October 23, 2017; Palo Alto, CA; Highlights – Draft

Executive Highlights

Hello from the Four Seasons Hotel in Palo Alto, where HITLAB’s 2017 digital health symposium (sponsored by Novo Nordisk) just wrapped up. Watch the entire morning session here and the afternoon session here. We have three notable highlights from today:

1. A keynote from Novo Nordisk’s VP of US Device Research Thomas Miller shared a valuable 90-second video of the company’s digital health “vision” – complete with a connected insulin pen (it looked like an auto-injector); a patient-facing app with dosing guidance, automatic carb counting, and real-time CGM; and a healthcare provider dashboard with clinical decision support. Watch the video here from 55:40-57:16; see below for our takeaways and screenshots.

2. HITLAB announced the five startup finalists for its 2017 World Cup, also sponsored by Novo Nordisk and exclusively focused on voice-activated solutions for diabetes. Five finalists were chosen out of 146 applications from 15 countries: Palette, T2D2, Light House Powered by ADA, My Diabetes Coach, and Proof Work. We’ll have to wait for November 29 to see demos and the winner at HITLAB’s New York City conference. The organizers said it will likely be webcast.

3. Glooko’s Marianne Berkovich and Linda Parks shared that the company’s mobile insulin dosing system (MIDS) for titrating basal insulin in type 2 diabetes remains under FDA review (“hoping to hear soon.”). Notably, Glooko now has data on over 1.5 million people with diabetes, up from “1+ million” as of the last update in January – wow!

• Who is HITLAB (“Healthcare Innovation and Technology Lab”)? Today was the first contact we’ve had with this consulting group. According to their website, “HITLAB helps organizations innovate through an iterative process of ideation, creation, evaluation and diffusion. Together, we generate new ideas, implement solutions, and test for real world outcomes.” HITLab’s expertise ranges widely: strategy, UI/UX, prototyping, business models, piloting, clinical trials, etc. From what we can tell, they are a sort of IDEO-meets-clinical research organization.

Top 3 Highlights

1. Novo Nordisk’s “Vision” For the Future of Digital Health – Connected Pens, CGM, Dose Guidance, HCP Decision Support

Novo Nordisk’s VP of US Device Research Thomas Miller shared Novo Nordisk’s compelling future vision of digital health-enabled insulin delivery. Watch the cool 90-second video here (55:40-57:16), which shows: (i) a connected insulin dose capture device, which appeared to be a single-use auto-injector with a 6-unit dose; (ii) a patient wearing CGM and receiving real-time insulin dose guidance, hypoglycemia prediction alerts, and carb counting advice from an app; and (iii) a healthcare provider dashboard with clinical decision support (“add bolus injections”), time-in-range, and an AGP. WOW! We’ve taken notable screenshots of the video and included them below. The talk did not share specifics on Novo Nordisk’s connected pen efforts (a pilot with NovoPen 5+ is ongoing in Sweden) or partnerships with Glooko and IBM Watson. (Mr. Miller even emphasized this video was a “vision” and included actors.) However, the commitment to capturing insulin dose data from pens is clear – for Novo Nordisk, we assume it’s more about “how” (e.g., disposable vs. durable, battery, cost, business model) and “when” vs. “if.” It’s unclear how far along things are – especially on the device front – but the video was a strong vision and we hope pieces can be achieved quickly. Novo Nordisk announced its partnership with Glooko in January (of which dose titration was a clear goal), and subsequently launched the Cornerstones4Care app powered by Glooko in July.

For those interested, we recommend watching the full ~20-minute keynote here (it starts at 39:26).

• Notably, Mr. Miller showed three concentric circles with “medicine” at the center, surrounded by “devices” (quality of life, usability) and then “digital services” (personalized, active guidance) on the outside. He also mentioned supporting an “ecosystem” several times, in line with the company’s digital health partnership strategy – we wonder if a glucose sensor partner is on the radar, given the prominent use of CGM in the video.
• In a later slide on “seamless integration,” Mr. Miller noted that insulin pen connectivity and smartphone pairing “needs to happen quite fast – if it takes less than a second, that’s seamless. If it takes 1-10 seconds, that may be OK, but then we need to add some feedback on the phone along the way. If it takes more than 10 seconds, people are not going to use it.” We love this focus on seamless and look forward to seeing the company launch more digital-enabled services.

2. HITLAB World Cup Finalists Focused on Diabetes+Voice Solutions: Palette, T2D2, Light House, My Diabetes Coach, Proof Work

HITLAB announced the five startup finalists for its 2017 World Cup, also sponsored by Novo Nordisk and exclusively focused on voice-activated solutions for diabetes. Five finalists were chosen out of 146 applications from 15 countries – none were demoed today, but see below for the details we could dig up. Demos and the winner will be announced on November 29 at HITLAB’s New York City conference. The organizers said it will likely be webcast.

• Palette – A smart placemat that sits on a dining table and tracks meals’ nutrition. We’re not sure what the voice component is, but presumably it allows for voice-based logging.
• T2D2 –A “virtual nutrition assistant” that uses machine learning to provide in-the-moment feedback, meal planning, and food/glucose logging. T2D2 was also a Merck/Amazon Alexa Diabetes Challenge finalist.
• Light House Powered by ADA – Expanded chronic care outside of appointments, including ongoing patient education and remote monitoring. From the website: “‘Alexa, ask LIGHTHOUSE what my priorities are for today?’ The power of voice puts your practice on your patients’ kitchen table.” It’s notable that the product is co-branded with ADA, a sign that professional association remains interested in digital health. Like other finalists, it’s hard to understand what this product specifically does.
• My Diabetes Coach – “Personalized feedback and suggestions based on actual behaviours.” It’s very difficult from the company’s recent press release to understand exactly what it does.
• Proof Work – A “voice activated block chain ecosystem.” We cannot find anything online about it.

3. 1.5 Million Glooko PWDs, 7,000 providers, 7 Billion Data Points; Glooko's Mobile Insulin Dosing System Still Under FDA Review

Glooko’s Marianne Berkovich and Linda Parks shared that the company’s mobile insulin dosing system (MIDS) for titrating basal insulin in type 2 diabetes remains under FDA review (“hoping to hear soon.”). Notably, Glooko now has data on over 1.5 million people with diabetes, up from “1+ million” as of the last update in January – wow! Most of the talk highlighted the fascinating distinction between usability and human factors testing – a critical area of diabetes and digital health as apps move to insulin titration. Ms. Berkovich provided a case example of Glooko’s Mobile Insulin Dosing System (MIDS), which we learned is still currently being reviewed by the FDA – it was in the FDA process in mid-May, implying there has been quite a bit of back-and-forth with the FDA. As a reminder, MIDS is a patient-facing insulin dose titration app for people with type 2 diabetes on long-acting insulin; a provider configures the treatment plan, and the app provides continually-updated basal dosing guidance based on fasting glucose levels (i.e., similar to Voluntis Insulia, Amalgam’s iSageRx). As part of the user experience, a simple screen prompts users to confirm their fasting blood glucose, along with instructions on how to edit the values if they are incorrect (click on the pencil; see below). The screen passed usability testing, but human factors testing revealed that users automatically confirmed values, even when supplied incorrect numbers. Some users didn’t understand that the pencil indicated the path to “edit” an incorrect value. Of course, such oversight could lead to potentially dangerous automatic titration decisions – e.g., if a post-meal glucose of 242 mg/dl was considered a “fasting” glucose (and not edited), the app would use it as the basis to up-titrate the dose. This kind of human factors testing will be critical as apps move to more ambitious support – and we thought the screen below looked fine! Ms. Berkovich beautifully emphasized that what is easy for engineers is not necessarily practical for the average consumer and cautioned against assuming user familiarity with terminology (e.g., “fasting blood glucose”). Points well taken!

--by Adam Brown, Maeve Serino, and Kelly Close