQ-9 was not significant (β = −.12, p = .113), but there was a significant interaction for age × diabetes relationship distress (β = .18, p = .043). As seen in Figure 2, the association between diabetes relationship distress and HbA1c was stronger for teens (ages 13-17) than for early adolescents (ages 10–12).

Table II.

Regression Analysis Predicting Adolescents’ HbA1c From Demographic Factors and Maternal Distress

Predictor	B	SE B	β	ΔR2
Step 1				.09***
Child race/ethnicity	1.61	.30	.28***
Child age	.37	.23	.08
Child sex	−.11	.22	−.03
Duration of diabetes	.00	.00	.06
Step 2				.11***
Child race/ethnicity	1.49	.28	.26***
Child age	.37	.22	.08
Child sex	−.05	.21	−.01
Duration of diabetes	.00	.00	.04
PHQ-9	.01	.03	.02
DDS-P	.61	.10	.33***
Step 3				.01
Child race/ethnicity	1.48	.28	.26***
Child age	.39	.22	.09
Child sex	−.06	.21	−.02
Duration of diabetes	.00	.00	.04
PHQ-9	.06	.04	.12
DDS-P	.35	.17	.19*
Age × PHQ-9	−.37	.23	−.12
Age × DDS-P	.48	.24	.18*

Note. Child race/ethnicity is dichotomized as non-Hispanic White or other race or ethnicity. PQH-9 = Patient Health Questionnaire; PDD-S = Parent Diabetes Distress Scale, Parent–Teen Relationship subscale.

^* p < 0.05.

^** p < 0.01.

^*** p < .001.

Figure 2.

Moderating effect of adolescent age on the relationship between diabetes relationship distress and HbA1c.

Discussion

Maternal depressive symptoms are a known risk factor for poor diabetes outcomes, but only recently has parental diabetes distress been examined in relation to adolescents’ diabetes outcomes (Hessler et al., 2016). Findings from the current study highlight that many maternal caregivers experience emotional distress related to the burden of diabetes management, particularly within the caregiver–adolescent relationship, and that this diabetes relationship distress is strongly associated with adolescents’ HbA1c.

Caregivers face additional stress as children reach adolescence and take on more responsibility for their complex diabetes regimen. In our sample, younger child age was associated with higher maternal depressive symptoms, which may be due to the greater levels of responsibility for diabetes management among mothers of younger children. The high levels of stress related to taking care of children’s chronic illnesses are linked with greater symptoms of depression and anxiety in caregivers (Jaser et al., 2009; Streisand et al., 2005). Further, caregivers experiencing an overwhelming amount of diabetes distress may prematurely transfer diabetes care responsibility to their adolescent (Carcone et al., 2012), so it is pivotal to assess and treat caregiver distress and family functioning throughout the transition to adolescence (Anderson et al., 2002; Cameron et al., 2007; Carcone et al., 2012). An earlier study of these factors found that maternal depressive symptoms were more strongly associated with HbA1c than diabetes relationship distress (Rumburg et al., 2015), but this study had a smaller sample size (n = 81) and a higher percentage of mothers reporting clinically significant depressive symptoms (25% as compared to 14% in our sample), which may explain why our findings differ. It is also interesting to note that the timing of the adolescent’s HbA1c value was not related to mothers’ distress scores. Given that the majority of adolescents are now using continuous glucose monitoring devices which allow caregivers to see glucose data, many caregivers have objective information about the adolescents’ diabetes management before learning the HbA1c value. Thus, if the HbA1c value was above the target range, it is likely that glucose values were also frequently above the target range, which could contribute to parental distress at the time of data collection.

Contrary to previous findings, we did not observe a significant association between HbA1c and child age in our sample. This may be due to the fact that the maternal caregivers who completed the screener were interested in participating in a behavioral intervention because diabetes management was not going well, regardless of the child’s age. However, we did find that age was a significant moderator of the association between maternal diabetes relationship distress and HbA1c, such that the relationship between maternal distress and HbA1c was stronger for teens than for early adolescents. This could reflect increased conflict and distress around diabetes management as older adolescents take over responsibility for more diabetes care (Young et al., 2014).

Limitations

The current analysis is based on responses from maternal caregivers interested in participating in a behavioral intervention, so this sample of mothers may be experiencing higher stress about diabetes and/or be more motivated to improve diabetes management than the general population. Since these data were collected from the screening portion of a larger study, detailed information about socioeconomic status and family functioning was not available. In addition, our screening survey did not include questions about diabetes device use, and we were unable to investigate how modern diabetes technology and glucose data sharing may impact maternal distress. This sample is majority non-Hispanic White—but representative of the larger diabetes clinic—and future studies are needed to replicate findings in more diverse samples. It is possible that cultural differences exist related to reporting depressive symptoms and/or relationship distress (e.g., Steptoe et al., 2007). Further, we did not include male caregivers in the current study, and they likely play a distinct yet important role in diabetes management (Young et al., 2014). Finally, this data analysis was cross-sectional rather than longitudinal, and it is unknown whether challenges in meeting glycemic targets lead to additional distress or result from distress.

In recent years, progress has been made in screening youth with T1D for psychosocial outcomes (depressive symptoms, barriers to diabetes care, and disordered eating) (Young-Hyman et al., 2016). PHQ-9 screeners are increasingly administered in pediatric diabetes clinic visits (Mulvaney et al., 2021), which are an indispensable checkpoint for the child’s mental well-being, and it has been suggested that screening parents may also help to identify youth most in need of support (Cameron et al., 2007). While we acknowledge that screening for parental distress could be challenging if pediatric diabetes clinics do not have the resources to meet parents’ needs, the current study's findings indicate that screening caregivers for diabetes relationship distress may be particularly important. For example, using only the PHQ-9 to identify families in need of extra support would have excluded a substantial number of mothers in our sample who reported clinically significant diabetes relationship distress without clinically significant depressive symptoms. Clinically significant responses on the DDS-P relationship subscale could be used to initiate discussions between pediatric diabetes providers and caregiver–adolescent dyads or trigger a referral from social work on the type of support options available to caregivers, with the potential to match families with psychosocial support or other resources that could improve both psychosocial and glycemic outcomes.

Author Contributions

Fayo Abadula (Formal analysis [lead], Project administration [lead], Writing – original draft [lead]), Sydney Garretson (Data curation [supporting], Project administration [supporting], Writing – review & editing [supporting]), Nkemjika Okonkwo (Project administration [supporting], Resources [supporting], Writing – review & editing [supporting]), Lauren M. LeStourgeon (Data curation [lead], Project administration-supporting, Software [equal], Writing – review & editing-supporting), and Sarah Jaser (Conceptualization [lead], Formal analysis [supporting], Funding acquisition [lead], Investigation [equal], Methodology [equal], Writing – review & editing [equal])

Funding

This work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (R01DK115545).

Conflicts of interest: None declared.

Data Availability

Data available upon request.