The incidence of pediatric type 1 diabetes mellitus (T1DM) is increasing wordwide 3% per year [1]. In the United States, the cost of diagnosed diabetes was estimated to be in the US $400 billion range in 2022 [2], much of this cost being related to complications of suboptimal adherence to treatment. Consistent glucose control is vital to avoiding long-term complications of T1DM, such as retinopathy, nephropathy, and cardiovascular disease [3,4]. Current guidelines suggest keeping hemoglobin A_1c (HbA_1c), the traditional gold standard metric for T1DM control, to below 7% in both children [5] and adults [6]. Another important clinical outcome in diabetes care is the prevention of diabetic ketoacidosis (DKA), a life-threatening complication of diabetes that continues to be the most common cause of hospitalization and death in children with T1DM [7,8].

The advent of insulin pumps and continuous glucose monitors (CGMs) has significantly advanced T1DM management in recent years. Children who use CGMs [9] and pumps [10] are more likely to attain HbA_1c levels below 7%, irrespective of socioeconomic status, insulin regimen, or duration of diabetes [9]. Despite these advancements, adherence to T1DM regimens (eg, consistent insulin administration, regular glucose monitoring, and appropriate compliance with dietary recommendations) continues to be suboptimal. The transition from childhood to adolescence has been established as a strong risk factor for poorer adherence and blood glucose control [11,12]. This developmental transition is characterized by shifting responsibilities from parents to patient, feelings of increased social pressure from peers, and fatigue from chronic illness management [13], factors that have been shown to further complicate this period and to be linked to comorbid depression. Depression is common in adolescence and has been linked to reduced adherence [14] and higher HbA_1c levels [15] via its negative impact on motivation, cognitive functioning, and self-efficacy.

Health insurance type is known to influence access to T1DM care and supplies [16]. Despite growing evidence that generous insurance coverage of diabetes technology improves outcomes [17], there continue to be significant barriers for youth with public insurance related to strict prior authorization requirements, high copays, and lack of access to specialized health care for both pumps and CGMs [16].

The aim of this study was to investigate whether diabetes technology use moderated the negative effects of teenage depression across 3 large health care systems in California. We also examined whether technology use mitigated the previously demonstrated disparities in glycemic control associated with public insurance. We hypothesized that having a history of depression and being publicly insured would predict higher HbA_1c levels and more frequent episodes of DKA, and that these effects would be reduced by the use of technology.

We found that use of diabetes technology was associated with improved glycemic control, attenuating the negative impact of depression and mitigating outcome disparities related to public insurance. Diabetes technology use was predictive of better glycemic control through both lower HbA_1c levels and fewer DKA events. While having a reported depression diagnosis was associated with higher HbA_1c levels and more DKA events, technology use attenuated this risk in an additive manner. These findings support the strategy of offering diabetes technology to all youth, regardless of comorbid mental health issues. However, providing an insulin pump to youth with depression has historically been considered risky given the potential for self-harm via exogenous insulin administration. Given the benefits of technology use in youth with depression, clinicians who continue to be hesitant to prescribe a pump to due to fears of “suicide by insulin” should consider using tailored screening of suicide risk rather than a broader depression or mood measure. The Columbia-Suicide Severity Rating Scale [18] appears to identify twice the number of T1DM patients at risk for self-harm than its more general depression counterpart, the Patient Health Questionnaire 9 [19], which was used in this study [20]. Our results also underline the importance of screening for depression in pediatric settings in order to minimize its impact on the health of individuals with T1DM. In fact, research shows that both pharmacological and psychotherapeutic treatments for depression are associated with improvements in glycemic control [21].

Although not examined in this study, it is also worthwhile to consider whether comorbid depression impacts regularity and consistency of technology use. In general, youth who perceive more barriers to technology use are less likely to use it [22]. In youth with T1DM, commonly reported barriers to consistent technology use include cost and insurance issues and wear-related concerns [22]. While depression does not appear to predict discontinuation of CGM use [23], more research is needed to confirm this finding. Data on technology use collected for this study reflect a dichotomous variable and therefore cannot be used to inform the larger literature on health outcomes as it relates to technology consistency and discontinuation. What is known in regards to CGM is that users consistently report more engagement in diabetes self-care and a better quality of life, likely because they feel more efficacious in day-to-day disease management [24]. Given that helplessness is central to the development and maintenance of depression [25], encouraging CGM use may present one way to help youth feel more in control of their health.

With regard to insurance type, our study demonstrates that outcome disparities between publicly and commercially insured youth are mitigated by both CGM and pump use. Prior studies of publicly insured pediatric patients with T1DM indicate that interruptions in the use of CGM were largely due to insurance gaps and were predictive of poorer outcomes [26]. Countries that have facilitated technology access have seen improved clinical outcomes in youth with T1DM. For instance, after Australia changed its national health policy to universally subsidize CGM use in youth with T1DM, both HbA_1c levels and DKA hospitalizations significantly decreased in youth [27]. There have also been documented differences in countries with and without universal health care. A large-scale examination of T1DM outcomes in American versus German youth found that disparities between socioeconomic status and HbA_1c were evident in both groups but that technology access was a significant covariate only in the American sample [28].

At the time of data collection in California (2016-2021), public insurance companies mandated that patients document 4 fingerstick blood glucose tests every day for at least 1 month to qualify for a CGM [16]. Even 1 instance of a missed test could disqualify families. Many other state Medicaid programs have had the same requirement [16]. Fortunately, as of 2021, pharmacy benefits for publicly insured patients in California have expanded, making it easier to obtain diabetes technology for publicly insured patients [29]. Our data support expanding pharmacy benefits to other states that have not yet made them publicly available.

Besides improved clinical outcomes, prior research has shown that facilitating universal access to CGM and pump use is associated with significant cost savings. While the cost of a CGM is estimated to be US $15.20/day (extrapolated to approximately US $5000/year) [30], 1 admission for DKA in the United States costs upward of $30,000 [31]. In fact, the cost-effectiveness ratio of CGMs in pediatric T1DM is well established, demonstrating not only short-term improvements but significant reductions in costs related to long-term complications [32], emergency room visits, and hospitalizations [33].

There are several limitations in our current study. First and foremost, as this was a retrospective, nonrandomized cohort, causality cannot be inferred from our results. Patients who use diabetes technology were self-selected to some degree, and thus may have had differences in motivation, self-efficacy, and other important variables that could contribute to improved outcomes. Patients and families more motivated to attain optimal diabetes control may be more likely to pursue technology, suggesting that social support is also likely at play. Therefore, future studies evaluating social determinants of health may be helpful to clarify the impact of social support on the benefits of diabetes technology. As laid out in the introduction, publicly insured youth with CGM may have had to document daily adherence to blood glucose monitoring in order to qualify for the device. As such, we cannot exclude the possibility that HbA_1c levels for this group are attributable to better overall adherence rather than solely contingent on subsequent CGM use.

Depression was collected as a binary variable, and so degree of depression and status of treatment were not known. Patients with milder or better-controlled depression might be more likely to agree to diabetes technology than those with more severe symptomatology (due to increased feelings of hopelessnes and helplessness). Furthermore, the depression diagnosis data were not collected uniformly, which could have impacted the validity of results. It is also possible that some patients in the sample may have had undiagnosed depression, while others may have said they had depression without having a formal diagnosis. Future research should seek to replicate these results using a uniform method of depression screening. Thankfully, the utility of universal depression screening in pediatric patients with T1DM is gaining momentum [34]. It is also worth considering preliminary data on the use of medical record mining algorithms that can effectively predict patients at higher risk of depression in pediatric settings, where adolescents may be likely to underreport difficulties due to stigma, socially desirable responding, and the desire to be autonomous in their health management [35]. The integration of mental health care providers in clinical settings can also be a value-added service that removes the burden of mental health screening from endocrinologists [36].

To conclude, this is the first study to examine the interplay between socioeconomic, psychological, and medical factors in elucidating differential outcomes for adolescents with T1DM. It also represents a large sample of youth across 3 major health systems from one of the most populous states, allowing for the examination of these complex variables at the state level. Future studies should aim to replicate the above model with a national sample to determine how insurance status interacts with sociocultural variables such as parent education, income, race/ethnicity, and health literacy, especially given established disparities in technology acceptance based on cultural factors [37]. Our study also supports universal depression screening of adolescents in pediatric settings to ensure that mental health issues are treated promptly and are less likely to impact medical trajectories. Finally, given the concerning trend that socioeconomic status–based disparities appear to be increasing in American youth with T1DM [28], we cannot overstress the importance of facilitating universal coverage of diabetes technology. Such efforts are vital in our fight to provide equitable care that will facilitate improved clinical outcomes across demographic groups.