User and Healthcare Professional Perspectives on Do-It-Yourself Artificial Pancreas Systems: A Need for Guidelines

Introduction

Technological advances, such as insulin pumps, continuous glucose monitors, and closed-loop artificial pancreas systems (APS), have aimed to assist in daily diabetes management by providing mechanisms to visualize and improve glycemic outcomes and reducing the associated burden for the person with diabetes (PWD). APS have consistently demonstrated in clinical trials efficacy and safety in improving time in target blood glucose range and reducing rate of time in hypoglycemia.^1-3 However, the development of such sophisticated technology and the associated rigorous clinical validation and regulatory process aimed at demonstrating safety and efficacy are lengthy and delay commercial availability. Frustrations about the lack of access to these novel devices have resulted in patient-driven, unregulated alternatives, specifically closed-loop systems created by technologically skilled PWD.

An online community known as the #WeAreNotWaiting movement was established in 2013 by a small group of PWD to advocate for the accelerated pace of innovation and to facilitate access to open-source APS technology through patient-built systems, also known as “do-it-yourself” artificial pancreas systems (DIY APS). ⁴ APS, DIY, or other commercial and investigational closed-loop systems enable an insulin pump and a continuous glucose monitor to communicate, and use potentially complex algorithms to automate part or all of insulin dosing. The number of PWD adopting DIY APS has been steadily growing in recent years with intriguing early outcome data for DIY APS that suggest possible improvements in self-reported hemoglobin A1c and glucose variability in this specific and unique subsample of adults and youth with type 1 diabetes.^5-9 Additionally, reduced diabetes management burden, improved quality of life, and perceived safety on DIY APS were reported by adult users and partners of PWD.^7,9 It is unclear, however, to what extent the demands of building and maintaining DIY APS may impact diabetes management because no data have been reported thus far. To date, the majority of studies on efficacy and safety of DIY APS have been uncontrolled, observational, or based on self- or parent-report data, with the exception of one published randomized clinical trial (RCT) that demonstrated the safety and feasibility of the open-source closed-loop AndroidAPS system compared to predictive low glucose management in a group of 22 youth participating in a 3-day winter skip camp. ¹⁰ Those results, consistent with the existing DIY APS literature, are promising yet had several limitations that impact the ability to draw definitive conclusions. Study samples thus far have been self-selecting, uncontrolled, and consisted of technologically savvy, highly motivated participants, so it is yet to be determined whether these benefits will be sustained over time or will generalize to the larger population of PWD. ¹¹

DIY APS are self-built using existing or outdated devices (eg, insulin pump and continuous glucose monitoring device) that have been modified by users to enable open communication. While outdated devices are already out of warranty, such adaptation of in-warranty devices invalidates said warranties and removes liability from device manufacturers and potentially leaves the user unprotected. No DIY APS are currently regulated or approved by governing bodies, such as the Food and Drug Administration (FDA). In 2019, the FDA issued a formal warning against the use of such systems after sufficient concerns about safety and liability persisted among stakeholders. ¹² These concerns have the potential to significantly impact routine clinical care because healthcare professionals (HCPs) are caught in an untenable position of wanting to provide best-practice diabetes care but being unable to openly support the use of DIY APS in their patients’ diabetes management. We believe this dilemma could be lessened if there were guidelines in place to assist both users and HCPs in better understanding the legal, moral, and ethical implications of collaboratively using DIY technology.

To explore these complex and controversial issues, we held a workshop at the annual Advanced Technologies and Treatments in Diabetes conference in February 2020 to obtain stakeholder perspectives on DIY APS in order to inform the development of guidelines for clinical practice. Approximately 60 stakeholders representing several countries worldwide participated in the workshop. Presentations were held by three key stakeholders: an HCP, a DIY AP user, and an ethicist. We report a summary of these perspectives and ensuing discussions, sharing considerations that are salient to the development of clinical care guidelines and standards of practice around DIY APS use.

User Perspectives

Many DIY APS users maintain that diabetes is inherently a DIY condition even with the various technologies that are available. In fact, HCPs have historically supported patient agency and autonomy in their diabetes care, particularly given that routine clinic visits occur only every few months. Users made the argument that the safety risks of DIY APS are no greater than the risks associated with human error in diabetes self-management (ie, calculating insulin dosing), regardless of technology use. Regarding preferences in routine clinical care, users reported re-evaluating the role of their HCP in their diabetes care once they began DIY APS use. Some users acknowledged experiencing changes in their perceptions of their HCPs based on HCP attitudes about DIY APS.

Despite a potential shift in the model of clinical care, the vast majority of users claimed to want continued medical support from HCPs regarding their diabetes management, including but not limited to obtaining supplies and prescriptions, consultations, and emotional support through active listening and an open, nonjudgmental attitude toward their decision-making. There was also emphasis on valuing HCP’s willingness to learn about DIY APS. Most users denied having any expectations that HCPs will provide technological support for the DIY APS, as they typically receive this type of support through peer mentoring.

HCP’s Perspectives

An important distinction was made that DIY APS only replaces the device, not the physician, in diabetes management. Specifically, there must be continued patient and physician engagement in diabetes management even with the use of APS technology in order for patients to achieve optimal glycemic control. In the absence of regulatory approval and sufficient efficacy data combined with use of off-label devices and the lack of formal guidance by institutions or licensing boards, nearly all HCPs indicated that they are unable to actively endorse or recommend DIY APS to their patients. Some HCPs continue to support patients by providing care as usual and additional DIY APS-specific resources (eg, websites, groups in the community) as part of a mutual commitment with their patients, while others described feeling cautious to do so given the lack of clarity around what they are able to communicate to patients about DIY APS and lack of official technical support and education.

Some HCPs reported having limited knowledge of and experience with DIY APS, which hinders their ability to confidently support their patients who use these systems. Regardless of personal beliefs and opinions of DIY AP use, the majority of HCPs cite liability and safety concerns associated with off-label use of component parts as the prominent barriers to supporting patients on DIY APS, even in countries with more liberal legislation.

Ethical Considerations

DIY APS use raises a number of ethical issues, including those that relate to device safety, liability, and potential impact on HCP–patient relationship. Regarding device safety, concerns focus on the use of out-of-warranty devices in DIY APS systems, which may increase the possibility of device malfunction, as well as associated risks of pump or sensor failure and/or inaccurate data. Such risks may lead to broader concerns about the quality of the data on the systems’ safety and efficacy. Additionally, there may be an increased risk with cybersecurity and, more specifically, unauthorized access to the system or one of its component parts, which could have significant and potentially fatal consequences.

These considerations taken together raise the question of who or which entity determines whether devices are “safe enough” to use in clinical practice. Although the risks of devices are measurable and quantifiable, whether or not these risks are worth the perceived benefits is a question each user will evaluate individually based on their own values, other personal factors, and without the guidance from the very agencies whose role it is to make such determination.

The component devices when used for DIY APS are considered off-label use, which has led to ethical and legal questions from stakeholders about liability should there be an adverse event. The lack of explicit guidelines from governing bodies on the clinical care of patients using DIY APS contributes to increasing concerns from HCPs about liability, particularly for HCPs who have limited knowledge about these systems. Without guidelines in place, some HCPs underscored the need for thorough documentation in the patient’s electronic medical record about DIY APS use paired with a statement of patient responsibility.

DIY APS have the potential to shift the clinical care paradigm, which inherently introduces some unique ethical and moral considerations in the patient–physician relationship. There is currently no precedent of responsibility in DIY APS use. Without regulatory approval or clinical guidelines in place, there may be uncertainty, discomfort, or lack of clarity about the degree to which HCPs are able to engage their patients in a dialog about DIY APS. Additionally, the patient–physician relationship is inherently composed of trust and patient expectations that HCPs will inform them of all treatment options, discuss associated risks and benefits, and help tailor interventions to the patient’s individual needs. Lack of discussion about DIY APS by either the HCP or the patient has the potential to undermine trust, which is critical in collaborative, patient-centered care.

Conclusion and Future Directions

DIY APS as a patient-led initiative are challenging the more conventional top–down model of health care. With any paradigm shift, it is essential to take into consideration the ethical, legal, and moral implications, particularly in the delivery of clinical care. User and HCP perspectives appear to align on the principle that DIY APS only replaces the technology, not the provider, in diabetes management, as the majority of users want to receive continued support from their HCPs and HCPs want to continue to provide high-quality patient-centered care irrespective of diabetes technology.

There continues to be concerns, which largely stem from a lack of clarity or sufficient evidence, around several critical issues related to DIY APS use that hinder more broad acceptance and/or support of these systems in clinical practice. These questions and concerns need to be carefully considered and addressed before formal clinical guidelines related to DIY APS can be developed and implemented. Ongoing research includes the OPEN Project, which may address some of these outstanding questions. ¹³

Based on the workshop presentations and discussions, HCPs, including those who positively regard DIY APS, continue to have many questions related to patient safety, provider liability, official technical support, and how these systems work. A combination of liability and patient safety, particularly the possibility of an adverse event, appears to be the prominent barrier to support/recommendation of DIY APS from stakeholders, which has historically led to users assuming full liability for their own systems with HCPs and peer mentors. ¹⁴ Additionally, despite the potential glycemic and quality-of-life benefits of DIY APS, studies to date are limited by highly selective samples, self-report data, and mostly observational reports and uncontrolled studies with the exception of one RCT.

Widespread engagement by HCPs with DIY technology will require some combination of professional guidelines for use of DIY, a less negative attitude toward this approach by regulatory authorities, and evidence that liability for malfunctioning equipment lies with the user rather than the HCP who is assisting them. In order to advance the development of clinical care guidelines and to garner formal support of DIY APS from stakeholders, including governing bodies, more robust RCTs examining system efficacy and safety, particularly in the home setting, are critically needed. Education and training programs to address questions about the various DIY APS and their functionality may be helpful to inform HCP decision-making in patient care and/or whether they elect to support DIY APS use for their patients. Similarly, on the premise that “one loop does not fit all,” it will be advantageous to systematically identify patient factors that are indicative of sustained or successful use of DIY APS. People with diabetes should have the choice to select the devices that they want to use and the closed-loop algorithm that will best suit their diabetes management needs. We believe that this is the time for stakeholders, including patients, HCPs, payers, and regulatory authorities, to come together to define the use of safe DIY technology.