Abstract
Background:
Bihormonal fully closed loop (FCL; “artificial pancreas”) reactively manages glucose levels by adjusting insulin or glucagon delivery based on sensor readings. Previous quantitative research has demonstrated FCL group-level efficacy in terms of glycemic and person-reported outcomes. Less is known about specific user experiences and the impact on different life domains.
Aims:
To qualitatively examine motives and user experiences related to this bihormonal FCL system among adults with type 1 diabetes in a real-world setting. This included (a) reasons for considering FCL use, (b) positive or negative impact of FCL, (c) the process of adjusting to the system, and (d) the perceived benefit–burden trade-off.
Methods:
After close of a 12-month single-arm trial (FREE 1), a subsample was interviewed (n = 12). Eight ongoing users were selected based on change patterns in the quantitative trial data. Four people who discontinued the system were selected based on duration of FCL use and discontinuation reasons. Data were analyzed using thematic and narrative analysis.
Results:
Participants voiced a combination of personal and altruistic considerations in the decision to start FCL. While there was a variety in individual impact, benefits (improved glycemic outcomes, feeling more like a person without diabetes) mostly outweighed burdens (e.g., hassles related to glucagon replacement, multiple components, alarms). FCL was a positive life changer for ongoing users, despite continued effort required for technology management. Finding a working relation with the system was challenging and took time, with participants reporting different coping strategies. Discontinuation reasons presented early, with an accumulation of negative experiences leading up to the decision to stop.
Conclusions:
FCL users experience improvements mostly in glycemic and often also person-reported outcomes, with continuation of use depending on reconcilability of technology hassles. To facilitate FCL adoption and long-term use, expectation management and continued support from the manufacturer and trained diabetes care teams seem key.
Introduction
Clinical guidelines recommend automated insulin delivery (AID) systems for day-to-day management of type 1 diabetes, based on overall superiority in medical outcomes.1,2 Commercial single-hormone hybrid closed-loop (HCL) systems partly automate pump-based insulin administration by using a sensor-driven algorithm for basal and correction insulin. 3 With HCL, randomized controlled trials and real-world studies have demonstrated an increase in average time in range (TIR; 70–180 mg/dL or 3.9–10.0 mmol/L) of 5%–17%, without increasing hypoglycemia risk.4,5 Users have also reported quality-of-life benefits, including reduced overall diabetes burden.6–9 Yet, these systems also have limitations, classified as “physiological” (e.g., sensor time lag related to measurement in the interstitial fluid), “technological” (e.g., system failures), and “behavioral.” 2 The latter mainly relates to situations where user action is still required for proper system functioning, such as manual mealtime and correction boluses and changing target glucose settings around exercise.6,10
Fully closed-loop (FCL) systems have been developed to decrease or obviate the need for meal and exercise announcements.3,11,12 Insulin-only FCL focuses on overcoming mealtime input by using ultrarapid insulin. 13 While users experienced reduced burden of management around mealtimes, they also varied in their postprandial glycemic outcomes. 14 A bihormonal approach is the Inreda AP® (Inreda Diabetic B.V., the Netherlands), delivering insulin lispro and glucagon reactively guided by glucose sensor readings. Results of a single-arm trial were published recently (FREE 1; Fully closed-loop glucose Regulation; Evidence and Experience in real life), 15 with an average TIR increase of 25% and self-reported diabetes distress decreasing significantly over the 1-year treatment period among 71 adults with type 1 diabetes (of an initial 79 who started the intervention) in a real-world setting. Ten trial participants (13%) discontinued FCL, slightly higher than the 5%–10% discontinuation rates found for recent single-hormone commercial or open-source systems.16–18 This signals the need to gain a more in-depth knowledge of user experiences.
Therefore, the present qualitative study among a subsample of FREE 1 study participants aimed to better understand (a) reasons for considering FCL use, (b) the perceived positive or negative impact of FCL use on the burden of living with and self-managing diabetes, (c) the process of adjusting to the FCL system, and (d) the trade-off between FCL benefits and burdens, leading most people to continue and some to discontinue using the system.
Materials and Methods
Study design and participants
Details on the FREE 1 study have been published previously. 15 In short, participants were recruited from eight hospitals in the Netherlands between June 2021 and March 2022. At recruitment start, only 1–2 participants could be trained in FCL use at the same time due to COVID-19 restrictions. Inclusion criteria were a type 1 diabetes diagnosis, age 18–75 years, use of real-time or intermittently scanned glucose monitoring for at least 3 months, and being willing and able to start with the FCL system as part of regular care. Exclusion criteria were (planned) pregnancy during the trial period and conditions that the local clinical investigators judged to interfere with participation or evaluation. The present qualitative analysis was a preplanned substudy of FREE 1, involving interviews about participant experiences with FCL treatment (January–April 2023).
FREE 1 followed the most recent version of the Declaration of Helsinki and received approval by the Medical Ethics Review Committee of the Martini Hospital Groningen (nWMO21.03.012, 2021-046). All participants provided written consent. In addition, written informed consent was obtained separately for the interviews to ensure that participants were fully informed. The FREE 1 trial is registered in the Dutch Trial Register (NL9578).
Procedures
After the 1-year FREE 1 trial was closed, a subsample was invited to participate in semistructured interviews. To maximize variety in experiences, eight people who had continued with the FCL (“ongoing users”) were purposively sampled by G.N., T.J.P.J., and M.K. based on different constellations of change in glycemic and person-reported outcomes as observed in the quantitative trial data. To this end, subgroups with similar quantitative experiences (from which interview participants were selected) were determined using K-means cluster analysis based on change scores of the glycemic and person-reported outcomes (difference between last follow-up and baseline assessment). Two of the initially approached ongoing users did not participate due to unreachability or personal circumstances; they were replaced by a priori selected participants with similar profiles as backup. At the time of the interview, all eight ongoing users were still using the FCL, as FREE 1 participants received continued reimbursement of the system after the trial as part of regular diabetes care. 15 Four people who had chosen to discontinue using the system during the trial were purposively sampled by G.N. based on duration of FCL use (0.75, 6, 7, and 10 months) and initially stated reasons for discontinuation. All the initially approached people consented. Compared with people from the FREE 1 cohort who were not interviewed, interview participants had a higher median baseline TIR with a larger interquartile range (66.9 [46.9–76.5] vs. 55.5 [41.9–65.1]). This fits with our purposive sampling strategy to obtain diversity in experiences, irrespective of the absolute number of people in the FREE 1 cohort with similar experiences. End-of study TIR was similar in both groups (mean 81% vs. 80%).
Assessments
The semistructured interview guide (Table 1) included mostly open-ended questions asking about expectations, (dis)advantages, and perceived trade-off, impact on different life domains, wearability, support, satisfaction, and recommendation regarding FCL use to other people with type 1 diabetes. The interviews with people who discontinued also zoomed in on the process leading up to the decision to stop using the FCL. All interviews were conducted by G.N., a diabetes psychologist and behavioral researcher experienced in qualitative methods. The interviews were conducted through Microsoft Teams; audio of the video recordings was transcribed ad verbatim. Interviews took approximately 30 min as per protocol. To characterize the subsample of interview participants, baseline medical record data were used.
Questions Included in the Semistructured Interviews with Ongoing Users and People Wh◦ Discontinued the Bihormonal Fully Closed-Loop System
FCL, fully closed loop.
Statistical analyses
Qualitative content analysis, narrative analysis, and the constant comparison method were used to analyze interview data (Atlas.ti Scientific Software Development GmbH, Germany), parallel to data collection. Themes were identified and labeled (open coding), and consequently grouped in categories based on their relatedness (axial coding). Overarching, central themes were then identified (selective coding). In coding, we used both pre-existing frameworks from the broader technology literature (deduction) and developed new notions based on the data at hand (induction). The single-coder approach (G.N.; diabetes psychology scientist-practitioner with technology expertise) accommodated limited study resources.19,20 This approach is common in the social sciences and allows deep interpretative insights. Strategies for safeguarding rigor were researcher reflexivity and discussion of results among the wider research team. Initial analysis results were discussed with T.J.P.J. and L.B.V., leading to a more balanced presentation of benefits and burdens and distribution of quotes across participants. The calibrated analysis results were then reviewed by F.J.S., J.H.d.V., and A.C.v.B., resulting in minor relabeling and reordering of data. The other coauthors critically reviewed the resulting report. Approximately half the coauthors are MDs working in clinical practice with people using FCL.
Results
A total of 12 FREE 1 participants were interviewed. Table 2 displays sample characteristics.
Sample Characteristics
As per the study inclusion criteria, all participants were using glucose sensor technology.
n = 2 used a single-hormone hybrid closed-loop system.
Reasons for considering FCL
Ongoing users and those who discontinued the system reported similar reasons for considering FCL. Participants generally expected and hoped for the system to make meaningful changes in their life (improved health, increased safety, less diabetes interference, less management burden, increased freedom).
That was the biggest motive to start. You are free, you can come and go as you please. You don’t have to think each night when you go out: what am I going to drink, how much do I have to inject, how often do I have to check each hour. That would all be gone. You could live like a king, so to speak. [P3, man, discontinued]
For two ongoing users, the decision to start was spurred by explicit dissatisfaction with their previous treatment modality or arose from the sentiment that FCL “could never be worse than that.” Often, hopes for personal gain went hand-in-hand with a sense of curiosity, a taste for innovation, or altruistic motives (“moving science forward,” “helping others”). The two initial HCL users (both ongoing FCL users) combined an affinity for technology and science with hopes of reduced diabetes burden, in particular related to managing meals and exercise.
Impact of FCL use on the burden of diabetes
All participants (except P7, discontinued after 3 weeks) reported at least some benefits (Table 3), all related to becoming “more like a person without diabetes,” Weaving through the narratives was the observed glycemic improvement, such as unprecedented HbA1c, TIR, reduced glucose variability, and excursions. Two participants mentioned improved hypo-awareness (which did not correspond with quantitative changes in GOLD scores; objective HbA1c increased).
Benefits Related to the Use of the Fully Closed-Loop System
TIR, time in range.
So now I don’t have any hypos anymore, almost no hypos, but when I do have them, I feel them acutely. I really feel like a bus is running over me [smiles] and I’m knocked out. So in terms of hypos, the FCL really delivers. [P1, woman, continued]
One participant mentioned it became easier to handle the destabilizing effect of the menopause on glycemic outcomes. Several participants stated that they were more able to lead the life they wanted, feeling less “ill” and more as themselves. People experienced more freedom, time, flexibility, and spontaneity. They felt more fit and able to keep up with others without type 1 diabetes. This translated into increased mobility and social participation. Participants also described to be less self-critical and less worried about hypoglycemia or glucose fluctuations (also for their family members). The hands-on and cognitive workload of direct glucose management became less, often with more satisfactory results. Most participants also described needing less contact with their health care providers, themselves becoming experts in FCL settings. If needed, they preferred troubleshooting with the manufacturer.
For several outcome domains, differences in impact across interviewees were mentioned. This included sleep (less hypos vs. alarms), weight, and eating.
I sleep much better. Now I can just lie down in bed and go to sleep. That wasn’t the case before. I always had to… I was a very light sleeper, because I was afraid of having a hypo in the middle of the night. When you’re not doing anything, why would you have a hypo? But I did have them. […] And my parents and people who haven’t seen me in a while say, “You look much better.” Because I used to have really dark bags under my eyes. I don’t have those anymore, because I’m sleeping. [P6, woman, continued]
At night, the alarms drove me crazy. [laughs] Because the system has two sensors, which have to… One checks the other, and when there is too much difference, the system says, “Please do a fingerprick, because it’s not right.” Or it needs to know if it’s right. [P7, man, discontinued]
A big plus is not having to bolus anymore. […] You don’t have to count the carbohydrates. […] After your meal you can simply do what you want, or clean up, or spend time with your kids. [P11, woman, continued]
For a fellow FCL user it worked so well, his glucose levels were so stable he could take regular coke and apple pie. And I was like “no, I really can’t eat that.” I don’t respond well to certain breads or grain products, my glucose struggles with it. So I still mind what I eat. [P9, woman, continued]
I did not use to feel limited in eating what I wanted due to diabetes. Now I feel it even more so. Because the FCL has a hard time dealing with meals, I feel more limited. Because I’ve lost my steering method. Previously, I could make adjustments and correct. And that’s not possible anymore. So I feel more limited now. [P1, woman, continued]
For ongoing users, expectations about benefits came true at least in broad terms while burdens were experienced as manageable. In contrast, people who discontinued the system mostly reported that their overall expectations were insufficiently met, often combined with the occurrence of unexpected burdens that proved irresolvable.
Making it work; the adjustment process to the system
The main focus of diabetes burden shifted to challenges related to technology management (Table 4). Participants described a challenging first period with the FCL system, stressing that the start was hardly “plug-and-play.” The experience of a rough first start was in contrast to expectations, partly fueled by overly positive portrayals in the media, which was explicitly mentioned by two people who discontinued. It took time to find the “right” settings for stabilizing glucose levels, incidentally overshooting to a severe hypo. Simultaneously, the system required consistent user maintenance.
Burdens and Challenges Related to the Use of the Fully Closed-Loop System
At a certain point I was close to crying, I didn’t know what to do anymore. The alarm went off again, and I thought, “I’ll never learn, I don’t know how to do it.” [P5, woman, continued]
From the outside, everything looks perfect. Also when you hear [inventor] on tv, it is almost perfect, but reality is somewhat, er… […] A lot of things are still needed.” [P10, man, discontinued]
I always say “I have to take proper care of the FCL, the FCL takes care of me. But it really is a diva. [smiles] A princess on the pea, so to speak. And glucagon plays a big part.” [P1, woman, continued]
A common thread running through personal narratives were the alarms of the system (e.g., related to replacements, connectivity, sensor discrepancies, occlusions), which this group of experienced sensor users found frequent, persistent, and disruptive, especially in the first period.
And when you’re walking, you can do it [respond to alarms], when you’re at home too, but when you’re in the car, and you’re just driving, at some point halfway through, that thing goes off, [swears], again needs a fingerprick… You have that to some extent with the regular pump as well, but not as extreme as with the FCL. [P10, man, discontinued]
In finding a working relation with FCL, obtaining positive results (“hard” glycemic data, life changes), a system working as it should, and growing self-efficacy provided a sense of trust.
If I do all those things [timely replacements, rotating application places], the FCL works for me. I think I am lucky. I don’t react to adhesives, my sensors work for a long time. My TIR is 95% or so. [P2, woman, continued]
Participants described the need for finding their own way with the system, which sometimes differed from the manufacturer’s recommendations. For example, using alternative adhesion sites, hastening material replacement, limiting on-the-road spare parts, switching off alarms, adding insulin, and eating extra. Practical solutions such as clothing accessories for carrying the device helped integrating the system in one’s life. Beneficial support was experienced mostly related to practical help and troubleshooting assistance from the manufacturer during people’s learning process. Psychological counseling was mentioned as helpful by people who stated to have had more difficulty engaging with their diabetes or had difficulty “letting the algorithm do the work.”
I felt with my back against the wall. I saw it went wrong, but I couldn’t interfere, because the FCL had to handle it. […] I talked with my psychologist, and she gave me some exercises, to experience that what I feared would not happen, that I wouldn’t pass out or anything. […] It gave some peace of mind, learning to adapt to it in a way. [P9, woman, continued]
Finding a workable way of relating to the system took time and effort, requiring persistence, hands-on learning and some level of tolerance for dealing with the disadvantages of the FCL.
You’ve got to have space. For it to wake you up in the night, for it to go completely wrong. Or for not bringing spare batteries when you’re far away. For always having to carry a bag with supplies. [P2, woman, continued]
Continued use took time to habituate, resulting in different forms of a human–system partnership. These ranged from collaborative to “love–hate” and distanced coexistence.
I received the FCL and thought “oh my Lord, what a happening and what a hassle. With those vials and oh… am I not putting the insulin in the glucagon…” But it is a matter of experience. And now… For example, if the alarm sounds, I take the FCL, while continuing chatting. Everything sorted again. [P2, woman, continued]
There are days when it’s drama all over, I’d really like to throw out the entire system, because there are many alarms. And there are days I can work all day, without a single alarm. Without thinking about my diabetes. Without thinking “I don’t feel well” or “is this diabetes-related.” [P1, woman, continued]
Everybody comments [on Facebook, hoping for FCL] “Then I can finally live again.” Then I think: “Then what are you doing now?” If I had to turn it in today, I would turn it in, and go back to pump therapy. Only, you hand in a piece of TIR. 50% in my case. [P12, man, continued]
Weighing of benefits and burdens
The importance of personally balancing the pros and cons of FCL was stressed.
Among ongoing users, people differed in how they appraised and dealt with the hassles and burdens. Six participants tolerated the burdens they encountered, or found them “hardly worth mentioning” compared with the benefits. They reported to cope with burdens by minimization, optimism, resilience, and benefit-finding.
There are hardly any burdens, you know. Only, just like I told you, glucagon is a little… But that is just because… I want to do it quickly, and you can’t do it that quickly. For me, there are no real burdens. [P5, woman, continued]
You have to work for it. But in return, there are major health benefits. [P4, woman, continued]
And I’m not the kind of person to get stuck in the negative. If I’ve had one bad night, I think “oh well, better luck tomorrow night.” [P9, woman, continued]
These were the ongoing users who did not want to lose the FCL, and would be willing to sometimes go to great lengths to keep the system.
It would be a disaster, seriously, if it would end tomorrow. Yes, I am being sincere. Then I would… I would become depressed [laughs]. In a manner of speaking. [Having to go back] … to something very impractical. To something very old-fashioned. To something… Back at square one. […] Of course that is not true. Because of course the latest systems… But that is not closed loop. [P2, woman, continued]
I have now worn my glucagon for 24 hours. When I am at home, I am ok with wearing it for several hours more, until it really starts to occlude. […] Just to save glucagon. [P5, woman, continued].
For two participants, benefits hardly outweighed the burdens, mostly driven by reflections on the future.
I will continue, because I know this is better for the future, also for my health. And what comes with it in terms of lower user friendliness and occasional alarms, and the daily [hassles], the changes you have to do every day, I need to go through those things for it, I try to endure them. [P11, woman, continued]
Glucose levels are of course more stable, so you feel more fit, your TIR is high. But I see more burdens than benefits. And still I have to choose my two benefits over my multiple burdens. […] And yes, at a certain time it was: do you choose “let the good times roll and I don’t give a ***.” Or do you choose to make sure your TIR is around 75%. [P12, man, continued]
Both ongoing users and participants who discontinued the FCL reported moments of doubt, often luxated by an accumulation of negative experiences.
In the beginning there were moments where I thought “what have I gotten myself into?” Especially when you were in the supermarket… and the alarm kept on going… and people were watching. And at inconvenient times. [P11, woman, continued]
One or two weekends, it can happen, but a third or a forth? Then I said [to the manufacturer] “I don’t know what you want, but get it off of me.” And within thirty minutes they were at my door with the [glucose] data, saying “it’s up to you.” So. Then I got a new device, this time working properly. [P12, man, continued]
For two ongoing users and three people who discontinued, the decision to (dis)continue FCL stretched over months, “balancing on a knife’s edge” [P1, woman, continued]
I thought, are these burdens really substantial, are they big enough to quit? Because no treatment is perfect. There is always something. [P8, woman, discontinued]
Among people who discontinued, the will to make it work was thwarted by early-onset untenable side effects (itch, scalds) or an accumulation of unresolvable burdens over time. These commonly included alarms hindering other people. Frustrations built up gradually. In longer periods of doubt whether or not to continue, hope was provided by alternation of difficulties with relative serenity, and anticipation of design improvements. However, in the end, these did not weigh up against the burdens.
So in the end, taking all situations together, we said after six months “it’s time to decide, what shall we do?” And I had to go abroad for my job. In one of the periods where the system wasn’t working as it should. You have to take a lot of supplies with you. No, let’s just say: “We’ve tried, it didn’t work, let’s quit, and continue as before.” [P10, man, discontinued]
A common theme among people who discontinued was a perceived system–person mismatch. This mainly focused on singularities in bodily functioning (delayed or irresponsiveness to food, glucagon, insulin) or situations where the system could not keep up with strong food- or exercise-related glucose fluctuations. One participant also described a conflict with the preferred coping strategies.
My diabetes is my problem. I have to deal with it and other people should not be bothered by it. And with the FCL, that is not so easy, because if there is an alarm, everybody around you is bothered by it. […]. [P8, woman, discontinued]
All 12 participants, including those who discontinued FCL, would recommend the system to others, wishing other people with diabetes a life with fewer complaints and a higher quality of life. They praised the system or stressed the importance of keeping an open mind, giving it a chance and staying positive. People who discontinued were satisfied with the decision to stop using FCL, acknowledging all systems have their own disadvantages. Participants did express some strong caveats in safeguarding positive impact. In considering the human–technology match, presuppositions focused on groups who might not be able to work optimally with the system (older age, less digital skills) or who would benefit par excellence (strong glucose fluctuations, high mental burden, younger age). Four participants called for careful personal consideration of FCL necessity, suggesting it might not be the obvious choice for all.
If you can manage things just fine with your current treatment, with the pump or injections, if you are satisfied with it, or achieve stable glucose levels, then I wouldn’t recommend it. And that’s purely because the FCL is somewhat less user friendly. [P11, woman, continued]
You have to be able to switch your mindset from strict personal control to releasing everything. You have to listen to the device instead of telling it what to do. [P10, man, continued]
Several participants stressed the importance of awareness of possible disadvantages before deciding whether or not to start FCL.
I think many people welcome FCL with open arms. Also judging from responses on Facebook. They all say “yes please, because then I will have my life back on track.” And then I think: “You don’t even know yet how this thing works, generally speaking.” [P12, man, continued]
In terms of results, participants recommended tempering of expectations and not expecting miracles or a panacea. Diabetes was still there, albeit glucose fluctuations were less common and easier to cope with. For one ongoing user and one participant who discontinued, benefits were limited to glycemic outcomes. Where glucometrics fell short of expectations (e.g., after-meal peaks, work-related hypos), people usually did continue when FCL effected positive changes in quality of life.
For people who discontinued the system, reconsidering FCL use would require improvements in design (smaller device, component integration, sensor adhesion), functioning (connectivity, faster measurement, faster hormone effectiveness), and personalization (input variables, combined lifestyle advice). Ongoing users similarly prioritized upgrades related to wearability and personalization (alarms, experimentation safety mechanisms). They also advocated for reduced workload (change frequency, supplies, ready-to-use glucagon) and increased user access to glucose readings.
Discussion
Among this first cohort of users of a bihormonal FCL system in real-world type 1 diabetes care, benefits mostly outweighed burdens. FCL represented a positive life changer, although technology management took continued daily effort. There were different strategies for coping with FCL burdens, with ongoing use depending on whether people found a personally acceptable working relation with the system. Discontinuation reasons were often present early and accumulated over time.
All participants using the system beyond the start-up phase described benefits, mostly in terms of unprecedented glucometrics and often also related to quality of life. This corroborates quantitative group-level results from the overall trial, finding improvement in glycemic as well as person-reported outcomes, including diabetes distress and general emotional well-being. 15 Similar positive impacts have been reported in previous qualitative studies among adults with type 1 diabetes using single-hormone commercial or open-source HCL or FCL systems.6,7,9,14,21 As only two interviewed participants had used an HCL system before FCL start (out of five people in total in the FREE 1 trial), the present findings do not allow a solid distinction between benefits that were FCL-specific versus general to AID systems. Indirect quantitative comparison of glycemic outcomes between FCL and HCL favors the former.4,5,15 The ongoing Dual-Hormone Fully Closed-Loop for type 1 diabetes (DARE) randomized parallel-group trial will provide data on a head-to-head quantitative comparison of FCL with usual care treatment, including HCL. 22
Notwithstanding overall positive notes, the interviews also highlighted that not all participants benefited in the same way. Improvements in glycemic and person-reported outcomes did not necessarily co-occur, making satisfaction with FCL dependent on the personal value attached to either. This signals the need to include both dimensions in technology evaluations. 23 Furthermore, benefits differed in the specific areas of improvement. For example, for several participants, this related to not having to announce meals and exercise anymore, while for others, meals still required user vigilance. Studies examining user experiences related to insulin-only FCL or HCL using meal announcements without carbohydrate counting similarly reported continued struggles for some participants, particularly related to specific meal and exercise schedules.9,14 With respect to diabetes burden, people described a shift in focus from glucose to technology management. Challenges mainly applied to AID systems in general,7,24 with some being more pronounced for FCL (e.g., wearability). Bihormonal FCL also introduced glucagon-related burdens, mainly related to hassle and for some people nausea, a known side effect occurring mostly in the first period of use.12,15 Some reported system drawbacks are essential for safe use of the FCL system, including daily replacement of glucagon infusion set and cartridge, the necessity to wear two glucose sensors, and a multitude of alarms. With two glucose sensors and two infusion sets, alerts for material replacements are more common for FCL than for other treatment modalities. Furthermore, the alarm-based safety approach includes alerts in case of sensor discrepancies to safeguard the quality of sensor input. User and safety experiences have allowed some alarm modifications in the most recent software version; further tracking of these experiences remains necessary to optimize the trade-off between user friendliness and safety. The abovementioned reports about technology burdens also stress the importance of supplementing general FCL (or HCL) education with broader support, in terms of technical troubleshooting as well as learning to cope with (as yet) unmodifiable system burdens.
What differed between people who continued versus discontinued using the system was the extent to which they managed to find a working relation with the FCL system. In the general literature related to wearing external devices, this is marked as a crucial stage in the adjustment process.25,26 It includes a “probationary” period, typically lasting weeks in which functioning is scrutinized and outcome depends on results. 27 For several of our participants, this process extended over several months, describing a continuous weighing of benefits and burdens, 28 and a variety of coping strategies to tip the balance toward the positive (e.g., burden minimization, focusing on benefits). Adjustment to technology is best considered a dynamic process, and what is workable in the human–technology tension field may differ across individuals. To secure longer term adoption, support initiatives need to allow for such interindividual flexibility and establish at least a basic level of human–device trust. 29 Potentially helpful elements include prevention of accumulation of burdens, for example, through easily accessible expert technological support, and making results tangible (e.g., visualization of glycemic parameters, discussing changes in life domains).
The process of learning to work with the FCL was described as challenging, and not always met a priori expectations. As described in other qualitative studies focusing on first-generation users of innovative diabetes technology,14,30 many of our participants combined hopes for personal gain with more altruistic considerations, which appeared to have fueled them to put in extra effort to make it work. This makes it all the more important to provide potential future users with a realistic picture of the (im)possibilities of the system. From the broader diabetes technology literature, it is known that expectations play a key role in the sustained adoption of devices. 28 Information provision may help in setting realistic expectations and avoid disappointment or feeling misled.10,31 In fact, several participants indicated to appreciate the attention to benefits as well as burdens given during the manufacturer training close to system start. However, to avoid or reduce disappointment and frustration early on, nuanced information about FCL use needs to be available and part of the shared decision-making process. 28 Furthermore, commercial messages about system benefits—particularly related to automation of glucose management and related messages of freedom—may feed into misperceptions if not combined with context about the user actions that are needed for the technology to function properly. 10 The Diffusion Innovation Theory posits that “success stories” may work for people who desire role models (often early adopters of technology), but not for people who prefer experience accounts from others who like them are more cautious in their assessments. 32 Results from this study on FCL experience underscore the need for nuanced and varied information, based on lived experience. As realistic expectations cannot fully be formed by a priori information provision alone, these may be complemented—where feasible—by temporary user exposure to the system, for example, try-out periods or virtual reality approaches. 33
Some participants voiced their own ideas about for whom the FCL might (not) be a match. However, predicting “success” is far from easy. The interviewed people who discontinued during the trial described a perceived FCL-person mismatch that became apparent only while using the system. Realistic expectations and being able (and learning) to resolve or cope with FCL-related burdens would appear important. In this light, the discussion around user selection—that is, who is “suitable” for FCL in terms of characteristics such as perfectionism and frustration tolerance—can better be reframed toward redesigning the level of coaching and support that can be offered as part of routine care. Some people will need relatively limited help and others ongoing support, for example, in finding an appropriate level of trust in technology. 29
Burdens and hesitations about whether to continue with the FCL were often present early, which stresses the importance of a personalized coaching period after system start. The present 2-day run-in training and the ensuing 8-week coaching trajectory focused on troubleshooting and building self-efficacy in working with the system by the manufacturer 15 were well appreciated by participants. There are feasibility and upscaling considerations for switching care and start-up responsibilities from the manufacturer to regular care, but this requires diabetes teams to have at least one team member with in-depth technological know-how as well as to add psychological expertise to technology pathways. Recommendations about the minimal training and resources needed for health care providers to effectively support use of AID systems in regular care include a formal training with a demonstration system for hands-on learning, conversation guidance for system consultations, education packages accommodating different learning strategies, and monitoring of person-reported outcomes. 34 Support in the process of psychological adjustment to FCL should be structurally available as well, including expression of frustration and worries, making cost–benefit analyses and changing of nonhelpful thoughts. Ideally this represents a structural part of the training, supplemented by peer-to-peer learning opportunities (validation, practical advice) and—as part of shared decision-making by people with diabetes and their diabetes teams—counseling with a medical psychologist. Particular attention is warranted for discussing the more experiential aspects of FCL use, that is, related to wearability, in particular relating to intimacy.35,36
The key strength of this work includes the in-depth information about user experiences with FCL, both from ongoing users and people who discontinued the system. This provides nuanced information for people considering FCL as a treatment option as well as delineates the roles and competencies of health care providers in supporting people in this process. One limitation includes the potential participatory bias in the larger trial. There were seven people at the initial screening stage who decided not to participate and a further three who did not start after following the 2-day FCL training. 15 This left a sample of 79 people who—at least at the onset—appeared willing and able to work with this novel technology. Furthermore, the specific lens of experience and knowledge in the single coding approach needs to be acknowledged.
Conclusions
People with type 1 diabetes using bihormonal FCL experience improvements mostly in glycemic and often also in person-reported outcomes. Ongoing use appears closely related to navigation of the multifaceted challenges that come with working with diabetes technology. To facilitate FCL adoption and long-term use, expectation management and continued support from the manufacturer and trained diabetes care teams seem key.
Authors’ Contributions
G.N., T.J.P.J., M.K., L.B.V., and H.B. were involved in the substudy design. M.O., G.D.L., B.J.J.W.S., M.N.G., A.H.M., M.A.R.V., J.P.H.v.W., and A.C.v.B. were the local clinical investigators of the trial. Statistical data analysis was performed by M.K., supervised by G.N. Qualitative analysis was performed by G.N. and discussed in two stages with (1) T.J.P.J. and L.B.V., and (2) F.J.S., J.H.d.V., and A.v.B. G.N. wrote a first draft of the report, which was edited by T.J.P.J., M.K., L.B.V., F.J.S., H.B., J.H.d.V., and A.C.v.B. All the other authors critically reviewed the report. All the authors agreed to submit the report for publication.
Footnotes
Acknowledgments
Author Disclosure Statement
G.N.: Received no renumeration other than (co-)authorship for the present project. All other, unrelated industry research discounts (Dexcom; appr. EUR 5000 on the purchase of sensors) and speaker fees (Dexcom, Sanofi) were paid directly to the employer Radboudumc. Employer Diabeter is an independent clinic that was acquired by Medtronic. The research presented here was performed independently and there are no conflicts of interest. T.J.P.J., M.K., L.B.V., and H.B. are employees of Inreda Diabetic. G.D.L. has received consulting fees from AstraZeneca, Boehringer Ingelheim, and Lilly, and payment or honoraria from Sanofi and AstraZeneca; was a member of the scientific Advisory Board of the Dutch Diabetes Research Foundation; received support for attending meetings or travel from Sanofi; and has received institutional grants from Boehringer Ingelheim, Sanofi, and ZGT Research Fund. M.O. was a board member of De Diabeteskamer, the association of Dutch diabetes internists. All other authors declare no competing interests.
Funding Information
The study was funded by Inreda Diabetic B.V. (Goor, the Netherlands). The funder participated in study design, analysis and interpretation of data, and writing of the report. The decision to submit the report for publication was included in the research protocol. As external researcher, G.N. supervised the research activities of the funder and had the final say in the report.
Prior Presentation
A brief selection of the data was presented by G.N. at the ATTD Congress in March 2025. They have not been submitted to a preprint server. The primary quantitative outcomes of the FREE 1 trial have been presented at EASD 2023 and have been published in a peer-reviewed journal. 15