Stressful Life Events, Parental Psychosocial Factors, and Glycemic Management in School-Aged Children During the One Year Follow-Up of New-Onset Type 1 Diabetes

Abstract

Objective:

To monitor occurrence of stressful life events, assess correlations with family functioning and parental psychosocial measures, and examine the impact of stressful life events on diabetes management in the first year after diagnosis of type 1 diabetes (T1D) in children using a mixed methods design.

Methods:

In a prospective study of 5-9-year-olds with recent-onset T1D (mean age 7.4±1.3 years, T1D duration 4.7±3.3 months), we monitored glycated hemoglobin A1c (HbA1c), income, job status, family health, and marital status at baseline and every 3 months up to 1 year. We measured coping, parental depression, and diabetes family conflict at baseline.

Results:

Of 128 families, 53.9% (n=69) reported 1+ stressful event, with 25.8% reporting income change (n=33) during this period, 23.4% additional family health changes (n=30), 22.7% job changes (n=29), 21.9% changes in child’s school (n=28), and 3.9% changes in marital status (n=5). Baseline active avoidance coping, parental depression, and diabetes family conflict correlated with a higher number of stressful life events (r=0.239, p<0.01; r=0.197, p<0.05; r=0.225, p<0.01, respectively). There were also cross-sectional associations between HbA1c and income decrease, school change, and job change at various time points in the study.

Conclusions:

Families can experience concurrent life stressors during the first year of T1D, which relate to coping, depression, and conflict. Consistent with existing literature, stressful life events relate to glycemic management. Future research should explore the individual’s or parent’s perception of stress and ways that diabetes centers can effectively assist families of youth with T1D and concurrent life stressors.

INTRODUCTION

Type 1 diabetes mellitus (T1D) is a life-altering diagnosis for both the individual and family affected, especially in school-aged children.¹ Incidence of T1D is highest in 10-14 year old individuals (25.9 per 100,000 person-years), followed by 5-9 year old individuals (22.1 per 100,000 person-years).² Although the incidence of T1D is high during ages 5-9 years,² research on parental psychosocial and family functioning in this group is scarce,^3,4 and no studies have examined how stressful family life events (e.g., change in income, attending a new school) may impact diabetes management in families of school-age children with recent-onset T1D, within the past year. Moreover, there is strong evidence that engagement with the T1D self-management regimen and family T1D-related behavior patterns are difficult to change after the first few years of T1D, indicating that prevention and intervention methods early in diagnosis could help with future T1D care and management.⁵

Development in early childhood is marked by physical and social growth, with potential changes in friendships, schools, schedules, and family dynamics. Children diagnosed with T1D in early childhood experience even more change due to the development of new routines with the introduction of insulin therapy, blood glucose monitoring, and alterations in diet and activity after diagnosis. Between the ages of 5-9 years, it is recommended that caregivers assume primary responsibility for their child’s T1D care.⁶ T1D caregivers often perceive burden related to long-term health concerns, medical treatment, and parental emotional distress.⁷ Therefore, it is important to monitor the relationships between psychological measures of the caregiver, diabetes self-management behaviors, and glycemic control in the child during the first few years of T1D diagnosis.

Stressful life events are experiences or changes that could cause or contribute to increased stress or alter one’s daily routine substantially. Examples of stressful life events include changes in family health, parental income, parental job, parental marital status, or child’s school. These stressful events are correlated with worsening health in general and poorer health outcomes.⁸ In T1D, research suggests that the occurrence of these stressful life events can interfere with self-care, T1D management, and adherence among adolescents.⁹ Specifically, Commissariat et al. (2018) studied the relationship between stressful life events, T1D management, and psychological measures in adolescents with T1D and found that there was an indirect association between the number of stressful life events and youth T1D adherence, perceived self-efficacy, and perceived quality of life, and a direct association with family conflict. However, the study was cross-sectional, localized to adolescents, and excluded youth with recent-onset T1D. There remains a critical need for research using a prospective study design to evaluate the effects of stressful life events on T1D over time among younger children, and among those who are recently diagnosed, because that is such a critical time for intervention.

The objectives of this study were to examine the impact of stressful life events on T1D characteristics of school-age children with recent-onset T1D, and to identify family psychological stressors correlated with greater numbers of stressful life events using a mixed methods design. Our first hypothesis was that parents of children with recent-onset T1D would commonly report the occurrence of one or more stressful life events (defined as >50% of the sample). We also hypothesized that the number of stressful life events that parents reported would correlate cross-sectionally with higher child hemoglobin A1c (HbA1c) and lower child self-monitoring blood glucose. Lastly, because psychological stressors, conflict, and poor coping mechanisms can affect daily living, we hypothesized that higher parental depression scores, higher diabetes-related family conflict scores, and higher active avoidance coping measures at the baseline visit would correlate with greater numbers of stressful life events in the 12 months of follow up.

METHODS

The current study was part of a larger 30-month longitudinal prospective study (TACKLE T1D) examining psychological, social, and T1D-specific outcomes in families of 5 to 9-year-olds with recent-onset T1D. For the current study, we used data collected within the first year of study participation. We collected data at baseline and at subsequent 3-month visits up to 12 months.

Participants:

We recruited children between the ages of 5-9 years with a duration of T1D of ≤12 months and one parent or caregiver to participate in this study. We excluded children with severe psychiatric disorders, non-English speaking families, and children with evidence of type 2 diabetes. Two study sites with multi-clinic networks in the mid-western United States participated. Participants and parents attended their T1D clinic appointments as part of their routine T1D care and research assistants collected parent survey data and child hemoglobin A1c (HbA1c) via fingerstick blood samples during these appointments.

Measures:

Demographic and medical information.

At the baseline appointment, researchers collected the following demographic and clinical information about the youth-parent dyad through self-report and medical chart review: age, child sex, child race/ethnicity, duration of T1D, caregiver relation to child, caregiver age, caregiver marital status, and family income.

Major life events.

Every three months following the baseline appointment, we asked caregivers to endorse the following stressful life events in a yes/no format: child attending different school, caregiver job change, family income change, marital status change, and family member significant health change. If the caregiver reported a job change, income change, marital status change, or family health change, we asked the parent to specify the change in a blank space provided by the item. Some common examples of family health changes include siblings diagnosed with T1D, maternal pregnancies, and hospitalizations or deaths of grandparents.

Treatment Engagement.

At every time point, we measured treatment engagement with T1D management by self-monitoring blood glucose (SMBG). To calculate SMBG frequency at each time point, we used 14 days of device downloads to calculate the average total number of SMBG checks completed per day. Continuous glucose monitoring device (CGM) use was also recorded for each participant. At least twice daily finger-stick testing was still indicated in this time period of the study for all FDA approved devices.

Child HbA1c.

We obtained HbA1c levels at every time point. We collected child blood samples from a finger stick via capillary tube and immersed in stabilizing solution for transport. Then we used a central laboratory to process all of the study samples using automated high-performance liquid chromatography (reference range 4.0-6.0% [3.8-7.0 mmol/l]; Tosoh Corporation, San Francisco, CA). Previous research shows this method aligns with Diabetes Control and Complications Trial standards.^10,11 In the case of missing or inadequate blood samples, we used children’s point-of-care HbA1c value. Previous research demonstrated an average association >0.90 (p< 0.001) between point-of-care and lab-based HbA1c values in the larger study.⁴

Diabetes-specific family conflict.

We used the Diabetes Family Conflict Scale (DFCS) to measure parent-child conflict related to direct and indirect T1D management tasks.¹² We used the revised 19 question scale with updated language. Given the age of children (5-9 years old), researchers asked for parent’s perspective of conflict rather than the child’s perspective. Parents rated each item on a 3-point Likert scale (1=never argue, 2=sometimes argue, and 3=always argue), generating a total score that ranged from 19 to 57 (19=no conflict to 57=high conflict).

Parental Depression.

We used the Center for Epidemiologic Studies – Depression Scale (CES-D)¹³ to assess for parent-reported depressive symptoms. Consistent with published directions, parent’s endorsed items using a 0 to 3 frequency score (0=rarely or none of the time to 3=most or all the time/5-7 days); higher scores correlated with higher parent depressive symptoms.

Coping Method.

We used the Brief COPE inventory to measure parents’ perceptions of their coping style.¹⁴ The Brief COPE asked parents to report on their coping style since finding out that their child had T1D. Parents endorsed each coping strategy based on the frequency with which they used that strategy versus how well they perceived the strategy to work for them. We scored the Brief COPE into four different styles of coping: active avoidance, problem-focused, positive, or religious/denial coping, based on published scoring criteria.¹⁴ Active avoidance is a method of coping in which an individual uses some other action or behavior to evade stressful stimuli (i.e., watching television to think of the situation less). In contrast, problem-focused coping includes strategies to remove or tackle the cause of the stressor (i.e. finding a job after being laid off). Religious coping and denial coping are both emotional coping mechanisms that do not directly address the problem (i.e. praying or believing the situation is not real). Positive coping includes behaviors or thoughts that focus on identifying the positive aspects of encountering a problem (i.e., recognizing that losing one’s job is an opportunity to spend time with family).¹⁴

Statistical Analysis.

We used descriptive statistics to examine all outcome variables and participant characteristics. We calculated family socioeconomic status (SES) using the Hollingshead Four-Factor Index, which computes a class score based on parents’ marital status, highest education level attained, job title, and employment status.¹⁵ Class scores were on a 5-point scale and ranged from low (1) to high (5). To test our first and second hypotheses, we used frequency counts and a priori determined that we would identify stressful life events as “common” if reported by ≥50% or parents. To test our third hypothesis, we used a series of one-tailed Pearson’s correlations to examine the associations between parent-reported stressful life events and children’s HbA1c levels, children’s SMBG frequency, and family demographic variables. Similarly, to test our fourth hypothesis, we used one-tailed Pearson’s correlations to examine the associations between parent-reported stressful life events and parent’s baseline depression score, diabetes-specific family conflict score and coping mechanism. Because we had specific hypotheses for each measure, we used one-tailed Pearson’s instead of two-tailed tests to detect associations in the direction we hypothesized. With these one-sided correlations, we only reported statistical significance for the associations that supported our a priori hypotheses. Due to exploratory nature of the study, we did not apply an alpha correction to reduce the risk of Type II error.

This study was approved by the Colorado Multiple Institutional Review Board and the Pediatric Institutional Review Board at The Children’s Mercy Hospital & Clinics.

RESULTS

Demographics.

Of 129 parent-child dyads consented, 128 (99.2%) completed the baseline visit, 119 (92.2%) completed the 3-month visit, 114 (88.3%) completed the 6-month visit, 111 (86.0%) completed the 9-month visit, and 104 (80.6%) completed the 12-month visit. There were a total of 128 dyads included in this study. At baseline, children had a mean age of 7.5±1.3 years and a mean diabetes duration of 4.6±3.3 months. Most parents identified their child as White (80.6%) and non-Hispanic (89.9%) and 55.0% identified their child as female. Caregivers were mostly mothers (83.7%) with a mean age of 36.7±6.4 years and 75.7% were married. Most families had an SES in the upper-middle class range (65.5% of Hollingshead Scores ≥ 4).¹⁵

Life Events.

Among 128 families, 53.9% (n=69) reported one or more stressful life events during the 12 months of follow-up, supporting our first hypothesis. The number of stressful life events ranged from 0 to 9 per family, with an average of 1.45 (IQR=0-2) life events per family. Specifically, 25.8% reported one or more income changes (n=33) during this period, 23.4% reported one or more additional health changes in the family (n=30), 22.7% reported job changes (n=29), 21.9% reported changes in their child’s school (n=28), and 3.9% had changes in marital status (n=5). Regarding the income changes, 16.4% of all the families experienced income decrease (n=21), 14.8% of all families experienced income change within the same bracket (n=19), and 3.9% of all families experienced an income increase (n=5). Participants specified that health changes either involved other family members’ health (n=23) or the child’s health (n=7). Job changes included the caregiver obtaining a different career/position (n=20), promotion (n=11), a parent quitting his/her job (n=7), and a parent losing his/her job (n=2). Changes in marital status consisted of divorce (n=2), separation (n=2), and death of a spouse (n=1). Some families experienced the same stressful life event multiple times throughout the study period. Figure 1 depicts the frequency of each reported stressful life event at each time point. Table 1 reports the number and percent of participants with each type of stressful life event recorded at each visit and their correlation to HbA1c obtained at that visit. Table 2 records average child age at baseline, duration of diabetes at baseline, and psychological measure scores for 0, 1, 2+ life events subgroups.

Figure 1.

Depicts the number of reports of a stressful event type, split into each visit time point. Some families reported the same type of event multiple times throughout the study (ie, income decrease at both the 3 month visit and the 12 month visit), and this graph includes all reports of events, even if it is from the same family

Table 1.

Cross-sectional HbA1c, incidence of life events, and correlation coefficient of life events to HbA1c and SMBG among youth 5-9 years old with new onset T1D.

	Baseline	3 Months	6 Months	9 Months	12 Months
Mean HbA1c, % (SD)	7.60 (±1.4)	7.73 (±1.3)	8.09 (±1.2)	8.25 (±1.1)	8.30 (±1.2)
mmol/mol (SD)	9.5 (±2.3)	9.7 (±2.1)	10.3 (±1.9)	10.6 (±1.7)	10.6 (±1.9)
Health Change, n (%)		14 (11.8%)	11 (9.6%)	8 (7.2%)	10 (9.6%)
Correlation to HbA1c		−0.065	−0.034	+0.008	+0.035
Correlation to SMBG		+0.138	−0.037	+0.050	−0.075
Marital Status Change, n (%)	75.7% married	3 (2.5%)	1 (0.9%)	0	1 (1.0%)
Correlation to HbA1c Correlation to SMBG	−0.02	−0.106	−0.065	--	+0.016
Correlation to SMBG		−0.055	--	--	+0.020
Income Decrease, n (%)		6 (5.0%)	12 (10.5%)	3 (2.7%)	5 (4.8%)
Correlation to HbA1c		+0.040	−0.072	+0.333**	+0.199**
Correlation to SMBG		−0.009	−0.015	−0.097	−0.243*
School Change		7 (5.9%)	12 (10.5%)	7 (6.3%)	12 (11.5%)
Correlation to HbA1c		+0.083	+0.016	+0.024	+0.200*
Correlation to SMBG		−0.010	−0.023	+0.074	−0.219*
Job Change		9 (7.6%)	13 (11.4%)	5 (4.5%)	14 (13.5%)
Correlation to HbA1c		−0.073	+0.009	+0.240**	+0.082
Correlation to SMBG		−0.155	0.038	−0.062	−0.117
n	128	119	114	111	104

^*p<0.05

^**p<0.01; Positive correlation coefficient indicates higher HbA1c value, or more frequent SMBG.

Table 2.

Baseline descriptive measures of participants grouped by number of life events.

Baseline Measure	0 Life Events n=59	1 Life Event n=25	2+ Events n=44	All Participants
Age (years)	7.00±1.3	7.79±1.2	7.39±1.3	7.5±1.3
T1D duration (Months)	5.36±2.6	5.25±3.6	4.40±3.2	4.65±3.2
DFCS Total Score	22.93±8.3	21.41±2.5	23.80±6.6	21.82±9.5
CES-D Total Score	9.40±6.6	8.45±8.6	12.16±9.2	10.8±8.6

Life Events and Glycemic Control.

Consistent with our third hypothesis, we observed associations between family income decrease and higher child HbA1c at the 9- and 12-month visit, as well as less frequent SMBG at the 12-month visit (Table 1), without any difference in CGM use (X² (2, n=104) = 1.4, p=0.23) between those who had a family income decrease and those who did not. There was also an association between parental job change at the 9-month visit and higher child HbA1c, while we observed associations between school change and higher child HbA1c and school change and less frequent SMBG at the 12-month visit (Table 1). There was no difference in CGM use between those who had school change and those who did not (X² (2, n=103) = 0.9, p=0.33). We did not observe any associations between marital status and health changes with child HbA1c at any time during the study.

Life Events and Psychological Measures.

In support of our fourth hypothesis, parent’s baseline active avoidance coping, CES-D score, and DFCS score correlated with presence of at least one stressful life event (Table 3). Baseline active avoidance coping was positively associated with additional family health changes, an income bracket decrease, child school changes, and parent job changes. We also observed a positive association between parents’ baseline CES-D score and an income bracket decrease within the year of follow-up. There was a positive association between parents’ DFCS scores at baseline and their report of marital status changes, income decreases, and child school changes during the year of follow-up. We did not see an association between income bracket increase and parents’ report of active avoidance coping, depression, or DFCS at baseline (Table 3).

Table 3.

Pearson Correlations between stressful life events and baseline caregiver psychological measures.

Baseline Measure	Active Avoidance Coping	CESD Score	DFCS Score
Any Stressful Change	0.239**	0.197*	0.225**
Health Change	0.248**	0.153	0.126
Marital Status Change	0.035	0.122	0.193*
Income Decrease	0.159*	0.165*	0.287**
Income Increase	−0.055	−0.056	−0.065
School Change	0.175*	0.132	0.199*
Job Change	0.185*	0.018	0.090

^*p<0.05

^**p<0.01

DISCUSSION

This is the first study to examine life events in school-aged children and examine the impact on T1D management via a prospective, longitudinal design. As the 5 to 9-year-old age group has the second highest incidence rate of T1D,² it is valuable to have research specifically examining factors that might complicate T1D management for this age group in order to identify approaches to improve diabetes self-management. Over half of the participants in our study reported experiencing at least one stressful life event, which shows that stressful life events commonly occur in the immediate years after T1D diagnosis for families with school-age children. This study confirms and extends previously published findings that stressful life events are common within the T1D population, that family health changes are among the most common, and that stressful life events are correlated with worsening diabetes management.⁹ Moreover, the stressful life events examined in this study potentially have a strong financial impact on the family, especially income change, which was reported by over one quarter of families.

Our study found an association between a decrease in income bracket and higher child HbA1c, consistent with data seen in TeenCOPE, a study that looked at the contribution of income to glycemic management and psychosocial outcomes in adolescents with established T1D.¹⁶ One theory for this association could be that the rising insulin cost and overall cost of T1D management limits those with low incomes from achieving T1D management goals.¹⁷ Over the past decade, prices of insulin have nearly tripled.¹⁷ Therefore, an income bracket decrease could result in more difficulty or higher perceived difficulty in obtaining medications needed for T1D due to lower expendable income or poorer insurance coverage. In our study, we also saw an association between lower income and less frequent SMBG, which supplements the research of Xie et al. that concluded there was a correlation between lower out-of-pocket cost of blood glucose strips and greater frequency SMBG.¹⁸

Other stressful events, such as school change, may emphasize the importance in continuity of care. The association of school change with both higher HbA1c and less frequent SMBG may highlight the significance of T1D care in school and how the stress of changing schools and school nurses may impact T1D management. Current studies support this theory and illustrate that school care in youth with T1D can be suboptimal, especially in younger age groups.¹⁹

Furthermore, the higher HbA1c results among individuals whose families experienced stressful life events may be heightened by differences in the timing of the end of the child’s honeymoon phase. In the current study, we only see correlations between stressful life events and HbA1c at the 9-month and 12-month follow-up time points. As the average T1D duration at baseline is 4.6 months in the present cohort, these time points correspond to 12 and 16 months after diagnosis, likely around the end of the child’s honeymoon phase. The honeymoon phase is a time when a person with newly diagnosed T1D (often starting 1-3 months after diagnosis) requires less insulin than originally used, or than that of others of the same weight and age, while maintaining a low HbA1c.²⁰ On average, the honeymoon phase lasts 7-9 months.^20-22 It is possible that the honeymoon period may provide a protective factor against the effect of stressful life events on blood glucose levels, as our findings suggest that higher HbA1c levels relate to stressful life events only in the months after the average honeymoon time is over. Future studies should capture honeymoon status estimates in patients (e.g., insulin dose per kilogram) to provide a better understanding of whether honeymoon status may serve as a protective factor.

There is evidence that T1D management in children and adolescents also relates to parental psychological measures.^3,23 Thus, it is important to consider some parental and family psychological characteristics that may relate to stressful life events. Similar to teens with T1D,⁹ we found a correlation between higher family conflict and stressful life events in the younger cohort (5-9 years old) examined by the present study. We also observed associations between stressful life events and baseline active avoidance coping, depressive symptoms, and family conflict and family income bracket decrease in the subsequent study period, suggesting that financial and psychosocial difficulties often co-occur among families of children with T1D. One could postulate that those with psychosocial issues are at higher risk of having difficulties balancing work and a new diabetes diagnosis, resulting in changes in employment. However, qualitative research is needed to further investigate the reasons behind the income bracket changes.

This study complements the current literature on the familial implications of T1D diagnosis in young children during the recent post-diagnosis period and enriches the research on stressful life events with a novel age group and study design. The prospective study design utilized the same survey format and same parent for each study visit, providing valid comparisons between visits and limited discrepancy in the data collection methods. We used the same laboratory technique, solution, and facility for HbA1c analysis thereby enabling us to standardize HbA1c and achieve a measure of child average glycemic control that was reliable and supports greater internal validity.

This study should be interpreted in the context of its limitations. Life events were parent-reported and may have been underreported. For both income change and job change, we did not collect parent perceptions of why the event occurred, thus it was not clear if parents perceive the changes as positive or negative. For example, one participant may consider a job change as a positive opportunity to leave an unproductive job and pursue other interests, while another participant may interpret a job change negatively. Supplementary evaluations of the effect of stressful life events in future studies should include child perception of the event as well, since research shows that even child perceptions of stressful events are related to health.⁸ Another limitation of our study was that our sample was predominately non-Hispanic, white youth and their mothers. Although this racial/ethnic breakdown is representative of the clinic populations where we completed our study, we need future studies with more racially and ethnically diverse samples to determine if our results can generalize to the broader T1D population. CGM use may influence the SMBG of the participants. Because we did not record the exact dates of the CGM starts, type of CGM, and the dates of the life events, this interaction could have persisted even with the similar use of CGM in the groups. As previously mentioned in the data analysis, due to the exploratory nature of the study we did not apply an alpha correction to reduce the risk of Type II error and this method decision is also a study limitation.

These study results indicate that conversation with families regarding all family members’ psychological health history, as well as how the family unit interacts and copes with stressful life events is important at diagnosis and throughout the early post-diagnosis period. In those with income and career changes, it may be beneficial to monitor financial need and access to T1D care and supplies. Though the current study measured changes in income over time, it did not record the numerical income or socioeconomic status at times other than at baseline. Thus for more accurate relationships, it is important for future studies to measure the change in socioeconomic status. Further, health literacy has evidence of having an important role in diabetes outcomes over time in young children with type 1 diabetes families.²⁴ Although it was not examined in the present study, talking about stress and health is an opportunity for early intervention in this group. Knowledge of the impact of stressful life events could additionally help providers and health care professionals provide early interventions with families to improve diabetes management.

Overall, this is the first study to examine life events in school-aged children and examine the impact on T1D management via a prospective, longitudinal design. This study indicates that stressful life events are related to diabetes management. Screening and identifying these additional stressful life events in the immediate years after T1D diagnosis in families with school-age children can lead to earlier discussion and intervention, which could help improve diabetes care. Additional research, including qualitative studies, can elicit more specific details of the relationship between stressful life events and diabetes management to further guide management of this patient population. This study suggests opportunities to conduct further research in diabetes management after stressful life events occur and provides insight into factors that may affect the parent’s ability to care for his or her child’s diabetes in the 5 to 9-year-old age range, especially in the post-diagnosis period. In order to control for time of diagnosis, future prospective studies could capture patients immediately at diagnosis or address the stressful events retrospectively. Finally, more focused studies among youth with T1D utilizing specific measures of stressful events, perception of those events and their interaction with T1D management may be essential to achieve a greater understanding of the relationship between stressful events and T1D management. Other prospective studies involving larger groups of participants may consider conducting repeated ANOVAs to determine significant changes over time and trajectories of T1D outcomes. Diabetes registry studies may be able to accomplish this by incorporating stressful life event questionnaires.