DIY Artificial Pancreas: Patients Aren't Waiting for Us

Jay H. Shubrook, DO; Rayhan A. Lal, MD


October 10, 2018

Jay H. Shubrook, DO: I'm Jay Shubrook, DO, a professor at Touro University, California, and a primary care diabetologist. We are going to continue our series, Everyday Diabetes: Practical Management for Primary Care. Today I'm delighted to have with me Rayhan Lal, MD, a senior fellow in adult and pediatric endocrinology at Stanford University. He also has a unique background in that he studied electrical engineering and computer science as an undergrad. You will find that this is relevant to our topic today, which is do-it-yourself (DIY) closed-loop insulin pump systems. This may not be familiar to you yet, but it is a topic that I think we are going to need to become familiar with. Dr Lal, thanks for joining us today. I know that the subject of diabetes is near and dear to your heart. Tell me about your diabetes story.

Rayhan A. Lal, MD: I've had diabetes for several decades now—I won't age myself by going into the specifics. I was always interested in technology from a young age, and as a result, I decided to pursue electrical engineering and computer science at University of California, Berkeley. During that training, unfortunately, my two younger sisters also developed type 1 diabetes. I decided I had to help all my brothers and sisters with diabetes out there. I switched gears and went to medical school. I'm not a super-decisive person. I could not decide whether I wanted to see kids or adults with type 1, so I decided to do med + peds (dual internal medicine and pediatrics residency) over at Los Angeles County Hospital, and I subsequently did a combined fellowship at Stanford.

Shubrook: That is great. You were quite passionate about this not only professionally but personally because it really touched your life.

Lal: Absolutely.

DIY Looping Systems

Shubrook: The treatment of type 1 diabetes has changed dramatically over time. There are so many different systems now. This new topic of a DIY system was discussed at both the American Diabetes Association 2018 meeting and the Stanford diabetes camp. It was all relatively new to me. What is this?

Lal: That is a great question. Whenever we talk about artificial pancreas systems, in general we're talking about three components. The first is a reliable insulin delivery system. The second is a way to monitor blood sugars on a continuous basis. And the third is something to control that insulin delivery—which used to be the person but now can be turned over to machines and algorithms, in some aspects. That third aspect [requires] a process controller that can decide: "For this stream of glucose data, how am I going to affect insulin delivery?"

We have had insulin delivery technology since the late 1970s. If you go back to those old papers, they talk about this need for continuous glucose data in order to design this closed-loop system. They were thinking about this way back when. It's only been in recent years that we've gotten reliable, accurate, continuous data from glucose monitors.

Shubrook: There are three parts: a delivery device, monitoring, and something to make these two work together in a continuous way. I've heard about insulin pumps. Why have I not heard about DIY, and how is it different?

Lal: Many commercial entities are trying to come up with the best algorithm, the best sensor, the best insulin pump. But a group of people in the diabetes community (#WeAreNotWaiting), would really love us to advance this technology at a more rapid pace than is currently happening in industry. The initial first insights into this came around 2011 or so. Someone who had diabetes realized that they could arbitrarily send commands to their insulin pump. When they presented this finding, they voiced the concern that it posed a significant security risk. As it turns out, you can send arbitrary commands to certain older-generation pumps, so it is indeed possible to take remote control over these pumps. If you have access to the sensor data and access to the pump, you can decide what algorithm you want to run in between the two of those. That is where this DIY comes in. Folks have come together and have figured out ways of reading data from their sensors and ways of controlling their pumps. They are writing code in order to help command the system and execute this artificial pancreas-type control that we talk about. This code is freely available.

Shubrook: This sounds like it's an adjustment to existing systems with outside code. Is this something that is common? Is it used a lot?

Lal: At last estimate, somewhere around 800 individuals were using these DIY closed-loop systems. But this is likely an underestimate. There is no mandatory reporting for using any of this stuff; it's open-source and you can download it freely. In addition [to the patients who are using it, many others are looking at it and are presumably asking], "Is this something I should consider?"

What Should Primary Care Clinicians Know About DIY Systems?

Shubrook: For primary care clinicians, this topic might not be very familiar. If a patient came in and asked me about this, what are two or three things you think a primary care physician should know?

Lal: First and foremost, this is not approved by the US Food and Drug Administration, so you have to discuss risks and benefits with the patient and family. One of the unique factors with these systems is that you can go online and look at the code that is controlling these systems. For insulin pump companies, you don't always have access to proprietary control algorithms used by manufacturers. However, if you can read the code, you can have a more a direct discussion on how the system is acting and make a more informed decision.

There are growing numbers of these DIY systems. The big ones currently in use in the United States are OpenAPS and Loop [A third system is Android APS.] OpenAPS is basically a black box at the side of the pump so all commands and actions that the user does are through the pump. Whereas with Loop, all interactions are with a phone, and the phone commands the pump and gets data from the sensor.

Shubrook: These sound like innovative and exciting systems. What I took away from your discussion is that if a patient approaches me as a primary care provider about one of these systems, I may not need to know all of these details, but I need to get them someplace where they can get the right support and knowledge to help them make an informed decision. It is important to know that there are people out there doing this, and we want to make sure they get their diabetes treated. Both are very important.

Lal: Absolutely. We have to remember that these are all individuals who are able to make informed decisions, and we have to be supportive just like we would be with anything else in medicine.

Shubrook: Thank you so much for coming on today. I really appreciate your insights and I really admire the work you are doing for diabetes. Thank you so much.

Lal: I appreciate it, Jay. Thank you again for having me on.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: