Comparative analysis of treatment modalities and health-related quality of life in pediatric type 1 diabetes in a predominantly Hispanic population

Study design and setting

This cross-sectional study was conducted in accordance with the ethical guidelines for medical research involving human participants, following approval from the Institutional Review Board (IRB). The study was conducted at a diabetes center within a freestanding children’s hospital in South Florida. This diabetes center is one of only 33 Association of Diabetes Care and Education Specialists (ADCES)-accredited centers in Florida, providing comprehensive care to over 1,000 pediatric patients with diabetes annually.

Population and sample

A total of 190 participants were recruited between December 2023 and March 2024. Eligible participants were pediatric patients aged 2 years or older who were diagnosed with T1D and were utilizing either AID or MDI for diabetes management. Patients with a diagnosis of type 2 diabetes were excluded from the study.

Procedures and instruments

To assess HRQoL, the study utilized the validated PedsQL 3.2 diabetes module, a 33-item instrument specifically designed to evaluate diabetes-related symptoms and management challenges among pediatric patients and their caregivers. This instrument includes two principal subscales: the diabetes symptoms scale and the diabetes management scale. Scores for each subscale were derived by summing the responses and dividing by the total number of items answered.

Subscales with more than 50% missing data were excluded from the analysis. An overall HRQoL score was computed by aggregating responses across all items and dividing by the number of items completed across all scales. Each item was rated using a 5-point Likert-type scale, which was reverse-scored and subsequently linearly transformed to a 0–100 scale (0 = 100, 1 = 75, 2 = 50, 3 = 25, and 4 = 0), where higher scores indicated better perceived HRQoL. [8]

Participants attending scheduled clinic visits were invited to complete the PedsQL 3.2 diabetes module while in the waiting room before their consultation. For those not attending clinic visits during the data collection period, the module was distributed electronically through REDCap, a secure online data capture platform. The method of completion varied by patient age:

• Ages 2–<7 years: The parent or caregiver completed the parent-proxy version of the module

• Ages 7–18 years: Both the patient’s self-report and the parent-proxy versions were completed independently

• Ages >18 years: Only the patient self-report version was completed.

In addition, a comprehensive chart review was conducted for participants who completed the PedsQL 3.2 diabetes module. Demographic information, including gender, race, ethnicity, and insurance type, was recorded. Clinical data, including the most recent HbA1c, TIR, and duration of diabetes, were also collected. If a patient had incomplete data, they were not included in the comparative analysis.

Data analysis

Descriptive statistics were used to summarize demographic data, reporting means, and standard deviations for continuous variables and counts and quartiles for categorical variables. First, the two variables assessing glycemic control (HbA1c and TIR) were examined for unadjusted differences between the MDI and AID groups using t-tests. Next, the unadjusted bivariate associations were evaluated in relation to both parent and patient diabetes symptoms scores, diabetes management scores, and overall PedsQL scores using t-tests to compare group means for these categorical independent variables. The demographic characteristic variables were then tested independently for an association with the PedsQL scores using t-tests or one-way analysis of variance. Finally, an analysis of covariance method using age as a covariate was used to determine if the treatment modality was significant after adjusting for heteroskedasticity found in the relationship between age and the PedsQL scores. Scatter plots and bar charts were used to visualize the data. Statistical significance was set at a P value threshold of 0.05. All statistical analyses were conducted using the Statistical Package for the Social Sciences version 28.0.

Demographics

A total of 190 patients with demographic characteristics are summarized in Table 1. The cohort consisted of 88 females (46%) and 102 males (54%), with a mean age of 12.5 years (standard deviation [SD] = 4.5). The majority of participants were identified as Hispanic (81%). Racially, the population was predominantly white (87%), with smaller proportions identifying as black or African American (6%) or from other racial backgrounds (7%). In terms of insurance, a higher proportion of patients using MDI were covered by public insurance (57%), compared to those with private insurance (43%). In contrast, the distribution of insurance types in the AID group was more balanced, with 49% of participants on public insurance and 51% on private insurance.

Diabetes control

Both the MDI and AID groups included most patients diagnosed with T1D for over 5 years, with 30% in the MDI group and 49% in the AID group. Notably, the highest proportion of patients diagnosed within the past year was in the MDI group (24%). Significant differences in glycemic control were observed between the two treatment modalities. The mean HbA1c in the AID group (7.9%, SD = 1.2) was significantly lower than that in the MDI group (8.9%, SD = 2.1), with a statistically significant difference of −1.07 (P < 0.001). This indicates that AID users achieved better glycemic control. In addition, TIR was significantly higher in the AID group (mean = 57%, SD = 20) compared to the MDI group (mean = 47%, SD = 26), with a mean difference of 10% (P = 0.004), suggesting that AID users spent more time within their target glucose range.

HRQoL

Analysis of HRQoL scores revealed no significant differences between the treatment groups [Figure 1]. For parent-reported QoL, the AID group had a mean score of 74 (SD = 15) while the MDI group had a mean score of 72 (SD =18). Similarly, child-reported QoL was marginally higher in the AID group (mean = 74, SD = 16) compared to the MDI group (mean = 73, SD = 18). A t-test confirmed that there were no statistically significant differences between the groups for both parent-reported (t = 0.640, P = 0.523) and child-reported PedsQL scores (t = 0.039, P = 0.969).

Close

Figure 1:

Comparison of diabetes management and quality of life (QoL) scores reported by parents and children in youth using multiple daily injections (MDI) versus insulin pumps (automated insulin delivery (AID) systems). Mean scores were slightly higher in the AID group for all four domains: parent-reported management (AID: 77 vs. MDI: 75), child-reported management (80 vs. 78), parent-reported QoL (74 vs. 72), and child-reported QoL (74 vs. 73). Error bars represent 95% confidence intervals. While trends favored the AID group, none of the differences reached statistical significance (P > 0.05 for all comparisons).

Export to PPT

Further analysis using a between-subjects effect test revealed that gender had a statistically significant relationship with child-reported diabetes symptoms, diabetes management, and overall PedsQL scores (P < 0.001, P = 0.002, P < 0.001) [Figure 2]. Still, none of the parent-reported scores were statistically significant. Gender accounted for approximately 9% of the variability in PedsQL scores, with males on average having a 10% higher score in all three domains. Age, ethnicity, race, and insurance were not independently associated with any of the PedsQL domains.

Close

Figure 2:

Comparison of child-reported symptoms, diabetes management, and quality of life (QoL) scores between females and males. Male participants reported higher scores across all domains: symptom burden (male: 74 vs. female: 62), management (84 vs. 75), and overall QoL (79 vs. 69). Error bars represent 95% confidence intervals. These findings suggest that male children may perceive fewer symptoms, better management, and greater overall QoL compared to female peers.

Export to PPT

Age was identified as a key factor in the analysis. Younger patients displayed greater variability in their diabetes management which may be related to challenges such as limited autonomy and greater reliance on caregivers. This was most evident in the child-reported diabetes management scale [Figure 3]. Age was included as a covariate in the analysis to ensure that differences in treatment modalities between the MDI and AID groups were not confounded by age-related factors.

Close

Figure 3:

Scatter plot illustrating diabetes management scores from the PedsQL diabetes module 3.2 reported by different ages (8–21 years) and stratified by gender (female vs. male). Left panel shows that female participants had a wider spread of scores across all ages with a tendency for lower scores compared to males. In the right panel, male participants generally reported higher diabetes management scores with a more consistent pattern of clustering across all ages. This suggests that males may feel more confident or effective in managing their diabetes compared to females especially as they age, while females tend to remain scattered throughout.

Export to PPT

To examine the potential impact of diabetes treatment modality on HRQoL, a between-subjects effect test was conducted, adjusting for the diabetes symptoms scale score. The diabetes symptoms scale score was found to have a highly significant relationship with HRQoL (F = 832, P < 0.001). After controlling for symptom severity, the treatment modality approached statistical significance (F = 4, P = 0.051), suggesting a potential trend where children using the AID system may perceive their QoL differently compared to those using MDI. However, the effect was not strong enough to draw definitive causal conclusions.

HRQoL and glycemic control

To further explore the relationship between glycemic control and HRQoL, two additional between-subjects effect tests were performed, focusing on HbA1c and TIR while adjusting for the diabetes symptoms scale score. The model examining HbA1c was highly predictive of HRQoL, with an R-squared value of 0.8 (adjusted R-squared = 0.8). The diabetes symptoms scale significantly impacted HRQoL (F = 814, P < 0.001), and after adjusting for symptom severity, HbA1c remained significantly associated with HRQoL (F = 5, P = 0.026). This suggests that poorer glycemic control, as indicated by higher HbA1c levels, was associated with lower HRQoL. In contrast, TIR did not significantly impact HRQoL (F = 0.3, P = 0.603) after controlling for symptom severity, indicating that while HbA1c serves as a meaningful indicator of HRQoL, TIR may have less influence when diabetes-related symptoms are considered.

Glycemic control and HRQoL

Our findings are consistent with the SEARCH for diabetes in youth study,^[3] which reported that AID users had improved glycemic outcomes but had a more nuanced impact on HRQoL. Our results demonstrate that AID systems are associated with significantly better glycemic control, as evidenced by a lower mean HbA1c and a higher mean TIR compared to MDI users. These clinical advantages align with previous studies, which have shown that AID systems improve overall glycemic control and reduce the incidence of hypoglycemia, thereby contributing to more stable blood glucose levels.^[1] The improvements noted in glycemic control between groups may be related to the significant difference observed in time since diagnosis between treatment groups, with a greater proportion of participants in the AID group having longer diabetes duration compared with those on MDI (P = 0.005). This suggests that individuals using AID were more likely to have lived with diabetes for several years, whereas those on MDI were more often earlier in their disease course. This imbalance may reflect clinical practice patterns in which advanced technologies are preferentially adopted by patients with longer disease duration, greater familiarity with diabetes self-management, or more prior exposure to intensive insulin therapy. Importantly, differences in disease duration could partially contribute to observed differences in glycemic outcomes and should be considered when interpreting between-group comparisons. Despite these clinical benefits, we observed no significant differences in HRQoL between the two groups, both in terms of parent-reported and child-reported QoL. This finding suggests that factors beyond glycemic control play a critical role in how patients and caregivers perceive their overall well-being.

Interestingly, although AID users achieved superior glycemic outcomes, their perceptions of QoL were similar to those of MDI users. This highlights the complexity of QoL, which encompasses more than just clinical indicators such as HbA1c and TIR. Previous research has underscored the multifaceted nature of HRQoL in chronic conditions, particularly in pediatric populations, where psychological, emotional, and social factors can significantly influence how the disease is experienced.^[7] For example, the burden of constant glucose monitoring, the emotional toll of managing a chronic condition, and the psychosocial challenges of T1D may offset the clinical benefits associated with advanced technologies.

Gender and HRQoL

An important finding in our study, similar to Naughton’s results,^[3] is the significant relationship of gender on child-reported HRQoL, with gender accounting for 9% of the variability in scores. This is consistent with previous studies that have highlighted gender differences in how chronic conditions, including T1D, are experienced. Female adolescents are more likely to report higher levels of diabetes-related distress and emotional difficulties, which could contribute to lower perceived QoL.^[2] These differences may be influenced by a range of factors, including earlier puberty, heightened body image concerns, and societal expectations, which disproportionately affect girls during adolescence.^[3] In addition, societal pressures and gender norms may exacerbate emotional distress, as girls often face higher expectations regarding appearance, behavior, and academic performance. These findings underscore the importance of adopting gender-sensitive approaches in diabetes care that address the unique psychosocial challenges faced by male and female patients and tailor interventions to meet their specific needs.

HRQoL and glycemic control: The role of HbA1c versus TIR

Our analysis also explored the relationship between glycemic control and HRQoL, with a focus on two key metrics: HbA1c and TIR. HbA1c emerged as a significant predictor of HRQoL, whereas TIR did not show a significant relationship once symptom burden was accounted for. This is consistent with the notion that while TIR is an important indicator of glucose stability, it may not fully capture the psychosocial impact of living with T1D. Conversely, HbA1c, which reflects long-term blood glucose control, may be more directly linked to how patients and caregivers perceive their overall health and well-being.

Age and variability in management

Age was also identified as a key factor influencing HRQoL, particularly in younger patients. As expected, younger children demonstrated greater variability in their diabetes management, which may be due to their limited autonomy and greater reliance on caregivers. The challenges of diabetes management in younger children—such as the need for constant supervision, the emotional toll on caregivers, and the potential for missed treatment regimens—can all contribute to increased stress and lower perceived QoL. These age-related factors must be considered when designing interventions and support systems that aim to optimize both clinical outcomes and QoL for pediatric patients with T1D.

Health disparities and the Hispanic population

This study’s focus on a predominantly Hispanic population highlights the unique challenges faced by minority groups in managing T1D. However, the study cannot address socioeconomic variables, which could be considered a limitation of the study. While our study did not directly assess these factors, it is important to acknowledge the potential role of cultural and financial considerations in shaping how families perceive and manage diabetes care. Future research should explore these barriers in more depth to ensure that advancements in diabetes technology are accessible and culturally appropriate for underserved populations.

Limitations

The study has several limitations that should be considered when interpreting the findings. First, the predominantly Hispanic population may restrict the generalizability of the results to other demographic groups, potentially limiting the applicability of the conclusions across diverse populations. In addition, socioeconomic status, aside from insurance type, was not measured as a confounding factor, which could influence the outcomes. The use of the PedsQL, which is completed by parents and patients through self-report, introduces the possibility of self-report bias, potentially affecting the accuracy of the results. Finally, the small sample size for subgroup analysis reduces the statistical power and may limit the ability to detect nuanced differences within subgroups.

Implications

• Future research should explore the factors that influence QoL in pediatric diabetes, including psychosocial factors, family support, and individual preferences

• Gender-specific interventions may be beneficial for enhancing QoL in pediatric diabetes care

• Age-related variability in diabetes management should be considered when assessing treatment modalities, especially in younger patients

• Strategies to manage diabetes symptoms effectively can have a significant impact on QoL.

Data availability statement

Datasets generated and/or analyzed are protected by the Health Insurance Portability and Accountability Act. While some datasets may be available from the corresponding author, due to the inclusion of sensitive participant information, data sharing is restricted to protect privacy and ensure compliance with IRB guidelines.