October 27-28, 2016; San Francisco, CA; Highlights – Draft

Executive Highlights

Greetings from our home city of San Francisco, where DiabetesMine’s intimate D-Data Exchange and Innovation Summit drew an impressive, technophile crowd on Thursday and Friday. We had a nostalgic “wow” moment as presenters shared the latest in smartphone connectivity, apps, and open data: it was only 4.5 years ago that Sanofi-AgaMatrix’s iBGStar launched in the US. And if you go back to that piece, “Bluetooth” is not mentioned once, since it was not a standard at the time. Fast forward to 2016, and we have several smartphone-connected BGMs and CGMs launched globally, a hybrid closed loop system approved in the US, and companies increasingly opening up data and building apps to make it more actionable. We are so glad Sanofi made that investment and although the timing was unfortunate with the iPhone charger, it’s key to remember the major investments that have been made. At the DiabetesMine Innovation Summit, the optimism is palpable at meetings like this, even though there is a ways to go before digital products become so compelling that the average patient and clinician cannot live without them.

See our top highlights from both days below!

Top 10 Highlights

1. FDA’s Dr. Courtney Lias shared that her team had monthly meetings with Medtronic for 2.5 years leading up to the 670G submission (speeding up availability by an estimated three years!), expressed optimism that the final BGM guidances will improve product quality, and jokingly unveiled the hashtag #WeAreStillWaiting for DIY groups like OpenAPS to reach out.

2. Dexcom announced the beta launch of developer.dexcom.com, a major win for accelerating an open diabetes data ecosystem. Developers can register on the site, access Dexcom’s APIs (retrospective data initially), create prototype apps, play with simulated data, learn how to become a Dexcom data partner, and even submit an app for commercial approval.

3. Talented software developer Nate Racklyeft keynoted Thursday, sharing lessons learned building Loop: an iOS app that substantially reduces the hardware form factor of the DIY OpenAPS. Loop provides a user interface for closed loop on the iPhone and requires just one small relay device (a “RileyLink”), a Medtronic pump, and Dexcom CGM. Very impressive.

4. We saw live demos of several in-development or launched diabetes apps: Medtronic/IBM Watson’s Sugar.IQ (similar to last month’s demo); Nutrino (food logging + CGM data); Suggestic (chat interface for food advice), Glucose Advisors with Engine 1 (T1D exercise advice), DarioHealth (smartphone BGM), and InSpark’s Vigilant (BGM pattern recognition). The field is gradually moving beyond gathering data to doing something valuable with it. It’s still early days, but we’re optimistic.

5. An outstanding design thinking workshop shared three key takeaways for diabetes devices: (i) it’s hard to pinpoint design problems in something you’re highly familiar with; (ii) the most mundane, everyday things need design improvements, but often require watching someone actually use the product; and (iii) delightful, fun experiences can make a big difference.

6. NightScout Foundation’s Wes Nordgren called on pump manufacturers to grant DIY developers communication access to control pumps remotely (e.g., via a phone). We wonder if it is possible for companies and FDA to allow an optional “experimental mode” – do-it-yourself users (DIYers) could toggle this on and assume all risks of communicating with a cleared device.

7. mHealth app consultancy Research2Guidance shared only ~1/3 of diabetes apps are “actively managed,” while the rest are clutter leftover from the 2008-2013 “there’s-an-app-for-that” boom. The unfortunate result, however, is a vicious cycle where previously popular, now-defunct apps stay on top of the list. Ugh!

8. Kaiser’s Dr. David Sobel gave an outstanding talk on the psychology of diabetes, arguing that “most of the time, people are not unmotivated; they are overwhelmed.” He urged the audience to “prescribe success” – understand what truly motivates someone, make it easier for them to do what they already want to do, and identify/celebrate what they are doing right. The latter is so often forgotten in diabetes! Dr. Sobel’s talk was a real standout and we look so forward to hearing him more in the future.

9. Diabetes veteran Ansley Dalbo shared lessons learned in developing Diabetes What to Know, a YouTube video channel and educational website for people with type 2 diabetes.

10. DiabetesMine’s Amy Tenderich announced the blog’s 2016 usability innovation awards: Dexcom’s G5 (people’s choice), Abbott’s FreeStyle Libre (best newcomer), TuDiabetes.org (best quality of life innovation), and OpenNotes for viewing a doctor’s notes from an appointment (honorable mention).

Top 10 Highlights

1. In an FDA innovation update, Dr. Courtney Lias revealed that her team had monthly meetings with Medtronic for 2.5 years leading up to the MiniMed 670G submission (speeding up availability by an estimated three years!), detailed measures the Agency is taking to improve test strip surveillance and quality (e.g., BGM guidances finalized two weeks ago), and jokingly unveiled the hashtag #WeAreStillWaiting (for DIYers like OpenAPS to reach out to FDA). We knew FDA and Medtronic had collaborated strongly on the 670G, but we weren’t aware it was that extensive. In the monthly meetings, FDA gained an understanding of the 670G well ahead of submission, and also gave the company advice and pushed it to move faster. The result was a much faster review process (just over three months) than anyone expected (see our coverage from last month’s approval). Dr. Lias wished that the FDA “could do this for every novel product,” and we’ll be interested to see how quickly other automated systems and novel sensors move through the agency. She noted that the 670G submission didn’t include the three-month pivotal alone, but a host of other studies: missed meal bolus, missed CGM transmission, exercise, PLGM alone, etc. (This was also a reminder of how much other companies may need to do on top of a pre-market pivotal.) Regarding the just-finalized BGM guidances, Dr. Lias called the FDA review of test strip lot release criteria a “huge deal.” The strategy is part of an Agency move to increase the quality of cleared meters, and companies will need to answer a series of questions: How many strips will be tested? How will they be tested? What are the criteria by which a company will decide if the lot is safe to release? Dr. Lias also turned to the OpenAPS community – which had numerous leaders in the audience – noting “#WeAreStillWaiting” for DIYers to reach out to the Agency. She followed this request with “Talk to us! We will only wait so long before we start reaching out in different ways.”  We didn’t see this as a threat – Dana Lewis and Scott Leibrand were sitting next to Dr. Lias for much of the day – but it was a reminder that the FDA wants to talk and come to possible “win-win” solutions for the Agency, patients, and the OpenAPS community.

• She concluded with three refreshing “next steps” for the agency: (i) continue to help device sponsors make new and innovative technologies available; (ii) work with the community towards device interoperability; and (iii) continue seeking input from patients on their technology needs. No argument there! Dr. Lias also asked attendees for help: “We want to hear from everyone. What features and functions do you want in future automated insulin delivery systems?” More from this positive and exciting talk below!
• Dr. Lias sounded truly excited about the products currently under FDA review, though could not say anything more about them. We’ll be interested to see how quickly the consumer version of FreeStyle Libre is approved following the 3Q16 submission.

2. Dexcom’s SVP of Data Annika Jimenez announced the beta launch of developer.dexcom.com, a major win for accelerating an open diabetes data ecosystem. Developers can register on the site, access Dexcom’s APIs (retrospective data initially), create and manage pre-commercial (prototype) apps, play with simulated (sandbox) data, learn how to become a Dexcom data partner, and even submit an app for commercial approval. Yes! The site has launched for sign-ups, though Ms. Jimenez said it will be fully available in “early 2017.” Dexcom partners will have access to retrospective data APIs initially (three hours old), but the hope is to reduce the delay over time in negotiation with the FDA (perhaps to two hours or even one hour). The picture (see below) showed Dexcom’s Clarity platform sitting next to third party apps, all built on top of a data platform and APIs. Developers will have access to quite a bit of information to develop novel retrospective data apps, including glucose data, statistics, device info, and calibrations (right side of picture below). Like Dexcom Clarity, this platform is a class I medical device (retrospective CGM data), and Dexcom will support data partners and share best practices through the developer portal. Partners will not be able to leverage Dexcom’s quality systems, as we would expect. Ms. Jimenez has significant Big Data experience from Silicon Valley (she spent years at Yahoo!), and noted that these are “OAuth2” and “restful APIs” – exactly what open data proponents like Tidepool’s Howard Look have desired for years. Dexcom has always been open to partnering on the data front, and this move to expose its APIs sets a new standard for the field. Ms. Jimenez shared several examples like Amazon, Apple, Google, GE, and Uber, noting that “a data platform simplifies our experience and raises the bar on what’s possible.” Dexcom clearly understands that it cannot and should not build all the useful CGM data apps, and we’re elated to see it enabling an ecosystem that could propel others. The recent hiring of patient innovators Ben West and Chris Hannemann from the OpenAPS developer community also speaks volumes about the company’s vision.

• One developer we talked to shared two unknowns with us: (i) What restrictions, if any, will be placed on apps that access this data? Will users be able to share the data to additional third party apps, as they can today with HealthKit? (ii) How does Dexcom plan to address the demand for real-time data?

3. Talented software developer Nate Racklyeft keynoted the D-Data Exchange, sharing lessons learned building Loop: an iOS app that substantially reduces the hardware form factor of the DIY OpenAPS automated insulin delivery system – see the pictures below and read the history of its development here. Loop provides a user interface for closed loop on the iPhone and requires just one small communication device (a “RileyLink”) that turns Medtronic pump RF into Bluetooth. Loop reads CGM data from Dexcom devices in real-time via a reverse-engineered Bluetooth protocol. The also allows remote Medtronic pump bolusing directly from the iPhone app (secured by Apple’s Touch ID fingerprint technology!), and the user interface is pretty compelling: 3+ hours of glucose history and predicted glucose, active insulin, bolus/basal insulin delivery, “active carbs”, and key status information on one screen (see below). Loop also has an Apple Watch app that enables bolusing from the wrist, and users can tell the algorithm whether they are “pizza,” “taco,” or “candy” carbs – very cool. Just like OpenAPS, Loop automatically modulates basal insulin delivery, and users set a max hourly limit for safety. Adam got the app set up on his phone, but definitely needed serious help from the programmer crowd – this is not plug-and-play by any stretch of the imagination (e.g., it’s not on the App Store, but requires compiling the source code using Xcode to install on a phone; technical experience definitely needed). Loop requires older Medtronic pumps, and communicating with the RileyLink relay devices requires entering the pump ID into the app. At least six people in attendance were already using the app, and it reminded us that burden threshold will be critical for this field – while the original OpenAPS configuration was too clunky even for some type A patients, lower carrying bulk with Loop has now reduced the burden substantially. Nate recently joined Apple’s Health team as a software engineer, and we’re glad to see some diabetes patient experience will make it into the tech juggernaut (he emphasized in his blog post that he has moved on from developing Loop for personal reasons). We thought this was an outstanding talk, and more importantly, a testament to how much the DIY community has accomplished and driven forward. We hope much of the learning makes it into commercial products!

• Nate’s talk also shared five lessons from Loop for anyone developing apps or devices: be interoperable, be connected, be transparent, be discrete, and be accessible. We loved his point that apps should optimize the user experience for spending less time in the app – the opposite of what most apps try to do. “I open Loop less than five minutes total on some days. Getting more eyeballs into our apps is not what users want. Strive to reduce engagement with users’ disease, even if that means less time in your app.” As algorithms get more complex, he also called on industry to “earn users’ trust by allowing them to explore [a] system’s complexity.” He pointed to the 2015 BMC Medicine article about poorly performing, unapproved smartphone bolus calculators – only 30% of apps even documented the formula they were using. We wholeheartedly agree that closed-loop algorithms should be very upfront about how they work, enabling users to understand how the system works and performs in different use cases (e.g., “What will the system do if I...”). Nate also highlighted the discretion enabled by Apple Watch and phone bolusing with Loop, a critical advantage for teens with diabetes. He praised the Apple HealthKit platform for enabling interoperability and data sharing between apps – for example, aside from Loop, he uses One Drop as a longer-term view of his diabetes trends. Last, Nate highlighted Apple’s strong accessibility features, allowing apps to be used by those with impaired vision or blindness. Unfortunately, his audio demonstration revealed that some apps design for this poorly or not at all.

4. We saw live demos of several in-development or launched diabetes apps: Medtronic/IBM Watson’s Sugar.IQ (no new updates since the demo last month); Nutrino (food logging + CGM data from Medtronic and Dexcom); Suggestic (chat interface for personalized food advice), Glucose Advisors with Engine 1 (exercise advice), DarioHealth (smartphone BGM), and InSpark’s Vigilant (pattern recognition for BGM users). The field is gradually moving beyond step one (gathering the data passively) to step two: doing something valuable with it. It’s still early days, but we’re glad to see so much interest in providing better real-time and retrospective guidance, particularly around CGM data and food. The big questions for every app are burden of entry, added value, sustained engagement over time, and business model. We continue to be impressed with Medtronic/IBM Watson’s Sugar.IQ, and will be interested to see how patients like it once the broader launch commences later this year. See the detailed table below for links, descriptions, and what we found notable.

5. An absolutely outstanding design thinking workshop led by Sara Krugman (Tidepool) and Caroline Arvidsson (Refugee Text) shared three key takeaways for diabetes devices: (i) it’s hard to pinpoint design problems in something you’re highly familiar with; (ii) the most mundane, everyday things need design improvements, but often require watching someone actually use the product; and (iii) delightful experiences count a lot. The workshop took attendees through two interesting hands-on, guided exercises: (i) watching a partner open and put on a Band-aid; and (ii) analyzing different user interfaces. Both had guided question sheets to help with specific design observations. The first exercise revealed how truly painful the design of Band-Aid packaging is, highlighting many obvious opportunities for improvement: easier to open, larger font, see-through packaging (to know whether it’s Barbie or Minions), cotton swab incorporated, etc. The volume of ideas was impressive, all from a room of people simply watching each other do an obvious, everyday thing. One person in our group had a miserable experience opening the Band-Aid (“This is not designed for me!!”), and his only highlight was, “Yay! I got Minions!” It was a great reminder that fun and personality goes a long way, and we hope diabetes devices start to add a little more delight into their design. The guided user interface workshop was also fascinating – our group picked apart unfamiliar user interfaces with serious precision (e.g., a complicated computer game display), but was immediately stuck when trying to improve the very familiar iPhone home screen. We loved the reminder that familiarity matters in design – it can be hard to zoom out and be critical of something you’re intimately working on. Adapting a designer’s perspective helped many attendees to step out of their comfort zone and identify weaknesses (and successes) in common interfaces, which we hope will trickle into products. User experience is easy to underestimate, and we hope design thinking continues to make its way into diabetes in a larger way – it will be needed to massively expand both technology and drug adoption.

• In the introduction, Ms. Krugman and Ms. Arvidsson emphasized that personal health is “one of the most intimate things about being human,” so healthcare designers must truly empathize and understand the user experience. They also posited that the context within which healthcare devices and products exist is critical – Ms. Krugman related a story about a psoriasis cream that proved ‘97% effective’ in clinical trials, yet customer retention was surprisingly low. The company sent a user research team to an elderly woman’s house, where they discovered that she couldn’t push the cream out of the metal tube due to crippling arthritis. Upon further research, they learned that 20% of people with psoriasis also have arthritis (!), and switched to plastic tubing.
• This short powerpoint, put together by Ms. Krugman and Ms. Arvidsson, contains five effective exercises that facilitate people-centered design. Each uses a different strategy to put the designer in the end users' shoes. Highly recommended!

6. NightScout Foundation’s Wes Nordgren called on pump manufacturers to grant DIY developers communication access to control pumps remotely (via a phone, for example). We wonder if it is possible for companies and FDA to allow an optional “experimental mode” – DIYers could toggle this on and assume all risks of communicating with an FDA cleared device. What’s happening now is a lot of wasted DIY time trying to hack into pumps, and of course, the additional patient burden of communication relay devices (like the RileyLink). Mr. John Costik, a founding member of the CGM in the Cloud group, told us that the first pump company to open up communication will “sell a lot of pumps and drive innovation at an even quicker pace than we’ve seen so far.” This presents a liability challenge for industry, but one that we hope is solvable with the proper disclaimers and working with the FDA and patient innovators – e.g., “By enabling experimental mode, you void your pump’s warranty and assume all liability risks.” Alternatively, we wonder if select innovator groups like OpenAPS could get “developer” access to pump communication protocols under certain terms and conditions. FDA’s longer-term vision of component AID systems (enabling users to swap different pumps, algorithms, and CGMs in and out of systems) would require extensive device interoperability anyways, and we assume companies could push this faster by working with the DIY community, FDA, and adopting recognized communication standards like Bluetooth. In reference to the preceding evening’s Hackathon, Mr. Nordgren formulated yet another hashtag: #Don’tMakeUsDoThis, a frustrated summary of the current wasted effort trying to decode pump communication protocols.

7. Ms. Audrone Skardziute of mHealth app consultancy firm Research2Guidance shared an astounding statistic about the app store: Only ~1/3 of diabetes apps on the store are actively managed, while the rest are clutter leftover from the 2008-2013 “there’s-an-app-for-that” boom. Most of these so called “legacy products” are the result of one-off attempts that have been abandoned – countless companies came out of the woodwork thinking they could monetize apps through advertising and in-app purchases, but this model did not turn out to be viable. A prime example is GlucoBuddy, which, despite not having been updated since 2012, is still consistently ranked highly on the app store. It manages to stay at the top due to a continuous cycle – it was launched when there was little competition and was highly ranked (the app store description says “Ranked #1 Diabetes iPhone Application by Manny Hernandez, Founder of TuDiabetes.com”). The high ranking continues to drive high download rates, though the app is not actively managed. This is disappointing, but luckily, the share of global downloads lines up with the percentage of actively managed apps (see picture below).  

• Between 2014 and 2016, the global share of diabetes app monthly active users grew from 2.2% to 3%. It was unclear exactly what the numerator and denominator were, but we interpret it as “monthly active users of diabetes apps” divided by “number of people with diabetes with smartphones.” The implication is there is still tremendous runway for apps to make an impact in diabetes! Ms. Skardziute commented that the US market is the “sweet spot,” boasting the second largest market of people with diabetes who own a smartphone: 13.1 million, only behind China. In addition, the US comprises one big unilateral market, so developers don’t have to tailor to specific markets, as they would in other zones like Europe. On the flip side, the US market is mature, so competition provides a high barrier to entry.
• Ms. Skardziute outlined four strategies of the current app market leaders, advising that developers “focus on one thing and do it very well.” The four strategies are: (i) Enhance and sell hardware – if selecting this vein, Ms. Skardziute advises keeping the app to a minimum, discrete device data display/transmission tool and managing the API strategy to connect to services and companies that build tools on top of device data; (ii) Provide patient services – in this case, the developer should be highly engaged with patients and prepared to incorporate their feedback; (iii) Gather data for clinical and R&D purposes, building tools to help make sense of the data and helping companies make better devices and software; and (iv) A health and fitness offering, focused on building dashboards and tools for companies.

8. Kaiser’s Dr. David Sobel provided a truly outstanding talk on the psychology of diabetes, arguing that “most of the time, people are not unmotivated; they are overwhelmed.” He urged the audience to look behind “boring” health education and think about “prescribing success” – understanding what truly motivates people, making it easier for people to do what they already want to do, and identifying and celebrating what they are doing right. To illustrate his point, he listed the best way to prescribe failure: undermine happiness and health; make a “general” prescription; focus on the long-term; and deprive people of something. Where chronic conditions need to go, he said, is towards the personal, specific, easy (70-80% confident), rapid and immediate, pleasurable, environment focused, and celebrating success. Hear, hear! He told three memorable stories to make the case: (i) a man who didn’t change until the benefits of caring for his diabetes were connected to his greatest love (Trout fishing); (ii) a woman who couldn’t remember to take her medication until she suggested her own solution (a sticky note reminder on the Tide laundry detergent); and (iii) an unexpected compliment from a mechanic that he had done “a great job of protecting his Tires (“I love to be acknowledged for something I’m doing well”). He summarized the motivation examples – Trout, Tide, Tires – beautifully:

• Trout: Discover their passions. What do you really enjoy?
• Tide: Invite their solutions. What do you think might work for you?
• Tires: Celebrate their successes. What a great job you have done!

9. Diabetes veteran Ansley Dalbo shared lessons learned in developing Diabetes What to Know, a YouTube video channel and educational website for people with type 2 diabetes: (i) lack of information is often confused with lack of motivation (people are curious about their diabetes, but their questions often aren’t answered); (ii) online resources can fill a significant void in the healthcare landscape; (iii) video is how people learn today (few want to read a brochure); and (iv) with digital solutions, we can measure, LEARN, and optimize (“With YouTube, we can see exactly when someone starts and stops watching a video. What is engaging? Where is it confusing?”). The team has taken a “ship and iterate” philosophy, using A/B testing often – e.g., videos without patient introductions did better than those with them, videos on more sophisticated topics have higher retention rates. The company launched in June 2015 and recently received a grant from Novo Nordisk to expand its marketing. The videos have received over 80,000 views in the past three months, including a 74% audience retention rate. Ms. Dalbo concluded optimistically about the future: “I am 100% sure that digital tools will change the experience for so many people living with type 2 diabetes.” We heartily agree – it’s so impressive to hear and see more about what the beloved Ms. Dalbo has been building!

10. DiabetesMine’s founder, the great Amy Tenderich, announced the blog’s 2016 usability innovation awards: Dexcom’s G5 (people’s choice), Abbott’s FreeStyle Libre (best newcomer), TuDiabetes.org (best quality of life innovation), and OpenNotes (honorable mention). The most surprising of these was Abbott’s FreeStyle Libre, a product that is not even available yet in the US (submitted to FDA in 3Q16; Pro version rolling out to HCPs now) – this could be a sign of the pent-up demand for this product, albeit from an engaged, early adopter community. We hadn’t previously heard much about OpenNotes, an online platform that allows patients to access their doctor’s appointment notes online. DiabetesMine selected these winners from a Spring 2016 online survey (n=485; convenience sample) based on consistently high scores on impact, participants write-ins, and multiple positive comments.

Detailed Discussion and Commentary

FDA Innovation Update

FDA Innovation Update

Courtney Lias, PhD (FDA, Silver Spring, MD)

FDA’s Dr. Courtney Lias gave an enthusiastic and meme-filled talk on the regulatory state of diabetes devices, touching on the exciting MiniMed 670G approval (“1^st in the world,” following monthly FDA-Medtronic meetings for 2.5 years!), the just-finalized BGM guidances (test strip lot release criteria are a “huge deal”), and encouraging the DIY community to reach out (“#WeAreStillWaiting”). We knew FDA and Medtronic had collaborated strongly on the 670G, but we weren’t aware it was that extensive. In the monthly meetings, FDA gained an understanding of the 670G well ahead of submission, and also gave the company advice and pushed it to move faster. The result was a much faster review process (just over three months) than anyone expected (see our coverage). Dr. Lias wished that the FDA “could do this for every novel product,” and we’ll be interested to see how quickly other automated insulin delivery systems and novel sensors move through the agency. Following optimistic comments on the just-finalized BGM guidances (see below), Dr. Lias shared three “next steps” for the agency: (i) continue to help device sponsors make new and innovative technologies available; (ii) work with the community towards device interoperability; and (iii) continue seeking input from patients on their technology needs. No argument there! Dr. Lias sounded truly excited about the products currently under FDA review, though could not say anything more about them. In particular, we’ll be interested to see how quickly the consumer version of FreeStyle Libre is approved following Abbott’s 3Q16 submission.

• In monthly meetings with Medtronic for 2.5 years (!) leading up to the MiniMed 670G approval last month, FDA proactively offered solutions, rather than solely reactive assessments. According to Dr. Lias, this strategy facilitated an IDE, allowed for planning of early facility inspections, and helped FDA learn about the device and the planned study protocols. Without the interactive dialogue, it would’ve taken a lot longer than just 104 days for the Agency to pore through the “thousands and thousands” of pages of information contained in the submission. Dr. Lias hoped that FDA “could do this for every novel product,” and it’s clear to us that other companies will continue to get lots of energy and time from the CDRH team. [Dr. Lias’ team worked nights and weekends to get the 670G approved so quickly – incredible commitment.] She also touched upon Medtronic’s three-month 670G pivotal trial (n=124), which “surprised everyone” by reducing A1c in a fairly well-controlled population. Between the speedy three-month, single-arm trial and multiple feasibility studies in over 160 patients (see picture below), the company identified safety, usability/human factors, and algorithm issues that they could address pre-submission, and more importantly, pre-market (see slide below). This was also a reminder that other companies will need to do much more than just a pivotal trial. We applaud FDA for all of its hard work to moved the closed-loop field forward, where there are numerous systems (hopefully) coming to market in the next few years (see our updated automated insulin delivery competitive landscape).

• Dr. Lias noted Medtronic’s 630G to 670G upgrade program. You can read more about the details of the program here. If patients purchase a 630G pump, they can upgrade to a 670G once it begins shipping in Spring for $0-$299. The challenge for current Medtronic pumpers is that upgrading to the 630G is very expensive: $599 or $3,100 out of pocket, depending on when the pump was purchased (the majority of the base is in the latter group). We hope the model of “getting a new pump” changes dramatically, as Tandem is doing with its software-updateable t:slim X2. Pumps should be like phones, enabling continuous improvement and over-the-air upgrades.
• Regarding the just-finalized BGM guidances, Dr. Lias commented that reviewing test strip lot release criteria “is a huge deal.” FDA wants to know how manufacturers will sample test strips: How many will they test? Where will the strips be taken from (front, middle, back, of the lot)? What tests will be performed? What criteria will determine whether the lot is ok to ship? FDA will use this information to improve the consistency between manufacturers, serve as baseline documentation (so inspectors can identify when these parameters change and determine if the adjustment is acceptable), and ultimately improve the quality of test strips.
  
  • In addition to test strip quality, FDA is working on Medical Device Reporting (MDR) guidance. She noted that some BGM companies send thousands of medical device reports every year, while others don’t send any. The goal of the guidance is to clarify which events should be reported. Furthermore, the coding of these reports will be more standardized – for example, the current coding system has a designation for “hypoglycemia” and another for “low blood sugar.” This leaves the potential for companies to report the same event under different codes, making it difficult for any large-scale analysis.
• Before wrapping up, Dr. Lias broke out the #WeAreStillWaiting hashtag (a joking parody of the #WeAreNotWaiting hashtag) to say that FDA is still hoping the DIY community will reach out. “Talk to us!” she said, adding that “we will only wait so long before we start reaching out in different ways.”  The world of DIY innovators was in the audience and we hope it made an impression. We didn’t see this as a threat – Dana Lewis and Scott Leibrand were sitting next to Dr. Lias for much of the day – but it was a reminder that the FDA wants to talk and come to possible “win-win” solutions for the Agency, patients, and the OpenAPS community. This is a tricky situation but it doesn’t have to be if the lines of communication are open – it is certainly clear how hard CDHR is working and how much change they are creating. 
• Dr. Lias showed off FDA’s fresh new logo. It feels far less stodgy than the older logo and aligns with the more collaborative and transparent (at least on the device side) FDA!

Live App Demos

--by Adam Brown, Brian Levine, and Kelly Close