What are open source decision support tools for diabetes?
While approaches differ, the goal is to be able to more easily access the data that is stored and transmitted from a glucose monitor and to be able to share it with family, carers and healthcare professionals.
‘Open source’ means software for which the original source code – the computer code that tells a program what to do – is made freely available and may be redistributed and modified. Essentially, this means that people have developed the kit and computer programmes that make these systems work for their own use, and have then made the designs or programmes freely available for others to use and adapt.
Nightscout is a popular open source, DIY project that allows real time access to a CGM data via personal website, smartwatch viewers, or apps and widgets available for smartphones.
Another system under development is Tidepool, which we are supporting with a small exploratory pilot study in the USA.
What are the potential benefits?
Open source decision support tools have proved popular with some parents of children with type 1 diabetes as they enable them to access continuous blood glucose measurement data remotely through a variety of methods and then support the child, or their carer, to check blood glucose and make adjustments to insulin accordingly. This can provide peace of mind for all involved, and may help improve blood glucose control because the child can benefit from the parent’s knowledge, even when they are not with them.
Some athletes with type 1 have also found it useful as they are able to easily check their blood glucose level on a smart watch while exercising, rather than stopping to do a finger prick test or get out a CGM.
What are the potential risks?
It is important to remember that currently, many of the open source systems available are unregulated. This means that they have not gone through the stringent testing and assessment that all drugs and medical devices must go through before they are deemed safe and effective and approved for use by people with type 1 diabetes.
Any use of unregulated open source decision tools are at the risk of the user and we can neither recommend nor give advice on their use.
Tidepool is a registered company which complies with the US FDA regulatory requirements for medical device software, in contrast to other open source approaches which are community led across multiple regulatory jurisdictions. In addition, the Tidepool software does not make any recommendation for treatment, such as insulin dosing.
What are open source artificial pancreas systems?
Some people living with, or caring for someone living with type 1 diabetes, ‘hack’ into their type 1 technology to create their own ‘artificial pancreas’: a system that allows a continuous glucose monitor and an insulin pump to ‘talk’ to each other and automate delivery of insulin in response to changing glucose levels.
Those who do this often use open source algorithms (computer code) shared online. Examples include OpenAPS, Loop, and AndoridAPS.
What are the benefits?
Many in the community that take DIY approaches do so because they feel traditional methods of developing such systems are taking too long to reach people with type 1. They are able to create automated insulin delivery systems similar to those currently only available in the UK through clinical trials.
By creating these systems at home, they can personalise them, tweaking them to match their own needs.
Those who use open source algorithms often report improved glucose control.
What are the risks?
The open source algorithms that people use to create their own DIY artificial pancreas systems have not been tested in rigorous clinical trials and are not regulated. Traditional research methods require new treatments – whether drugs or technology – to be tested on large groups of people to be sure they are safe and effective, because the risk to life of a fault in the technology is high.
Creating DIY devices is complicated, and when dealing with insulin – which in high doses can be very dangerous – any mistakes or miscalculations could have serious outcomes.
Those who hack their technology can also risk losing any warranty for their devices, preventing them from getting replacements or repairs when required.
We cannot recommend anyone use open source algorithms to create DIY artificial pancreas systems, or offer advice to those who do.
What is our interest?
We are interested in the emerging and rapidly changing landscape of decision support tools and ‘artificial pancreas’ systems to aid people affected by type 1 in managing glucose levels. This approach to personalised medicine for people with type 1 shows potential and we are exploring where we can add value to this field of research.
We recently announced our Open Protocol Automated Insulin Delivery Systems Initiative. Our aim is to explore how we can support a second technology development pathway for automated insulin delivery systems – the artificial pancreas – in addition to the traditional, proprietary approach of established medical device manufacturers. The proprietary approach means that brands can dictate which devices the technology they manufacture can interact with, and prevents those who wish to from personalising their technology to meet their needs. Open protocol innovation should open the door to more personalisation of technology, allowing different devices to interact with each other.
With this new initiative, JDRF will work with the type 1 community, the DIY community, device manufacturers and regulators to support open protocol innovation with new technical, regulatory, and legal safeguards.
Our goal is to make this approach more accessible to a wider group of people with diabetes, allowing them another option to manage their blood glucose levels better, more safely, and in a way that works best for them.