Tidepool Loop, with Howard Look and Brandon Arbiter: 30 January 2023
Loop and Learn Speaker Series: Transcript
**READER WARNING: ALL TARGET RANGES DISCUSSED IN THIS CONVERSATION ARE IN mg/dL**
JM= Joanne Milo
LnL= Loop and Learn
HL= Howard Look
BA= Brandon Arbiter
JM: Hello, everyone, and welcome to Loop and Learn’s T1D Speaker Series with a special event. Today, we’re here with Tidepool Loop’s Howard Look, who’s co-founder, and CEO, and Brandon, Arbiter, VP of Business Development. What’s so unusual about this is this is a fairly historic approval by the FDA. Approving an open-source product done by a nonprofit. That’s the uniqueness of this. That’s why this is so exciting and new and it’ll take some adjustment time.
But you’re gonna hear some, some out fun and interesting things you’ll learn about their journey, and if you have questions, please put them in the chat. Please keep yourselves unmuted. And this yes, this will be recorded, and will be up on the YouTube Channel. We always do our disclaimer. So the DIY Loop app is a do it yourself closed loop algorithm. And now Tidepool Loop is an FDA approved version. Yay! So we always do this, and we say, if you are using a diy loop, you take full responsibility for building and running the system, and you do so at your own risk, and you consult with your health care professionals regarding your diabetes management and yes, this is hot news so… tomorrow you’ll be able to listen to podcasts. Stacey Sims recorded her podcast; it will be released tomorrow if you want to hear more and if you want to even hear even more, Scott Benner on Juicebox Podcast, will be releasing his… so you can just like overdose on Tidepool and still be…
HL: I’m not sure that’s my favorite metaphor.
JM: So let’s talk about Tidepool, because what is it? What do you mean? It’s a nonprofit. It’s a 501 c3, nonprofit committed to empowering the next generation of innovations and diabetes management, so last Tuesday they received FDA clearance for Tidepool Loop. It’s an automated insulin dosing app intended for the management of type one
diabetes in a person, 6 years of age and older. It is a patient-led project which is most unique, and it is the first application of its kind to receive clearance from the US FDA, and, it’s also the first to enable insulin delivery from a compatible Apple Watch.The algorithm technology and type for loop is intended for use with compatible, integrated, continuous glucose monitors and alternative controller enabled pumps. So that’s insulin pumps and each to automatically increase decrease and suspend delivery of basal insulin based on iCGM readings and predicted glucose values. You’re all pretty much familiar with this. But now it’s been approved by the US FDA. Howard Look, who is CEO and co-founder acknowledges the extraordinary support, and this is really important to the entire diabetes community, including the US FDA, and the contributions of the DIY community and participants in the Loop observational study conducted by the JAEB Center for Health Research. This closed loop has been studied more than any commercially available closed loop on the market. Don’t let anyone tell you that it’s new… nobody knows about it. Yes, they do. They have a lot of data, and that’s thanks to all of you who contributed your data.
Talk about Howard. He’s focused on developing and bringing to market Tidepool Loop, and he’s done it. And the key words here: open-source, not-for-profit system, and his background is very interesting: VP Of software, TIVO, Pixar Animation Studios, Amazon. I met him at Asante Pump company. He has a BS In engineering from Carnegie Mellon University, and his teenage daughter was diagnosed with type, one diabetes in 2011.
HL: That’s an old bio. She’s a 23 now. She’s all grown up.
JM: So just that understand how much work Howard has done in this community.
(Shows photo.) You in 2015. So yes, we’re going back away. White House. Champions of Change Award for Precision Medicine. He is working for this, and it’s wonderful. We’re very fortunate that he’s on our team on this one.
Brandon Arbiter met him way back in Nightscout days. VP of Business Development, he’s focused on making diabetes, technology smarter and more accessible to patients everywhere. He built a new generation app. (Shows photo.) Is this old Brandon about then? …He’s been an outspoken member of the patient community, advocating greater access to data and expanding the role of software and health management. Prior to joining Tidepool, he was VP at Fresh Direct, where he built the company’s data management analytics practice. He was diagnosed at age 27, with type one diabetes. With a BS in engineering management systems, industrial engineering operations research, economics, earth, and environmental engineering from Columbia University. I’m so very grateful that we were able to do this today, and I’m going to turn this over to Howard Look. Thank you, both of you, for being here tonight. Take it away.
HL: Thanks so much, Joanne. I really really appreciate everyone coming. It’s really good to see you all. I see some friendly faces that I haven’t seen in awhile, and some names of folks that I started interacting with, like Lane Desborough, back in, I think, 2012. The year after my daughter was diagnosed. So hi everybody. Hi Lane! So thank you all for coming, and it’s been a pretty wild week here at Tidepool.
First, let me just tell you a little bit about the company. As Joanne said, we are a 501 c3 nonprofit. We’re actually a completely remote distributed team and have been since 2017. We have about 30 employees right now. Got a few more folks coming in on board over the upcoming months, but still a relatively small team. We have 2 major product platforms, and we’ll mostly talk about Tidepool Loop today. But just to remind everybody Tidepool started as a company devoted to liberating diabetes data, and we did that with the Tidepool data management platform. We developed it through 2013 and in 2014 we grew through a series of organic partnerships with clinics like UCSF and Stanford. We eventually supported a study run by the JAEB Center for Health Research, which helped us grow even further and we’ve been giving away the Tidepool data management platform for free as part of our mission and a 501c3 nonprofit with the goal of helping people living with diabetes and their healthcare providers get access to all their data… to enable to see your CGM data, your pump data all in one place. A funny thing happened during the pandemic. We grew dramatically and I certainly wouldn’t have planned it that way. But it turns out when you can’t see your endocrinologist in person, you need a way to upload data from home, and we were a very, very popular choice. And over the course of that growth, we got lots of feedback from larger health systems and larger clinics saying, hey, there’s functionality that we’d love for you to add to your data management platform. We really love this. So we started adding features and functionality like a single sign-on and integration with things like active Directory and Ldab, for those that know what those are, and we added a population management dashboard and remote monitoring workflows and we launched all that new functionality, formally, last summer at ADA, and we call it Tidepool Plus. And this is really building upon all the work that we did from 2013 through 2022, adding this new functionality which people have really told us is helping them improve how they can see and manage their diabetes and in terms of health care providers it’s helping them manage large populations of diabetes patients so we’re super excited about that. Tidepool Plus is a paid tier of our Tidepool data management platform. And that’s really great, because we as a nonprofit, would love to become self-sustaining. We’ve been really fortunate to have some super generous donors over the years and organizations like the Helmsley Charitable Trust and JDRF and the Goldsmith Foundation support us, but over time we actually hope that the revenue from both Tidepool and what we’ll talk about in a minute, Tidepool Loop, will help us to become a self-sustaining nonprofit.
So the thing we’re here to talk about today is Tidepool Loop, which, as you all know… Loop is an automated insulin delivery system. Tidepool Loop is our version of Loop. We started with the DIY source code and brought it into Tidepool and put it through our regulatory quality system. And we’re very excited that, as of last Monday we received FDA clearance for Tidepool, and as Joanne said, that was very much because of the participation of this community and I’m sure many people here participated in the DIY Loop observational study. I want to give huge credit to the FDA. When we first started talking to the agency about Loop in 20- late 2017, 2018, one of the things they said to us is we love real-world data and what better way to get real-world data than to ask people using DIY Loop to participate in an observational study, and this is one of my favorite things to say about this study, and I think it’s really a testament to the incredible motivation of folks participating in the #WeAreNotWaiting movement and the DIY Loop movement to want to pay it forward. There were over 1,000 people enrolled in the observational study of DIY Loop, over 850 of which contributed from 6 months to over 12 months of data. So if you think about all of that participation in terms of days of use, studied in the observational study, the DIY Loop study had over 3 times the amount of data compared to at the time- the AID systems that were available, the pivotal studies for those AID… so we’re super proud of this. And frankly, so incredibly grateful to the do-it-yourself community, because our submission of Tidepool Loop was completely dependent on this clinical data. We simply could not have done it without the participants in that study. So for those of you here, that participated, thank you. If you know folks that participated in the study, please pass on our thanks. It was really just an incredible testament and desire to pay it forward.
All right, so for those that don’t know, let me spend just a minute talking about what it means to be an interoperable automated glycemic controller that the FDA calls an iAGC. So Tidepool Loop is an example of an iAGC, and the iAGC is one of 3 components that create this interoperability pathway that the FDA created starting back in 2018, when they designated the Dexcom G6 as an integrated CGM or an iCGM. Then, a year later, in 2019, they gave the Tandem T:Slim X2 the designation of alternate connector-enabled pump or ACE pump, and then they also gave both Basal IQ and Control IQ, and then, later, the Omnipod 5, the iAGC designation, which is the interoperability pathway. So Tidepool Loop is one of those. We are an iAGC, and what’s great about the pathway that the agency created is that it means we came up with a set of plans and processes and -regulatory nerd speak- they’re called SOPs or Standard Operating Procedures- and what that means is, we can add a new iCGM or a new ACE pump and we can keep adding new iCGMs and new ACE pumps without doing an additional regulatory submission. So we came up with these SOPs and iterated on them, and agreed to them with the agency, and the agency, basically with this clearance has said, “As long as you follow the plans and procedures that we agree to you can continue to add new iCGMs and new ACE pumps over time without a new submission.” We think that’s a pretty amazing, and groundbreaking thing, and that makes Tidepool Loop the first truly interoperable system. There are other iAGCs and other ACE pumps, but, as you know, they come from the same company, and they’re integrating with their own controller or their own pump. With Tidepool Loop it is built from the ground up, based on the DIY Loop source code as a truly interoperable system, so as new ACE pumps and new ICGMs come along that meet our requirements- as long as we follow those plans and processes, then we can add them to the system without any new filings. It also means that everything that you’re about to see when we talk about the details of Tidepool Loop. Now that we have clearance, we become what is known as a “predicate device”, and in FDA jargon that is what you need to do when you submit your 510k. The 510k is basically an application where you’re saying to the FDA, “We think our device is substantially equivalent to a prior device.” So our predicate device was Control IQ. We showed substantial equivalence to Control IQ. Now that we are cleared, future device makers can actually use Tidepool Loop as a predicate device, and that’s going to become important in just a few minutes when you see some of the features and functionality and things like where the target range is.
BA: So yeah, we’ll you know, we’re gonna talk about some features of Tidepool Loop. Since most of you are familiar with Looping, they won’t be surprises to you. We’ll talk about where things are with partnerships, and then we’re gonna open it up for interactive dialogue for Q&A. So that’ll be, I imagine, most of our time.
So first off, I’ll just give you all a little bit of personal context. Back in, I think, 2016. I was running Open APS and I walked into an Open APS meetup in San Francisco, and a friend of mine, Mark Wilson, said, “Dude, you’re not using Loop…” and that day I got set up with Loop, and a few months later I was controlling my diabetes from my Apple Watch. And yeah, it was amazing. And I think that to me personally, the DIY community has just raised the bar so much on what managing life with diabetes can be. And, as I think, towards Tidepool Loop, and this FDA submission and this FDA clearance, that’s the thing that excites me the most is just continuing that work to raise the bar to take what the DIY community- what the Loop community- has done, and turn it into what’s now, as Howard described a predicate device.
So feature number one that is now FDA-cleared and becomes a predicate that other future device companies can point to is Apple Watch control. Tidepool Loop in our FDA submission, we did include Apple Watch control as it is in the source from the DIY app, and that is part of our clearance, so we’re really excited to say that is, that’s in. And now future device makers working with the FDA can say, “They did it, and now we want to do it.”
The next feature that I’m particularly excited about is customizable target ranges. So in particular, I think there’s, you know, a bunch of questions about well, what does customization mean? How can configurable is it? Tidepool Loop allows a user in our FDA clearance to set their correction range as low as 87 and as high as 180. So again, this is a first of its kind in the diabetes community. This becomes a predicate, a precedent that other feature device companies can point to and say, “We can now have greater customization, configurability on our correction ranges.”
Other features that were, I think, innovated and pioneered in the DIY community that hadn’t been in any commercial system before… one is the pre-meal range- this is something I use all the time as a DIY Looper myself. Phenomenal innovative feature that lowers your correction target temporarily for the next hour or until you enter carbs, so that rather than doing a premeal bolus, which, as all of us living with diabetes know, you do that and you go to the restaurant. You order your food, and it takes an hour and a half to come out, and then you go low. The pre-meal range is such an interesting new take on that, that again was pioneered by the DIY community. And emojis, the pizza bolus. I can’t believe somebody thought to just use a pizza emoji for her pizza bolus. That’s great. This is, you know, to me the beginning of the end of all of those percentages, and extended bolus like, how much do you want up front? How much do you want later? This simple concept of like: ‘Hey, this isn’t about insulin timing. This is about carb timing’ and using the real estate of a smartphone, the interface of an iPhone, and the features that we’re so used to like these emojis to tell the system what I’m eating. So again in the Tidepool submission and now cleared.
And the last feature I want to talk about- and it’s funny- I just used this like 30 minutes ago, because I went and had dinner about an hour ago, forgot to bolus, came upstairs, turned on my Zoom, and was like, “Oh, I forgot to bolus.” So I entered the carbs and I just changed my carb time by 30 minutes into the past. This is not something that you can do on any commercial system. And what reminded me to do this? And again, this is DIY Loop that I’m using. As I started hearing my insulin pump give me insulin, and I realized my blood sugar must be going up, and I forgot to bolus. Well, now, do I? You know, do I enter all of the carbs that I ate? If I was on a conventional commercial system I wouldn’t have much of a choice. I could like lie to the system and tell it I ate fewer carbs than I actually did, but I might not know exactly what to do. In the case of Loop, enter the carbs you ate, and then put in the time that you ate it, and that’s it. So that also part of the Tidepool Loop submission and part of the Tidepool Loop clearance. This is now another precedent created by the community created by the Loop community and now cleared with Tidepool Loop by the FDA.
So what isn’t in there? There is one feature in particular that I just want to name honorable mention that I know is really. really popular in DIY, and that’s overrides. Overrides were not in DIY Loop at the time that the observational study started.
We at Tidepool are working on the integration of overrides now. The human factors work the clinical work behind it, and the design and development work. So that’s happening now. That was not including our FDA submission and is not in our FDA clearance. It is something that we’re all very excited about. And, as Howard mentioned, part of our FDA clearance is a process for how to submit changes to the FDA moving forward. And if we want to submit changes for overrides, that doesn’t mean we need to re-review everything that we’ve done already. But we have a process set up with the FDA to review those changes, so we are very excited to be working on that. Another feature that I see, actually, Alan is asking about in the chat, is autobolus. That was another feature that was not available in DIY Loop when the observational study happened and was not in our submission, but that is something that we are thinking about for upcoming releases of Tidepool Loop and upcoming submissions. Okay. I’m gonna pause there and I’m gonna hand it back over to Howard. Howard do you want to talk about devices and device integrations?
HL: Absolutely. Thanks, Brandon, and thanks for all the great questions everyone. I hope we can get back to all these great questions in the chat. So when we announced the Tidepool Loop project back in 2018, we announced that our initial development partner was Insulet makers of the Omnipod, and then, in 2019, we announced that both Medtronic and Dexcom had also been signed up as development partners. I’m very pleased to say that Dexcom is still very much an active development partner with Tidepool. Unfortunately, both Medtronic and Insulet have let us know that they will not be
launch partners for Tidepool Loop.
I can tell you that we are working with another ACE pump manufacturer. I, of course, can’t say who they are yet. It is a company that you’ve all heard of and a company that we believe everyone knows and loves, and we’re very excited that we will have an integration with them going forward, but unfortunately they’re not ready to talk about their plans publicly yet. So I’m sure many folks are disappointed. We are, too, of course, that neither Insulet nor Medtronic will be launch partners. However, we’re continuing to work as closely as we can with multiple other device partners. I know there are a lot of questions about who they will be. The thing that we’re excited about, as I said earlier, is now that we have clearance, we can add new devices without a new filing. And just to address the obvious question, “Well, if you can’t tell us who your launch partner is, can you tell us when the launch will be?” and unfortunately our launch partners aren’t ready to talk about that yet, either.
So the good news here is, we have clearance. We have FDA clearance for Tidepool Loop as an interoperable system, we will continue to work both to add new devices, and as Brandon said, this makes it much, much, much easier to continue to iterate on and improve Tidepool Loop, adding new features and functionality, and frankly catching up with where DIY Loop is with things like overrides and autobolusing. Alright. I think that’s all the content we wanted to cover formally. And Joanne, should we just start at the top and start plowing through questions?
JM: Yeah, I’m going to turn it over to Tina Hammer, and Cassidy Robinson also admins with Loop and Learn and we’re just gonna bounce questions to to you. So I’ll put them on the spot. Okay, Tina, are you ready to go?
LnL: Does Tidepool Loop include the autobolus option?
HL: So the version of Tidepool Loop that we submitted to the FDA is actually from mid-2020, and at that time the autobolus wasn’t a feature and was not part of the functionality. It was also not part of the clinical or the observational study of Tidepool Loop. So the version that got clearance does not include the autobolus feature. But as both Brandon and I were saying earlier, now that we have clearance, making new making changes and iterating upon what was cleared will be much more straightforward.
LnL: Can you guys expand a little bit on the potential timing of those changes? You’ve developed the SOP’s with the FDA. So what do the change timeframes look like? Is it quick, or are we expecting, like 6-month plus, until these changes can be pushed out to consumers?
HL: Yeah, that’s a really great question, Cassidy. And unfortunately, it doesn’t have a simple answer. The answer is, it depends. It depends on lots of things, straightforward changes, like changes to training or, you know, minor changes to the user experience, cybersecurity changes the agency wants us to push those out quickly and without even interacting with them. Changes that require additional clinical evidence will take more time, and so, and the reason is, and it’s a very good reason. The agency wants to know that the changes we’re pushing out are safe and effective, based on clinical evidence. And so a fundamental change to the insulin delivery algorithm or the glucose prediction algorithm would require clinical evidence. Overrides is going to require clinical evidence. The good news is now that we’ve done one study it is much easier to do future studies. We have now collaborated with the JAEB Center for Health Research as an incredible clinical research partner. There are so many people using Loop that we believe that it’ll be easy to recruit new people, and then, just ongoing we think it’ll be a much more straightforward to get studies done, but it will take some time. We will need evidence for substantive changes to demonstrate to the agency that it’s safe and effective.
LnL: Once you can demonstrate, and the agency approves, what’s the method for pushing the changes? Are we just app store updating? What does that look like for us as the user?
HL: That’s exactly right. Tidepool Loop will be in the app store, and we should probably
have talked about that. So one of the big benefits of Tidepool Loop is, you won’t have to build it yourself. It will be an app in the app store that you download just like any other app in the app store. It will require a prescription. So we’ve created a health care provider portal where your doctor, your Endocrinologist will log into the portal and create a prescription. You’ll then receive an activation code through email or text message, and that activation code is what unlocks the features and functionality of Tidepool Loop. It also will download your initial set of settings that, in collaboration with your health care provider, get created. But then you can go in and modify those settings. Because, as we all know, diabetes is a self-managed disease.
BA: Okay. Oh, Howard, here’s a great question. Is there anything that was in the submission that we had to change like Dexcom having to take out the silence all feature in the G7?
HL: Wow! That’s a great question. We did have to make changes based on agency feedback. I will say the agency was very reasonable when it came to alerts and alarms, and what we came with. And I hope everybody will be pleased. It will be possible to mute alarms for long periods of time, for example if you have a situation, whether it’s a you know, a wedding or a concert, or whatever you’ll be able to make sure alarms are really, really silent for a period of time. There were definitely things in the submission that needed to change based on agency feedback. They had all kinds of questions for us over our what’s called V and V: verification and validation activities. Cybersecurity we had lots of iteration on how we presented the clinical data to them. But in terms of the core functionality of Loop, really the things that Brandon showed are the fundamental changes. So there are constraints on the target range between 87 and 180. And we did do some work on alerts and alarms, like I just described.
LnL: Can you clarify? Because this was in the chat…the pre-meal setting? Are we sitting at 80, like we are in in 3, most of us now, or is that at 87 as well?
HL: I wish I can answer that off the top of my head. Do you know that Brandon?…
BA: So actually the lower limit is limited by the glucose suspend number, and the upper limit is limited by what you set your lower level of your correction. The lower limit on pre-meal range is 67, or your suspend threshold. Yeah. So if your suspended threshold is higher than 67, then your suspend threshold will be your lower limit, otherwise it’ll be 67.
LnL: Yeah, and that’s the same as Loop 3.
LnL: Can you talk a little bit about the training that that there will be for users, please?
HL: Yeah, absolutely. So we have developed in-app curriculum, which actually, we have published as part of our commitment to open source. So some of you may have seen that in app training and we’ve also developed, I’ll call it hardcopy, or online training, you know, more traditional training, both for people living with diabetes and health care providers. So over the course of the upcoming months we’ll start making all that available on we’re actually permitted to publish all that now that we’re cleared. And so we’ll now that we have clearance, we’ll start making all that available.
LnL: That is very exciting. I have, I have a follow up question about that, because that is probably one of our biggest challenges in Loop and Learn is training and onboarding and helping people understand how to drive this Ferrari. That’s a powerful device. Is your training- or are you able to say- whether it’s interactive? Can you take it by steps, and is there support behind that? For people that seem to be going into little ditches.
HL: Yeah, so right now, it takes 2 forms, what we submitted to the FDA was the training that is in-app that you go through during the setup process of Loop and that is modestly interactive. And it also takes the form of a more traditional user guide and health care provider guide. And those are just, you know, documents that are hundreds of pages similar to the types of user manuals that you’ve seen with other products. Now that we have those cleared, we can create other forms of training based on that same content. And I would expect us to do that as well.
LnL Will patients be able to access the health care provider documentation?
HL: As far as I know, there are no restrictions on us publishing the health care provider guide.
LnL: Okay, that’s great. I want to shift this to partners. I’m sure that’s kind of in the forefront so you haven’t talked much about CGM announced partners.
HL: So as I mentioned, Dexcom is still an active development partner with us, and we’re very excited to be working closely with them. I saw some questions float by in chat about Abbott and Eversense. We have no other CGM partners to announce at this time. I will say- I often joke our dance card is not full- and we will gladly add other compatible iCGMS and again, because of the standard operating procedures that we agreed upon with the agency. We can do that without a new submission as long as they meet the agency’s requirements for an iCGM and our requirements for an iCGM.
LnL: Is there anything our community can do to help you push CGM manufacturers?
HL: Just let them let them know what you’re interested in seeing
LnL: Okay, we can do that. Okay. Thank you. I know all this is in development and changing by the moment. You’re U.S.-based, is this is your first and primary focus right now?
HL: Yes, the clearance we just received is U.S. FDA clearance. We absolutely positively want to bring Loop to the whole world, and we will get there as fast as we possibly can. And same answer to Android. Obviously the DIY Loop system was based on iOS and written for an iPhone. So that was the obvious place for us to start. But our mission, our goal as a nonprofit is to make our technology as a broadly available in the most equitable way possible. And so obviously being available on Android is really important to us, too. But nothing to announce in terms of availability. I mean it’s gonna take more development partners and more funding partners to make that happen. But we’re absolutely committed to porting to Android as well.
LnL: Okay another question about settings: can multiple profiles be used, such as for different times of the month or different activities that you have regularly scheduled?
BA: In the version of Tidepool Loop that is cleared we don’t have, as I mentioned before, we don’t have overrides in there yet. The idea of multiple profiles, you know, beyond what you would see in DIY today- we have not diverged from that path. So we’re kind of following the same pathway to date as what you’ve seen in the DIY app.
LnL: Are you considering, including the Custom Type One patches that many of the loopers are using now? That includes the switcher patch where you can switch from temp basal to autobolus. Obviously, you’d have to get autobolus approval first, but it allows you to switch between those at a certain blood glucose level that you choose.
BA: As we investigate, as we design these updates into Tidepool Loop, for instance, overrides, auto bolus… we spend a lot of time interviewing folks who may have used similar features, or be using similar features, or NOT using similar features to understand how they’re using it. And so you know, DIY Loop has certainly been an inspiration for us, and it was the starting code base that that we started everything with, but we put our own design, and rigorous human factors process behind it, and I hope- and I believe- what comes out the other end is improvements. What’s amazing about the DIY community is how it’s been able to have so many different variants, different features, different ideas, and some of them really take off… some of them don’t. It’s an incredible natural, organic way to see what’s the cream of the crop, what floats to the top, what makes sense and then when we start to understand how these things are being used that’s all part of our design process is doing these user experience interviews. So there is certainly an opportunity, Ellen, to look through all of that and see how we take the best of all of these worlds and turn it into the user experience that we’re gonna put out there.