HCP perspectives

There have been a number of quantitative and qualitative studies examining HCPs’ comfort with supporting PWD who choose to use DIY AID [5,10,11]. Morrison et al surveyed 204 Canadian HCPs (33% registered dietitians [RDs], 32% registered nurses [RNs] 28% endocrinologists) in June 2021 [5]. Respondents expressed greater experience with commercial systems (median 6 to 24 users per practice), relative to DIY AID systems (1 to 5 users), in practices with a median of 100 to 500 individuals with T1D. Comfort levels with these technologies were commensurate with experience; 73% were comfortable supporting commercial systems, relative to just 22% with DIY AID systems. Specific barriers highlighted to DIY system use included lack of exposure, lack of high-quality published data, and medico-legal concerns. Both HCP and user education, in addition to clinical practice guidelines, were felt to be required in order to improve confidence in recommending AID use to individuals in their practice [5].

Another often cited barrier is the lack of regulatory approval for DIY AID, exacerbated by the ADA standards of care [2] and US Food and Drug Administration (FDA) warnings against the use of such technology [12]. However, as presented by the ATTD workshop on DIY AID in 2020, T1D is inherently a DIY condition, and all therapies, whether approved or not, require the user to make real-time decisions that come with inherent risk. Diabetes Canada supports autonomy, and self efficacy is a key skill that is further enhanced when individuals (or families) choose DIY AID. The peer mentoring inherent to the online community provides much more frequent touch points and support for users than traditional models of care where HCP visits occur every few months. The workshop concluded that there is a risk that PWD will stop attending care altogether if their HCPs do not support the use of DIY AID [12].

Some clinicians may feel threatened by the role reversal where the PWD is the expert in DIY AID, as described by Dinneen and McMorrow in “The expert patient will see u now, Doctor” [13]. Using DIY AID as a model, they describe how the dynamic needs to shift away from the physician expert towards a long-term therapeutic relationship focused on meeting the individual’s goals. Further, clinicians have biases regarding whom they think are best suited for AID, with educational level/cognitive ability as the most prominent factor [5]. Strikingly, in a separate study, HCP views changed when interviewed after 6 months' experience in supporting PWD using a commercial closed-loop system. Educated, technologically competent individuals tended to over-interact with the system. It was individuals who HCPs had assumed would struggle to understand and use the technology who seemed to benefit more by allowing the system to operate without interference. These observations led to the conclusion that all individuals should be given the chance to try AID [14]. Clinicians need to be conscious of their assumptions and biases that may result in inappropriate gatekeeping, thus preventing PWD from accessing helpful technologies. We suggest that the evidence and experience to date indicate that clinicians should discuss AID, including both commercial and DIY options, with all people living with T1D.

Individual perspectives

DIY AID systems were created by passionate and technologically adept individuals with lived experience out of a deep personal need for more effective and less burdensome solutions for safe glucose management. Since no commercial AID systems were available until 2016, PWD and caregivers used their skills to create DIY AID systems and made them publicly available for free; technology built for the diabetes community by the diabetes community. Current opinion is that DIY systems can add some degree of individualization of parameters compared to regulatory-approved systems. The nature of open-source insulin delivery systems is that the developers rapidly iterate adding new “branches” of code that allow for improvements in the system much faster than commercial AID systems, which require regulatory approval prior to making changes to algorithms and features.

DIY AID systems have become increasingly popular, with more than 9,000 users worldwide [10]. Although there are no precise estimates of how many DIY AID users live in Canada, there are 2,575 Canadians in the Looped Facebook group (an online community-driven support group comprising PWD, caregivers, HCPs, and researchers) [15]. The popularity of DIY AID in Canada may, in part, be due to the large number of PWD who do not have access to HCPs with expertise in T1D or private or public insurance coverage for commercial AID systems. The online DIY AID community has filled some of these gaps by acting as a supportive resource for PWD not only to troubleshoot the technology, but also in managing T1D in general—a recurrent theme identified in qualitative analyses [16].

According to respondents of a 2021 survey of 662 DIY AID users (99 from Canada), people choosing DIY AID do so for the following reasons: transparency (they want to see how the system works), interoperability (they want options with regards to the device, algorithm, and continuous glucose monitoring [CGM]), desire for open-source software, limited commercial options available or accessible, customization (including the ability to set lower blood glucose targets compared to commercial systems), user-led design features, and the ability for DIY systems to iterate faster than their commercial counterparts [17,18].

DIY AID users report feeling empowered by the ability to customize their systems and from the greater understanding of their diabetes and the physiology they have achieved by building their system [10]. Parents and caregivers share that DIY AID systems provide improved quality of life, sleep quality, and glycemic management, all of which were motivating factors for using DIY AID [19] (Textbox 1). Typically, DIY AID users have been highly educated, with access to disposable income, and have a high level of computer competency [17,18]. The major barriers to building DIY AID systems reported by parents and caregivers are their limited technical skills, lack of support from HCPs, and a perceived inability to maintain DIY AID systems [20]. Discontinuation of DIY AID systems is often because of difficulty fine-tuning settings, which could be mitigated by improved HCP education and support [21].

Glycemia

Improvements in both glycated hemoglobin (A1C) (0.3%–0.9%) and time in range (TIR) (0%–23%) are seen with initiation of DIY AID [4,23–41] in self-report and observational studies (Table 1), and a recently published RCT (Table 2) [42]. Ninety-seven PWD with previous experience using insulin pumps (n=48 children aged 7 to 15 years and n=49 adults; mean baseline A1C 7.5% and 7.6%, respectively), were randomly assigned to DIY AID using AndroidAPS or a sensor-augmented pump (SAP) therapy for 6 months. Significant improvement in TIR (74.5% vs 64.7% adults and 67.5% vs 57.4% children, p<0.001) relative to baseline was seen only in the AID group [41]. Compared with the SAP controls, DIY AID users spent an additional 3 hours 21 minutes per day in range, with 52% of users meeting international consensus targets of >70% TIR and <4% time below range (TBR), relative to 11% in the SAP group. DIY AID was particularly effective overnight (adults TIR 85.2%±12.7% overnight vs 70.9%±12.7% daytime, and children 76.8%±15.8% vs 64.3%±11.7%), while no differences in TIR were seen between day and night with SAP. The glycemic improvements with DIY AID became apparent within 2 weeks from randomization, and were maintained throughout the 24-week trial period. Importantly, unlike the majority of real-world user data to date, this RCT highlighted glycemic benefits with DIY AID introduction in users from varied ethnicities and socioeconomic backgrounds [42].

Quality of life

Improved quality of life has been observed in multiple studies that have used validated quantitative scores with the initiation of DIY AID. This includes improvements in diabetes-related distress (as measured by the Diabetes Distress Scale) [32,43], fear of hypoglycemia (Hypoglycemia Fear Score) [32,43], and diabetes impact and device satisfaction (DIDS) [4]. A qualitative study of Loop users (n=72) additionally highlighted the substantial impact of DIY systems in decreasing the mental or behavioural burden associated with diabetes management, with notable improvements reported, particularly in overnight glycemia and sleep quality [10].

Safety of DIY AID

Despite lower mean glucose levels with DIY AID use, this has not been accompanied by an increase in TBR [4,24,25,31–33,35,38,40,44]. In the 6-month RCT, no episodes of severe hypoglycemia or diabetic ketoacidosis occurred in AndroidAPS users [41]. Occurrence of severe hyperglycemia and ketosis (capillary glucose level >17 mmol/L; ketones >1.5 mmol/L and symptomatic) per 100 user-days in the 24-week trial period was 0.10 in the AID group and 0.07 in the control group [42]. Indeed, the lived diabetes experience of DIY AID builders and their immediate need for a safer way of insulin delivery was a strong motivating force contributing to DIY AID development.

Accessibility and costs

Access to insulin pumps and CGM devices varies substantially between Canadian provinces. Some provinces have broad public coverage for pumps, others have broad coverage for CGM, others cover both or neither, while some coverage may be restricted to children and young adults. The prospect of losing coverage when reaching a certain age results in anxiety and financial burden for families. Even with private insurance (e.g. through employers), reimbursement is usually subject to an annual cap and/or co-pays. The resulting out-of-pocket costs may mean that commercial AID systems are not affordable. In contrast, DIY AID can be assembled with out-of-warranty devices that are often less expensive to obtain or already in hand. PWD should be aware that ongoing costs of CGM and infusion set supplies are the same whether using a commercial or DIY system.

The fictional yet realistic cases in Textbox 2 illustrate the complex financial challenges facing PWD in different provinces, even for those with private benefit plans.

Pregnancy

Pregnancy in PWD is associated with an increased rate of both obstetrical and neonatal adverse outcomes. Stringent glycemic targets are imperative, but are often unmet, or can only be achieved with an increased rate of hypoglycemia [45]. The pre-set glycemic targets of currently approved commercial AID systems in Canada are not sufficiently stringent to reach targets recommended for pregnancy. In contrast, customized glucose targets appropriate for pregnancy can be selected by DIY AID users. To date, the literature is limited to small case series and case reports [46–51] that have demonstrated higher percentage TIR with lower percentage TBR compared to previous studies of CGM alone. Safe use of DIY AID at time of birth and postpartum can reduce management burden at a critical time compared to users' pregnancy experience with open-loop insulin pump with CGM [51]. The Expert Working Group’s clinical experience (n=4) also suggests that DIY AID can be used safely and effectively in pregnancy and may yield more time in the target range without severe hypoglycemia, which would be expected to result in fewer adverse pregnancy outcomes.

In-hospital

There have been no published studies on the use of DIY AID while PWD are admitted to hospital. Clinicians should follow local institutional practices regarding self-management in hospitals. There are published protocols that can be used to support safe self-management of diabetes in non-critically ill, psychologically stable PWD [52], and clinicians could apply protocols developed for continuous subcutaneous insulin infusion (CSII) users to individuals who are capable of self-management and using DIY AID.

Pediatrics

The advantages of DIY AID over SAP for glycemia in a 6-month RCT described earlier applied equally to pediatric individuals (aged 7 to 15 years). The DIY AID systems have shown improvements in glycemic management comparable to adult groups [3,10,53]. Additionally, findings reveal improved sleep quality, quality of life, physical health, and less worry about hypoglycemia at night and outside the family home, leading to increased autonomy for the child [53]. Parents and caregivers repeatedly report receiving practical and emotional support from the online community #WeAreNotWaiting [3,53].

As with adult groups, it is vital that pediatric individuals and their parents/caregivers are supported by their diabetes care teams. This should include addressing the need for support in schools and other care settings that is individualized and equitable.

Medical ethics

HCPs have a professional obligation to serve the public good by making the health and well-being of the individuals in their practice their first consideration. A second key aspect of the HCP–PWD relationship is that of respect for individual autonomy. These ethical principles require PWD to be provided with information about available therapeutic tools (treatment options) that can reasonably be expected to be beneficial for them, including those that might not be recommended (for instance because they are less effective or have more side effects, or because of costs and coverage).

PWD can only make fully informed decisions when they have been provided with information about all treatment options that can reasonably be expected to be beneficial to them. Withholding information undermines autonomy, a key value supported in the Canadian Medical Association Code of Ethics, as well as the provincial colleges of most regulated health professionals [54–56]. All members of the diabetes care team play a role in discussions around therapy choices, and will often be the first point of contact for troubleshooting. The same ethical principles apply to all HCPs in the unique scenario of supporting PWD who choose to use DIY therapy.

Medico-legal considerations

Autonomy and equitable access to health care are overriding legal rights of individuals in Canadian law and jurisprudence. There is a compelling legal argument that HCPs who support PWD who are using DIY AID are acting within the law in Canada. Conversely, refusal to provide support without making an effective referral to another HCP who can and will support the individual—and is accessible to the individual—may, in fact, be in violation of the law. Refusing to provide medical support for PWD who choose to use DIY AID is a form of gatekeeping that leads to the effective denial of equitable access to health care, creating a risk of significant bodily harm [57–59].

Another concern for HCPs is potential liability for negligence if they are providing medical support to PWD using DIY AID. A review of the legal test to determine negligence discloses that there may, in fact, be a greater liability from a refusal to provide medical support to people using DIY AID than to supporting use of DIY AID. A claim of negligence requires the plaintiff to prove on balance of probabilities that (1) the clinician had a duty of care, (2) the clinician breached the standard of care, (3) the plaintiff sustained damage, and (4) the plaintiff’s damage was directly caused by the defendant’s breach of the standard of care [58]. All 4 arms of the test must be met in order for a clinician to be found negligent, which is more likely to be the case if the PWD was denied access to care due to their own choice to use DIY AID.

For HCPs who prefer not to provide ongoing medical support for PWD who choose to use DIY AID, it is advisable to provide a referral to a clinician with experience. The use of virtual care should be encouraged to facilitate PWD from rural and remote communities to access clinicians with experience using commercial and DIY AID systems. Overall, a review of the legal literature, liability seems to be a greater risk if clinicians refuse to discuss, support, or refer on for DIY AID rather than the often cited concern that supporting individuals using these technologies opens clinicians up to liability claims.

The Canadian Agency for Drugs and Technology in Health (CADTH) defines ”off-label,” or expanded use, as any use beyond that which Health Canada has reviewed and authorized to be marketed in Canada. CADTH deems off-label prescribing to be safe, provided there is strong scientific evidence to support it, with a balance of risk and benefit considered [60]. While the CGM and insulin pumps used in DIY AID systems have Health Canada approval, the computer code that makes insulin adjustments does not. However, as discussed above, there is ample evidence of effectiveness and all discussions should review the benefits and risks in the context of the PWD’s goals. The Canadian Medical Protective Agency has recommendations for physicians when recommending the use of off-label drugs and devices [61] (Table 3).

To be clear, we recommend that PWD take responsibility for the technical hardware and computer coding. This should not be the responsibility of the clinician. If a clinic or third party does the initial build, the PWD may lack the necessary troubleshooting skills. HCPs should inform PWD that there is no helpline to call when there are technical failures with DIY AID and when there are questions about system features the online community is where answers will be found. Another concern raised is about the use of out-of-warranty devices in some DIY AID systems, but this is the reality assumed by many pump users whose warranty may expire before they are eligible for a new pump, depending on insurance funding cycles. As with all individuals using an insulin pump, clinicians should advise DIY AID users to have a backup plan for insulin delivery in the case of system failure. While we recognize that not all clinicians will be comfortable with AID, we feel that if AID is going to be offered then DIY AID should be mentioned alongside commercial options. If clinicians are comfortable supporting individuals on commercial systems, they should be comfortable supporting individuals on DIY systems as well. In the future, we hope to see the various DIY systems receive approval from regulatory bodies to open more options for interoperability and allow for more widespread use, as was done by Tidepool Loop, which received FDA clearance in January 2023 but has yet to announce its pump and CGM partners as of the printing of this position statement [62].

Effectiveness of DIY AID

1. DIY AID can be an effective approach to support individuals with T1D to increase TIR and reduce A1C without increasing hypoglycemia, all while reducing the burden of diabetes self-management and improving quality of life.
2. The benefits revealed by qualitative studies and lived experience should be carefully considered alongside glycemic benefits when discussing the risks and benefits of DIY AID. A net assessment of total risk should include the goals and concerns of the PWD and not overlook the risks inherent in all methods of insulin delivery.
3. No comparative studies have evaluated commercial versus DIY AID systems, such that no assumptions can be made about the superiority in either direction. Individual needs and preferences should direct choices.

Role of HCPs

4. HCPs have an obligation to discuss all available treatment options that have evidence of benefit with individuals living with T1D. DIY AID should be discussed alongside commercially approved options.
5. DIY AID may be more accessible to PWD in Canada than commercial systems due to costs, variation in coverage, and slow approval of commercial systems. Out-of-pocket costs should be considered even for those with private insurance.
6. HCPs should be aware of their biases and the potential for inappropriate gatekeeping (e.g. HCPs making assumptions that an individual or family is not equipped to manage DIY AID). Traditional “prerequisites” for pump use may not be relevant in DIY AID (e.g. precise carb counting, frequent finger sticks).
7. HCPs should make every effort to support PWD who choose to use DIY AID and seek to maintain the therapeutic alliance. HCPs should facilitate effective referrals if an individual’s needs exceed the expertise of the HCP.
8. It is recommended that PWD should build and install their own AID software. HCPs may support PWD in building their own systems, for example, by providing clinic space for peer-led mentorship programs (e.g. “build parties”) and encouraging engagement with peer-to-peer technical support networks.
9. HCPs should continue to provide ongoing support and education on core diabetes self-management skills, and safe use of insulin pump therapy and CGM.
10. HCPs should assist with optimizing pump settings and coaching on the behaviours linked to improving glycemic outcomes and meeting personal goals.

Role of the Community

11. Where possible, HCPs should connect interested PWD to trusted online, social media, and in-person resources. Peer support has been an effective tool both for people seeking to start using DIY AID and for providing help and support to existing users.
12. Additional adjustable parameters in newer branches of DIY algorithms may include terminology that does not build on existing CSII settings language, making it difficult for HCPs to support individuals in optimizing settings. An effort by the community of developers to consult with the diabetes education community would advance translation of DIY AID into clinical practice.