Expect the Unexpected: Adolescent and Pre-teens’ Experience of Diabetes Technology Self-Management

Abstract

Objective:

Only 17% of adolescents with type 1 diabetes (T1D) are currently meeting their glycemic targets despite advances in diabetes technologies. Self-management behaviors and challenges specific to use of diabetes technologies are insufficiently studied in adolescents. We aimed to describe the experience of diabetes technology self-management, including facilitators and barriers, among preteens/adolescents with low and high A1C.

Research Design and Methods:

Youth (10–18 years of age) with T1D who use insulin pump therapy were recruited from the larger quantitative cohort of a mixed methods study for participation in semi-structured qualitative interviews. Maximum variability sampling was used to recruit youth with A1C <7.5% (n=5) and A1C >9% (n=5). Participants’ personal insulin pump and continuous glucose monitoring data were downloaded and served as a visual reference. Interviews were analyzed using a qualitative descriptive approach.

Results:

Participants were 50% female with a median age of 14.9 years and 80% used CGM. The sample was predominantly white (90.0%). Analysis produced four major themes, Bad Day, Expect the Unexpected, Nighttime Dependence and Unpredictability, It’s Really a Team and interconnecting subthemes. Youth characterized “Bad Days” as those requiring increased diabetes focus and self-management effort. The unpredictability (“Expect the Unexpected”) of glucose outcomes despite attention to self-management behaviors was considerable frustration.

Conclusions:

Diabetes devices such as insulin pumps are complex machines that rely heavily on individual proficiency, surveillance, and self-management behaviors to achieve clinical benefit. Our findings highlight the dynamic nature of self-management and the multitude of factors that feed youths’ self-management behaviors.

Background

Only 17% of adolescents with type 1 diabetes (T1D) are meeting their glycemic targets despite increased use of diabetes technologies.[1] Diabetes technologies, including insulin pump therapy, rely heavily on individual proficiency, surveillance, and self-management behaviors to achieve clinical benefit,[2] that may account for varying results.[3] The multiple self-management behaviors necessary to achieve good glycemic control often conflict with the preteen and adolescent desire for spontaneity, athletic activities, and peer relationships, leading to poorer treatment self-management and wide variation in outcomes in this age group.[4] Adolescents face multiple obstacles to successful diabetes management including social pressures, developmental behaviors, increased insulin resistance associated with puberty, and a shift in diabetes management from parent to child.[4 5] Adolescents with T1D experience increased levels of stress surrounding their diagnosis and self-management behavior and often exhibit maladaptive coping strategies.[6 7] The transition to diabetes self-management from parent to adolescent is consistent with multiple other adolescent transitions, but is often awkward and halting which leads to declines in glycemic control.[8] Higher degrees of family support and involvement correlate with better glycemic control for adolescents,[8] yet the use of insulin pump therapy is associated with increased self-management autonomy and decreased parental supervision for adolescents.[9 10]

Self-management behaviors and challenges among youth who use diabetes technologies has not been sufficiently studied.[11 12] In this manuscript, we report the qualitative results from a mixed methods study examining self-management among pre-teen and adolescent insulin pump users. In this work, we sought to describe the experience of diabetes self-management including facilitators and barriers, among preteen/adolescent insulin pump users.

Methods

The Children’s Hospital IRB approved the research protocol. We conducted a sequential mixed method study conducted in two phases. Phase 1 quantitative results have previously been reported and included 80 youth (10–18 years of age) with T1D > 1 year on pump therapy for >6 months[13] In phase 2, semi-structured interviews were conducted with youth between June 28, 2019, and February 5, 2020. The participant’s personal insulin pump and CGM data were downloaded and served as a visual reference for participants to analyze and describe their self-management behaviors and experiences. Interviews were analyzed using a qualitative descriptive approach.[14]

Sampling and Participants

Participants were recruited from the larger phase 1 sample. Eligibility criteria included diagnosis of T1D for >1 year, and insulin pump use for >6 months. We purposively selected 10 interview participants based on glycemic control (A1𝐶𝐶≤7.5% and A1𝐶𝐶≥9.0%) as well as age, length of diabetes diagnosis, insulin pump brand, and sex. The youth’s A1C at the time of quantitative data collection was used to identify potential participants. Maximum variation sampling on measures of glycemic control allowed for analysis in a wide range of cases and was chosen to facilitate analysis of patterns and variations among youth with low and high A1C.[14] The parents were contacted by phone to invite youth to participate in qualitative interviews. Phase 1 of the study included 32 youth with A1C≤7.5%, 15 youth with A1C≥9.0%, and an additional 33 youth with A1C between 7.5–8.9% who were not eligible for participation in phase 2 qualitative interviews. Youth from the larger quantitative phase were recruited using convenience sampling during their regular endocrinology clinic appointment. A detailed description of participant recruitment for phase 1 were previously reported.[13 15] Sixteen parents were contacted and 10 parents and youth ultimately agreed to participate. Parents were reminded of the interview 1 week and again 1 day prior to the scheduled meeting. Written informed consent/assent was obtained at the time of the interview.

Data collection

Data for this study were obtained from semi-structured interviews conducted by a single interviewer (ERF). The Self- and Family Management Framework,[16] along with the literature review, guided semi-structured interview development. Mock interviews were conducted with two College Diabetes Network (CDN) members to refine the interview guide. The interview began with a “grand tour” question, “Tell me about the things you do each day to manage your diabetes with your insulin pump (and CGM for participants who used CGM)?” (Appendix A). The participant’s insulin pump and CGM data were downloaded using Tidepool™ (Palo Alto, CA) software and a 5-day daily report was printed as a visual reference to facilitate and validate the individual’s reflection about their self-management experience. Youth were encouraged to mark or write on the download to better emphasize events. Participants were asked to identify a day they felt captured their best day as well as a day they felt was poor. The interviewer drew attention to particular events (e.g., hyperglycemia, advanced feature use), prompting the participant to describe the circumstances surrounding the event (i.e., “tell me about this time; what was going on?”). All interviews were audio recorded, professionally transcribed, and reviewed for accuracy. Interviews lasted a mean of 51.9±15.4 minutes (range 27 to 80) with most (80%) extended beyond 45 minutes.

Data analysis

Demographic data were analyzed using STATA 12.0 (College Station, TX). Two coders, primary (ERF) and secondary (MBH), analyzed the transcripts using a qualitative descriptive approach.[14] Interview data, fieldnotes, and memos were imported into NVivo™ for data management and analyzed using qualitative content analysis[14 17 18] and constant comparison techniques.[19] Portions of text were coded with terms that were low inference (“data close”)[14] and descriptive of the youth’s experience. Device downloads were analyzed with the transcribed text to better understand the circumstances, meanings, and strategies youth discussed.

We grouped related codes into thematic categories and subthemes. Properties of overlapping dimensions and subthemes were identified within each theme. Multiple iterations of diagrams were constructed by ERF and MBH in an effort to assess strength and direction of themes across cases and to understand the hierarchy of concepts within each thematic category and subtheme.[20] Case exemplars and extreme cases were constructed to gain an understanding of the range of values within themes and subthemes. Interviews were transcribed and coded concurrently with data collection. By analysis of interview number 8, similarities were identified within the data. Two more interviews were conducted beyond this point to ensure redundancy and data saturation. Saturation was achieved and data were determined to be adequate with the emergence of no new patterns, themes, subthemes or codes during interviews 8, 9 and 10.[14] [21]

Finally, selected quantitative outcomes from the larger descriptive study were contrasted and compared to further dimensionalize main themes and patterns.[22 23] Rigor was achieved by constructing date-stamped, analytic memos documenting coding decisions, preliminary ideas, categories and insights; through weekly group analytic meetings and dual coding of all interviews by ERF and MBH. A multidisciplinary team of investigators and people with diabetes (pediatric endocrinologist, endocrinology nurse practitioner, diabetes educator, CDN members) reviewed themes and subthemes to ensure analysis “rings true” with their clinical expertise and lived experience.

Results

Demographic, socioeconomic, and clinical characteristics of the participants are presented in Table 1 stratified by A1C. Half of the participants were female, median age 14.9 years (range 10–17 years), with 6.6 years (range 1–15 years) since diabetes diagnosis, 4.6 years of pump use (range 0.5–11 years) and 8 used CGM. Nine of the 10 participants identified as white.

Table 1.

Interview Participant Demographic, Socioeconomic, Household Structure, Academic Performance and Hospitalization Characteristics

Characteristic	Total (N=10)	A1C≤7.5% (N= 5)	A1C≥9.0% (N= 5)
Female vs. male, N	5	2	3
Age (years), Median, range	15.8 (10–17)	16 (14–17)	15.5 (10–17)
Duration of diabetes (years), Median, range	7.5 (1–15)	4 (1–8)	8 (3–15)
Duration pump use (years), Median, range	3.5 (0.5–11)	3 (0.5–4)	7 (2–11)
A1C (%), Median, range	8.2 (6.0–9.9)	7.0 (6.0–7.3)	9.1 (9.0–9.9)
Race/ethnicity
Non-white vs. White, N	1	0	1
* CGM
CGM (yes vs. no), N	8	4	4
Parental Education
High school or less, N	1	0	1
Some college, N	3	1	2
Bachelor’s degree, N	4	2	2
Graduate degree, N	2	2	0
Insurance
Private, N	8	4	4
Government, N	1	0	1
Private & Government, N	1	1	0
Household Characteristics
Income
$28,000 – $50,000, N	2	1	1
$50,000 – $100,000, N	4	1	3
>$100,000, N	4	3	1
Home environment
Single parent, N	1	1	0
Married parents, N	8	4	4
Parent & step-parent, N	1	0	1
Hospitalization within 1 year
*DM hospitalization/*ED visit, N	4	1	3
Hyperglycemia w/o *DKA, N	2	1	1
DKA, N	2	0	2
Hypoglycemia, N	0	0	0
Pump implicated, N (site issue)	3	1	2

^*CGM, continuous glucose monitoring; DM, diabetes mellitus; ED, emergency department; DKA, diabetic ketoacidosis

We identified four major themes, Bad Day, Expect the Unexpected, Nighttime Dependence and Unpredictability, It’s Really a Team, and interconnecting subthemes in the interview data. The themes, subthemes and their properties are discussed below.

Bad Day

Each participant was asked to highlight a day on the insulin pump download that they would characterize as a “good day”, and then, in contrast, a “bad day.” The youth’s characterization of “good” and “bad” depended on the degree of effort and time they invested in managing diabetes. On “bad days,” youth dedicated increased cognitive and physical effort to regaining control of their diabetes. This lack of control most typically manifested as hyperglycemia or high glucose variability. The youth’s characterization of a bad day, the interwoven subthemes, and dimensions of each subtheme are depicted in Table 2.

Table 2.

Themes, subthemes and dimensional codes

Bad Day
Definition: Bad day vs. good day characterized by the amount of effort and time invested in diabetes management. On bad days youth increased time and cognitive and physical effort for self-management
Subthemes:	Relationship to bad day:	Dimensional codes:
Feeling gross	Symptoms, “feeling gross,” compelled increased self-management time and effort. Hyperglycemia or fluctuating glucoses precipitated feeling gross.	Nausea
		Fatigue
		Abdominal pain
Focus on diabetes	Focusing physical and cognitive effort on regaining glycemic control often requiring youth to withdraw from their activities.	Take myself out of current or scheduled activity to manage diabetes
		Time and effort invested in regaining control
		Waiting for glucose response after treatment
Setting matters	Setting influenced the degree of attention youth dedicated to regaining control and their ability to focus on diabetes management. Setting influenced symptom distress.	Physical location: school, home, athletic event/practice
		Environment inhibited/facilitated management. Symptoms (i.e., confusion, irritability, nausea, fatigue) more concerning in larger groups or outside the home
Expect the Unexpected
Definition: Glucose outcomes described as unpredictable and characterized by unexplainable low or high blood glucose values. This unpredictability in glucose outcomes was incredibly frustrating.
Subthemes	Relationship to Expect the Unexpected	Dimensional codes
Personal history	Previous experience used to predict future outcomes. Youth recounted previous positive and negative outcomes and attempted to mimic successful or avoid unsuccessful self-management behaviors in similar circumstances.	Inconsistent with previous outcomes
		Inability to predict outcomes
		Unable to control extraneous factors
It’s just luck sometimes	Acknowledged luck in diabetes outcomes. Youth ascribed variability and unpredictability to luck.	Day to day inconsistencies in glucose
		Disconnect between actions/self-management and glucose outcomes
		Personification of diabetes and insulin as if they had objectives and a will of their own
Rollercoaster of treatment highs & lows	Inability to control outcomes in response treatments for high or low glucoses resulting in the reciprocal effect and a cycle of treatment.	Rebound glucose after treatment of low
		Fear of reciprocal effect after the treatment of highs or lows
Knowing why	Outcomes were especially concerning when youth were unable to ascribe reason (explanation) for a high or low glucose.	Unexplained and unexpected glucoses after mapping out or planning for event
		Limited self-management; reason for high or low helped dictate next self-management action (i.e., missed bolus, increased physical activity)
Unplanned technology management	Technology interactions (i.e., sensor change, pump site change) especially disconcerting when unplanned or unexplained. Otherwise they were not described as bothersome.	Technology inaccuracies (i.e., false CGM high or low)
		Device failure (i.e., sensor, site failure)
		Setting matters; unplanned events especially bothersome when occurring outside home
Nighttime Dependence and Unpredictability
Definition: Youth relied heavily or exclusively on parental for overnight management despite describing independence during the day. Overnight glycemic control was unpredictable and youth indicated their inability to hear or respond to alarms was predominant reason for parental involvement.
Subtheme	Relationship to nighttime dependence and unpredictability	Dimensional codes
Parental dependence	Youth with A1C<7.5% reported parents almost exclusively performed overnight diabetes and technology self-management. Often slept through parental overnight management. Participants with A1C>9.0% did not mention parental overnight management.	Reassuring to have parents surveilling glucoses and performing management
		Correct highs with insulin pump boluses or increased temp basal rate.
		Treat lows with carbohydrate intake (juice most often mentioned) sometimes with temp basal rate reduction or insulin suspension.
Fear of hypoglycemia	Overnight hypoglycemia very concerning for youth. Many described previous severe overnight hypoglycemic events	History of overnight lows
		Preparations for night to prevent highs and lows
Inability to react	Inability to “hear” CGM alarms primary reason for parental involvement overnight. Youth indicated they didn’t know why they weren’t able to hear or respond to overnight alarms.	Limited responsiveness due to hypoglycemia likely a factor
		Inability to hear alarms was primary reason for parental involvement
Device connectivity	Cloud-based CGM allowed parents to surveil glucose trends and receive alarms on their phones overnight. Youth stated their parents routinely woke up at a set time overnight to check glucose trends and assess the need for intervention.	Facilitated parental surveillance
		Phone used to view CGM trends/values
		Alarms for low, high, and predictive low overnight heard/viewed on parents phones
		Beyond the home, parents surveilled overnight CGM values/trends
It’s Really a Team
Definition: Youth described strong reliance on others for assistance, advice, and support (i.e., parents, teachers, coaches) despite describing their self-management and independent or nearly independent.
Subtheme	Relationship to It’s really a team	Dimensional codes
Parental support	Parents continued to play a pivotal role in diabetes and device management for youth. Most found the involvement to be beneficial and credited as pivotal to successful self- management however not all youth responded to parental advise.	Device connectivity facilitated parental involvement when youth away from home
		Solicited advice more critical/novel situations (i.e., illness) or technology management (i.e., alternate basal patter)
		Unsolicited advice sometimes in response to shared CGM data
		Night parental management of technology and diabetes was comprehensive for many
Community support	Team of support extended beyond parent; Most youth mentioned the support of others especially in the school environment. Youth described teachers who either had personal experience with diabetes or took a special interest	School is where received most additional support
		Teachers and coaches aware of the youth’s diabetes and supportive
		Accommodate technology (i.e. phones) that were often against school policy

Feeling Gross.

Interestingly, in hyperglycemia events, it was the glucose value in concert with the presence of symptoms, characterized as “feeling gross” or “crappy”, that triggered the youth’s attention. Youth described feeling “gross,” “crappy” or “yuck” in association with high and fluctuating glucoses. Symptoms of nausea, abdominal pain, and fatigue accompanied periods of hyperglycemia and labile glucoses.

There wasn’t really a whole lot of time that went by that day that I wasn’t thinking about it [glucose]. Mostly because I didn’t feel good, and I knew that was why. … when I don’t feel good, obviously I can’t stop thinking about that...That, that was not a good day. (17-year old girl) (Supplement 1)

Conversely the presence of hyperglycemia without symptoms did not necessarily incite concern or action.

Like here I was like 376, but I felt pretty much okay. I just knew that my blood sugar was high. And so there are times like that where I don’t really feel like I need to really take myself out of it,…. (a 17-year-old girl)

Focus on Diabetes.

The symptoms of “feeling gross” triggered concern about the glucose value, and compelled participants to withdraw from typical activities (i.e., “take myself out”) to focus efforts on their diabetes management. Several participants preferred low to high glucoses because it took less time to regain control whereas highs took several hours and more effort to correct. A 16-year-old boy described feelings of embarrassment, anger, and frustration over being forced to stop and manage his diabetes during football practice.

… and I go low during practice…I’ll have to sit for about 15, 20 minutes. And maybe my pump failed, and my sugar went - skyrocketed during practice…It’s happened before….I’ll be running. Then beep, my pump will go off. It’s very frustrating. Like it’s kind of embarrassing too. ‘Cause you try not to let it bother you, but it does. It always gets to me. And like it gets me mad, really mad. So, I have to go in the locker room, mess with my stuff, come back and all that. It’s really frustrating.

Setting Matters.

The setting in which the adolescent must manage their diabetes influenced the degree of attention and effort they dedicated to management. Symptoms were perceived as more distressing when experienced in an environment in which the adolescent was less comfortable or familiar (e.g., a social or sporting event).

Expect the Unexpected

Each participant described the unpredictability of their glucose referring to multiple instances in which they had little explanation for a low or high blood glucose value. Table 2 depicts the overlapping subthemes and dimensional codes comprising “expect the unexpected”.

Personal history.

They considered circumstances ahead of time, devising a reasonable plan based on prior experience, only to be bewildered when outcomes were quite different than anticipated. As one teenager in the low A1C group explained, “I mean, I keep all my factors the same, but my body is just different every day.” Despite efforts to think through previous experiences critically, they each shared multiple instances when the outcome was inconsistent with their personal history.

I ate some food, but usually when I get into the pool, I’m - like I go low so quick. That I was like I’m just gonna eat food - And then I think I’ll be fine …. And then I ended up being at 430, which was shocking because that doesn’t usually happen. (17-year old girl)(Supplement 2)

It’s just luck sometimes.

This unpredictability in glucose values was incredibly frustrating. Some explicitly acknowledged the role of luck, while others looked for rationale behind unexpected glucose values. A 16-year-old boy reflected that it often doesn’t matter what he knows as the actions he takes are often inconsequential and not directly tied to the subsequent outcomes.

But like sometimes it’s luck. Like it will just go sky high for no frickin’ reason or go - bottom out for no reason. It just - it happens…Just sometimes you have bad luck one day.

Within this element of providence or luck, participants described a personification of the diabetes or of the insulin administered - as if the diabetes was an entity with objectives of its own that could not be controlled.

And then, right here I gave one, two, three [boluses] within the span of maybe an hour and a half and it still spiked and came back down. So, sometimes it does it by itself. (16-year old boy)

Rollercoaster of treatment highs and lows.

The perceived unpredictable nature of glucose management was in part related to a reciprocal or rebound effect after treatment of high or low glucoses. A 15-year-old boy in the low A1C group described this rebound effect. “…you do one thing to fix a problem. Then another problem builds up and all that…..It’s a roller coaster.”

Knowing Why.

Unexpected outcomes were especially troubling to youth when they couldn’t ascribe a rationale for the high or low glucose outcome. Not knowing the cause also presented a self-management challenge in that the cause offered valuable information to help dictate their next action. A 17-year-old with low A1C described a period of hyperglycemia she could not account for.

So, if I know why it’s happening [a high glucose], then it doesn’t make me worry…if I forgot to bolus –I can look back and say, “Oh, I forgot to bolus. That’s why. Uh, oh, I’m sick. That’s why I’m high.” Then I’m like, “Okay. Well, we can just bring it down, and it’s fine.” It’s the highs that I’m like, “Where did this come from?” that makes me anxious.

Unplanned technology management.

Technology interactions like sensor changes or pump site changes were not considered bothersome when they were anticipated and planned for. Unexpected technology issues, however, were very disconcerting to adolescents especially when they were unable to control the setting.

Nighttime Dependence and Unpredictability

The overnight hours were marked by a sense of unpredictability and the unknown. Table 2 depicts the interwoven subthemes and dimensional codes in “Nighttime Dependence and Unpredictability”.

Fear of hypoglycemia.

Youth took considerable time and effort in preparing for the overnight period. They were particularly concerned about overnight lows; many recounted previous significant overnight lows.

And when I’m sleeping, getting woke up by my parents to drink something. I want to sleep. I don’t want to do all that. I’ve had a seizure when I was - during my sleep. My dad walked in when I was having a seizure… (15-year-old boy)

While youth described near independence in their daytime use of technology and self-management, parents performed most overnight diabetes management. Youth expressed solace in knowing that their parent was assisting with overnight management, explaining this as a “team” approach to management. Parents performed overnight correction boluses and set temp basals, often while their child remained asleep. Several participants stated that any overnight bolus seen on the download was undoubtedly performed by their parent.

Right here it looks like I must have went low. When my alarm goes off when I’m sleeping, I don’t wake up. So, my parents usually come in, and they make sure I don’t bottom out, and they wake me up, and I probably drink juice and all that.

Interviewer: So, who do you think set that temp basal? (Supplement 3)

Uh, it was either my mom or my dad. (15-year-old boy)

Inability to react.

Youth universally described their inability to “hear” CGM alarms overnight and cited this as the primary reason for parental involvement. They didn’t know why they weren’t able to hear or respond to overnight alarms.

Interviewer: Can you tell me about a time typically during the day or during the week where you’re worried about hypoglycemia?

During the nighttime, ‘cause, like my phone’s close to my bed. So, you would think I would hear the alarms, but I don’t, really…. So, my parents have to wake me up. (16-year-old boy)

A 17-year-old boy described an almost euphoric state associated with a recent morning low that slowed his ability to respond to the hypoglycemic event.

I remember one time I was just sitting there. I was, like, “my sugar is low.” And I just sat there for the next, like, 15 minutes. And it’s, like, “I got to go get something, like, now”….It’s just when my sugar goes low, I get kind of, like, numb and tingly and I just kind of, like, sit there a little bit. So, I’ll get up and I’ll push myself downstairs just because it’s like a really odd feeling. It’s like you - you’re just, like, really relaxed.

Device Connectivity.

The presence of cloud-based CGM played a pivotal role in overnight management. The youth’s cloud-based CGM allowed parents to surveille glucose trends and receive alarms on their phones and facilitated parental monitoring and intervention even when the youth was not in the home overnight. Youth stated they knew their parents were monitoring their trends overnight as an “extra pair of eyes”, which was reassuring to them.

She always wakes up in the middle of the night, she doesn’t, like, set an alarm or anything and get up. But, like, if she wakes up, she would always come in and poke me. Um, but now she can just look. And then, uh, obviously if she needs to feed me or whatever, then she’ll come in and take care of that. (17-year-old girl)

Sunday, that morning when I was at a friend’s house, uh, this is around 7:00, it looks like. My mom called me. And I was - I was pretty low obviously. I was 61. I tested again. I was 49. (15-year-old boy) (Figure 1)

Figure 1. Nighttime Dependence and Unpredictability, Device Connectivity, and Corresponding Device Download.

A 15-year old boy with A1C <7.5% described a hypoglycemic event while spending the night at a friend’s house. Earlier in the overnight period he had miscalculated carbohydrates for snacks causing his glucose to rise initially into the 300’s and then above the threshold of detection (>400mg/dl). His self-monitoring blood glucose (SMBG) fingerstick at that time was 448mg/dl. He boluses twice, about an hour apart and additionally set a temp basal increase in an attempt to correct the high glucose. He indiated he went to bed shortly after the second bolus. His mother, having received a low threshold alarm on her cell phone, called him around 7:00am, alerted him of the hypoglycemic event and promted him to treat the hypoglycemia. He indicated he was awaken by his mothers phone call and stated he had not heard his personal low threshold alarm. After treating the low by drinking juice and setting a temp basal reduction, he waited until the glucose rose to 96mg/dl and then stated he went back to sleep until noon.

Interestingly, while all of the youth in the low A1C group mentioned parental overnight involvement, none in the high A1C group mentioned nighttime parental involvement. A 16-year-old in the high A1C group discussed both an overnight high and low that appeared on the 5-day download. He did not seem bothered by overnight lows indicating they weren’t worrisome because he would just sleep through them. He recounted overnight hyperglycemia, a glucose peak of 599mg/dl in which he was awakened by symptoms of nausea, checked his glucose, and administered a bolus without ever informing his parents.

It’s Really a Team

While every youth characterized their management as independent, there still existed a strong reliance on others for assistance, advice, and support. Table 2 depicts the theme of “It’s Really a Team” along with overlapping subthemes and dimensional codes.

It’s a team effort. It’s a huge team effort. Like especially during school, my parents, people at school. And it - it’s also myself. (16-year-old boy)

Parental support.

Parental involvement was predominant and comprehensive in the overnight period but more supportive during the day. Youth acknowledged that the data sharing and connectivity between themselves and their parent often compelled them to reach out, acknowledge the glucose event or trend, and at times inform them of their self-management behavior.

Usually - like let’s say I go low, and I’m not with my parents. I know they’re getting alerts. So, I text them and say, I’m - I got it. I’m taking care of it. And they’ll say, okay, good job. (15-year-old boy)

Parental involvement was nearly ubiquitous with CGM device connectivity. Text messaging was mentioned by all youth as the predominant form of communication. Most youth characterized this surveillance as reassuring and positive (“.. an extra pair of eyes in case I miss it”) in assisting them in remembering self-management tasks. Parental input lessened during work hours. Even if the parent could visualize their child’s CGM, the youth felt their parent was often too busy at work to communicate with them

Several participants indicated that their parents’ ability to view their data also presented challenges. They were often faced with a barrage of communications alarms would inevitably trigger simultaneous text messages from their parent. One 17-year-old girl, who started CGM a few months prior to the interview, reflected on both the benefit and burden the device had for her mother.

I feel like she stresses out more that it’s constantly right at her fingertips. …. I think that it’s causing her more stress because she’s seeing it all the time. I was like, “Mom, I would eat and spike every time I ate for seven years. You just didn’t see it because I didn’t have this constant monitor. And now you’re losing your mind because you know.”

Independence in diabetes management was related to age, with older participants more likely to describe their management as fully independent. Parents intervened in more critical instances. One 17-year-old girl described a period of prolonged hyperglycemia during her school day in which her mother was in frequent contact via text message.

This morning when I was over 300 for, like, six hours –she [her mother] was constantly, like, correcting it and correcting it and correcting….And she’s like, “Just give yourself an injection. It’ll go faster and that way we can make sure that you get it” –

Parental communications did not, however, stimulate all youth to complete self-management behaviors. One 16-year-old with a high A1C discussed his mother’s communications “Yeah. Always calling me. “Check your sugar today?” Like I could probably go through here [his phone] in the last week, I probably have 25 text messages.” When asked if he then checks he stated: “I mean, yeah, if I’m at home, but if I’m busy I’ll just put it back in my pocket and do it later.”

Community support.

Assistance and encouragement extended beyond parental involvement to teachers, school personnel, coaches, and friends. Several participants mentioned additional difficulties self-managing at school and during extracurricular activities especially during periods of hyperglycemia or glucose variability. They affectionately mentioned teachers who provided additional support or empathy.

It’s all good. My coaches know everything. My teachers know. Like I said, during school I had xxx [a friend], and I have another teacher.... She’s the one that had my crackers and my juice and stuff like that. And so, I can just walk in there, get my juice, walk back out, and it’ll all be good. (15-year-old boy)

Some youth reported school policies that were antithetical to diabetes self-management, such as prohibitions on cell phone use and carrying snacks. In most cases, the school or individual teachers amended such policies to aid the youth. Several mentioned being singled out by other students for carrying phones or diabetes technologies that resembled phones.

And they would get mad at me being like “C has a phone out”…Because we’re not allowed to have our phones out at school. They like to tell on me for having a phone if I’m doing something for diabetes. (10-year-old girl)

Discussion

The results of this qualitative study of adolescent and pre-teen’s diabetes device self-management provides valuable insight into the way youth vary their behaviors in response to glucose values. Our examination of symptomatology within the context of everyday life for youth with T1D is critical in furthering understanding of self-management for this population.[24]

Youth’s descriptions of diabetes self-management during “bad” days are substantiated by our previously published quantitative findings from the larger sample of pre-teen and adolescent insulin pump users with CGM devices.[13] When daily time in range (TIR) was examined each insulin pump self-management task was, in fact, negatively associated with TIR.[13] On days when the youth experienced greater TIR (i.e., better glucose control), they performed fewer self-management tasks. Thus, when the youth experienced less TIR (i.e., poorer control), they increased their self-management efforts. These qualitative findings provide context, explanation, and illuminates the corresponding experience of increased effort on “bad days.” The qualitative analysis further contextualizes this relationship between self-management effort and glucose control as youth’s response to their glucose may be mediated by symptomatology (“feeling gross”) and setting. These important findings are valuable in establishing that youth are in fact responding to their glycemic trends and offer a potential avenue for future intervention aimed at improving glycemic control. While the focus of inquiry centered around technology self-management, youth’s experience captured within “bad day” could seemingly be generalized to non-technology users and future inquiry should extend to the larger population of youth with T1D.

The theme, “expect the unexpected,” helps to illuminate the youth’s experience of daily glucose variability. In Strand et. al. (2018), youth described the phenomenon of glucose fluctuations as a ‘roller coaster’, which is consistent with the characterization by youth in our study.[13] This phenomenon of unpredictability is reflected in residual variance analysis from the quantitative portion of this mixed methods study. The TIR for youth in this study varied significantly from day-to-day with a standard deviation (SD) of 18.6% (4 hours and 40 minutes) (SE0.455, 95% CI [17.690, 19.473]). Thus, even on days that might be considered typical days for that youth, their percent TIR fluctuated by nearly 5 hours.[13] The qualitative findings provide an understanding of what this experience of daily glucose variation and unpredictability feels like for youth. These novel findings highlight the burden of disease self-management.

One unexpected finding was the degree of similarity in self-management experience among youth with high and low A1C. Maximum variability sampling was chosen in part with the expectation that self-management experiences would differ between the two groups. In fact, youth’s description of self-management including their characterization of good vs. bad days and their frustration over unexpected outcomes were strikingly similar. While youth’s described experience was similar, the corresponding quantitative data from phase 1 of this mixed methods study showed frequency of self-management behaviors (i.e., bolus, SMBG, advanced feature use) varied significantly between youth with high and low A1C.[13] Further exploration of this difference between youth’s experience and application of self-management behaviors is warranted. It may be possible that youth with low A1C are simply responding to glucoses at a lower threshold or perhaps youth with low A1C experience symptoms at a lower glucose value prompting earlier action than youth with high A1C who experience more frequent hyperglycemia.

Although youth universally described their management as independent, they relied on parental surveillance and management in the overnight period. Schilling et. al. (2006) described parents’ overnight self-monitoring of blood glucose testing as a form of ‘pinch hitting’ in which parents allowed youth to take a respite from the burdens of self-management.[25] In Bergner et. al’s (2018) research, the majority of parents reported checking their teen’s glucose at least once overnight; however, specific overnight management (i.e., insulin administration, hypoglycemia treatment) was not reported.[26]

Use of Cloud-based CGM allowed parents to survey trends and respond to alarms received on their personal phones. The inability of youth with diabetes to hear overnight alarms was the predominant reason youth cited for parental involvement. It is likely that a degree of cognitive blunting or impairment associated with hypoglycemia plays a role in the inability to respond to alarms. In recent work examining adolescents’ and caregivers’ experiences with CGM, alarms were a prominent point of discussion; however, the inability to hear alarms was not discussed.[26 27]

It was interesting that only youth in the low A1C group described overnight parental glucose monitoring or management. It is possible that overnight parental surveillance is simply a proxy measure for the degree of parental supervision and co-management or that overnight management by parents in the correction of highs and lows had a meaningful impact on the overall glycemic control. Certainly, the burden of overnight surveillance and management experienced by these parents is significant and substantiated by previous research.[26] Hybrid closed loop systems have been shown to be especially effective in improving control in the overnight hours[28 29] and may decrease overnight parental burden in the future.[30] The degree to which youth in the study relied on parental intervention and surveillance overnight also speaks to the difficulty experienced by youth as they transition to young adulthood.[31]

Other studies have examined the distribution of self-management and transition in diabetes management from parent to youth.[29 32–35] Continued parental involvement throughout adolescence is associated with better glycemic control.[33 34] Our findings related to the role that device connectivity plays in youth’s management and parental involvement provide important insights into the way parents and youth interact with diabetes technologies and work together to manage the condition. In recent qualitative research by Lawton, et. al. (2016) examining patient and caregiver experience with CGM, the availability of data for the caregiver was discussed but how data facilitated or changed self-management behavior was not reported.[27] In a meta-synthesis of qualitative CGM literature, parents reported CGM decreased parental stress but youth reported parental surveillance could lead to negative interactions.[36] Future work should focus on quantifying the degree of parental involvement and the effectiveness of such measures as well as the long-term implications of continual parental supervision into emerging adulthood.

The lack of minority representation is a clear limitation of the study and results may not be generalizable non-white populations. Participants were recruited from the larger cohort of quantitative participants who were also predominately white (92.5%). While T1D does disproportionately effect whites,[1] the homogeneity of our sample is reflective of current clinical practice in which minority youth with T1D are prescribed diabetes technologies at significantly lower rates after controlling for other demographic, socioeconomic factors and A1C.[37] In the future, convenience sampling should be augmented with targeted recruitment of minority youth to achieve sufficient representation.

Our inclusion of diabetes device downloads as a visual and illustrative aid in this qualitative work is novel. The downloads elicited detailed descriptions of the youth’s self-management behaviors and daily activities, and evoked discussions about their emotional response to glucose trends. During interviews youth, initially expressed some mild anxiety, fearing that their data might not be useful or “good enough” for research. It was important to clearly communicate that for research purposes, we were seeking download data that would help us understand the real-world experiences of youth with diabetes, and the intent was not to critique their self-managed as you might in clinical practice.

Conclusions

Our findings highlight the dynamic nature of diabetes technology self-management and the multitude of factors that feed into a youth’s self-management choices. Our findings surrounding the way youth vary their behaviors in response to glucose values is novel, warrants further investigation and offers a potential avenue for future intervention aimed at glycemic control and technology self-management.