Do-It-Yourself Artificial Pancreas systems (DIY APS)

The way forward to closed-loop metabolic control in diabetes care?

The challenge

Diabetes represents a huge and multidimensional challenge. Despite the availability of numerous treatment options, many patents still fail to reach their treatment goals. Administration of  the right amount of insulin at the right time still poses a great challenge for the self-management of many people with diabetes.

Accordingly there is a huge need to implement new innovative products and services improving both the effectiveness of diabetes care and the quality of life for people with diabetes. Particularly user-centered products and sercvices co-created with stakeholders including people with diabetes may have a high potential to increase treatment adherence thereby reducing the enormous pressure on healthcare systems.

Artificial Pancreas Systems (APS)

In the last 10 years considerable progress has been made in the development of Artificial Pancreas systems (APS), also known as closed-loop or automated insulin delivery systems [1]. At the heart of APS an algorithm informed by the individual patients‘ continuous glucose monitoring (CGM) data regulates rates of a continuous subcutaneous insulin infusion (CSII) executed through a wearable insulin pump.

In case of a fully automated design the APS would take over full control over the patients‘ blood glucose homeostasis. When using socalled hybrid APS the user would be required to take action for adjusting insulin boluses at mealtime. Despite a long development history of APS and although large clinical trials demonstrated a safe and effective operation of AP systems in T1D [2] until today only the MiniMed 670G hybrid closed-loop system [3] received marketing authorisation for the treatment of people with T1D in the US and in Europe. Additional systems realising different design options such as the diabeloop DBLG1 system are on the way of achieving marketing authorisation [4].

Recognising end-user requirements: open source do-it-yourself (DYI) APS

Surveys that investigated user preferences for diabetes research [5, 6]  as well as the high commitment of a very enthusiastic open source do-it-yourself (DYI) APS innovation community (#WeAreNotWaiting) [7] document that users perceive a high need for the availability of technical solutions unburdening diabetes self-management including APS. In the DYI APS community technically versed people with diabetes connect existing insulin pumps and CGM sensor systems by using „homemade“ algorithms for the automation of basal insulin administration [8]. DYI APS users have reported impressive outcomes in terms of a decrease in HbA1c levels and an increase in time in glucose target range [9]. The open APS community is open to share their solutions and to collaborate with different stakeholders including diabetes associations and medical device companies [7].

In view of the lengthy development and regulatory approval processes and the high pressure exerted by end-users the question arises whether DYI APS solutions could pave the way for a wider and more rapid adoption of APS in diabetes management.

Potential issues with a wider DYI APS usage

A wider implementation of DYI APS systems has been critically reflected recently from different stakeholders‘ viewpoint [8, 9]:

1. Healthcare professionals run into a dilemma when encouraging patients to use DYI APS. On the one side they feel committed to provide maximum support tailored to their individual patients. On the other they might become exposed to liability issues – similar to the hypothetical situation of prescribing a drug not approved for the disease indication under treatment.

2. People with diabetes are quite heterogeneous along the dimensions of disease status, self-management capabilities and enthusiasm for technology, to name a few. While DYI APS could work well in highly engaged and tech-savvy users who are highly motivated in building and maintaining their DYI APS the risk- and cost-benefit balances of operating DYI APS in a vast majority of people with diabetes may be less favourable. Probably no medical device manufacturer would take liability for the consequences of failure of a component integrated within a DYI APS.

3. In case a DYI APS-experienced person would set up a DYI APS for someone else there could be a high risk of a limited joint understanding of what that would mean in terms of risks and liability in case of possible operator errors or malfunction of the DYI APS. Again this raises both ethical and legal issues.

4. DYI APS developers need to act with a great responsibility for others. While there is indication that DYI APS and the used algorithms can perform well under real-world conditions it seems obvious that the developer sets up and uses the DYI APS at his/her own risk. DYI APS developer should be aware of that the systems are not proven to be safe and effective in a broader population of people with diabetes.

5. Most probably payers would not reimburse costs for medical devices and suplies that would be used as part of a DYI APS. Further it seems questionable whether insurance companies would pay for emergencies and health issues related to the use of an unauthorized APS.

Merging the best of different cultures – for open and collaborative APS innovation communities     

APS may encounter a plethora of different needs and requirements but also hopes, fears, wishes, attitudes, expectations, and projections – within and across different stakeholder groups. Therefore an open and collaborative multi-stakeholder approach as currently implemented by the pan-European CLOSE EIT Health innovation consortium could help to accelerate a wider implementation of user-centric APS in diabetes care at attractive risk- and cost-benefit balances.

Such a co-creation approach integrating innovation, education and business creation merges the best of different innovation cultures. The engagement with patients, nurses, diabetologists, general practitioners, and payers will provide valuable insights about how they perceive diabetes and diabetes management and what are their expectations on current and future diabetes technologies including APS. Capturing provocative views from entrepreneurially‐minded people who take diabetes as an outstanding example for rewarding investments in digital care will help to test new business models related to the use of closed-loop control in diabetes care in a more comprehensive meaning.  

A the same time the experience and capabilities of powerful companies can help to promote the implementation of clinical trials large enough to provide sound evidence for a safe and effective use of APS in relevant patient populations and care environments. This will make APS more eligible for reimbursement. Of course meeting all regulatory requirements and getting marketing authorisation by regulatory bodies is of key for a wider and ethically justifiable distribution of APS for the benefit of all stakeholders.