Memorandum

Helmsley Charitable Trust launches $5 million Diabetes Data Innovation Initiative, requesting proposals for advanced decision support software – June 1, 2015

Executive Highlights

  • The Helmsley Charitable Trust has launched a Diabetes Data Innovation Initiative, requesting proposals to develop advanced decision support software for type 1 diabetes. The goal is to help patients and clinicians take meaningful therapeutic action from disparate data sources. DDII is $5 million program and will fund one or more projects over a two-year period. It’s a big deal not only that Helmsley is sending $5 million in funds, but that it is stepping into the digital health ring – while these funds will help type 1 directly, we believe they will have an indirectly positive impact on digital health more broadly speaking, including type 2. 
  • The RFP form is here; letters of inquiry are requested by July 20, and those invited to write a full proposal will have an October 5 deadline. The process is designed to move quickly; funding should begin by 1Q16. An FAQ is here.

This morning, the Helmsley Charitable Trust (HCT) launched the Diabetes Data Innovation Initiative (DDII), requesting proposals to develop advanced decision support software. Enabling patients and clinicians with the ability to take meaningful action from the overwhelming data now being collected across diabetes and consumer devices represents the primary and overarching goal. DDII is a $5 million program that could fund projects over a two-year period, and HCT’s goal is to support projects that have a clear path to commercialization.

Projects might include (though are not limited to): (i) mobile platforms and applications; (ii) electronic medical/health records; (iii) embedded wireless sensors and communications; (iv) big data analytics; and (v) integration with consumer, patient, healthcare professional devices/platforms (“likely essential”). Priority will be given to projects that offer a complete system (rather than part solutions) to help patients and clinicians take meaningful action from the various data resources – for example, something that takes existing device data, plugs into medial records, and delivers recommendations instead of just another app. 

Anyone can apply for funding, including for-profit, non-profit, and international organizations – a major aim is to expand beyond those already working on diabetes data. We list some likely candidates below, though we view this as a starting point! HCT plans to award at least one grant, and the total number will depend on the quality of submissions, according to trustee David Panzirer. The RFP form is here; a letter of inquiry is due by July 20; and those invited to write a full proposal will have an October 5 deadline. Funding will begin by 1Q16. An FAQ is here – some notable takeaways on interoperability and business plans are below.

Notably, the Helmsley Charitable Trust has engaged two highly regarded technology experts as advisors to help evaluate proposals – both give us major confidence that funded projects will be meaningful for people with diabetes, useful for clinicians (and ultimately researchers), and adhere to the very best practices in digital health.

For HCT, a critical element of the initiative’s success will be determined by applicants’ ability to navigate the regulatory hurdles necessary to provide meaningful, advanced decision support software. Though the technology largely exists to do this today, a big challenge is getting something safe and effective through the FDA, particularly when related to insulin calculations. We assume the best candidate projects will likely require a 510(k) or PMA submission. 

We are excited to see this proposal, as this is exactly where diabetes care must go – toward solutions that can scale. There simply aren’t enough endocrinologists and primary care doctors to manage patients, and those available aren’t being compensated to interpret the slew of data now being generated. We hope that this initiative produces projects that make life easier for patients and providers (e.g., personalized trend identification, help with insulin dose titration), as well as valuable population level solutions that save the healthcare system money (e.g., identifying risk of severe hypoglycemia, particularly those requiring emergency room visits or stays or ambulance calls). 

Below we provide more background, discuss who might apply to this initiative, and share why HCT didn’t fund any proposals for the previous $30 million Automated Insulin Delivery Initiative (in short: Helmsley did not see compelling evidence that the applications had a clear pathway to getting a product to market).

DDII Background

  • HCT will evaluate the initial letters of inquiry (due by July 20) based on:
    • Ability to reduce the burden of managing T1D;
    • Technical and scientific merit, and novelty of approach; and
    • Organizational capabilities, core competencies, and track-record.
  • “A sustainability plan and/or go-to-market model is required.” The RFP makes this point very clear, as it should – business models have historically been a weakness for digital health companies. Reimbursement is not easy, and to date, the enterprise (B2B) model has been a go-to for many companies (e.g., Omada Health, Livongo). WellDoc has perhaps the most unique business model, as its BlueStar software is reimbursed like a drug, though the novelty of its approach (i.e., a prescribed app) has its own paradigm-changing challenges. Applicants must also specify how the market will be made aware of the technology and how it will be distributed (operationally and financially). We’re glad to see that the RFP asks this of applicants, since great solutions are meaningless without business models that detail awareness, education, communication, distribution, etc.
  • “It is expected that newly developed technologies will embrace data interoperability and easily share data,” notes the RFP. As a reminder, JDRF and HCT funded University Health Network (Toronto) to develop standards for CGMs (published in December 2014), pumps, and BGMs.
    • In discussing DDII, the HCT team reminded us that Bluetooth-enabled does not mean interoperable – the latter is much more complicated and requires standardization of data formatting, among other things.

Who might apply for DDII?

  • Below, we list the organizations that first came to mind when we heard of this proposal – this is intended to be food for thought and by no means an exhaustive list (let us know if you can think of others!). The below is broadly listed in order of who is already working DDII’s goals.
    • Tidepool: HCT funded Tidepool in late April ($900,000, 18-month grant) to help build out its platform, and its device uploader. Tidepool has talked in the past about building decision support software (“Sonar”), and we imagine the non-profit might well apply for this RFP as well. As of our last update, Tidepool’s Blip and Uploader were in beta and set for public release this summer. Tidepool is really working on the most basic problem with diabetes data – getting it downloaded easily and all in one place. We assume DDII could help it take its work to level two: actionable recommendations from the now-collected data.
    • Glooko: Glooko raised $16.5 million in March to add CGM and pump data to its device-agnostic platform (Medtronic was a first-time investor in this round). Glooko planned to use the funding to add personalized predictive algorithms (e.g., analyzing blood glucose data alongside exercise data; making provider suggestions on how insulin delivery settings should change; and bringing more analytics around infusion set changes). These aims would be in line with DDII.
    • OneDrop: The company launched its One Drop diabetes logging app for iPhone on the app store in April, which includes plans for future pump/CGM integration (as long as the device sends data to Apple’s HealthKit, which Dexcom’s current Share Receiver and Gen 5 will do). One Drop founder Jeff Dachis has significant experience in building businesses on Big Data. Previously, Jeff was CEO/Founder of the Dachis Group, which built a proprietary social analytics platform to monitor over 10 billion social behaviors in near real-time. The company was acquired by Sprinklr in 2014 for an undisclosed sum. At the time, the combined companies had raised over $95 million in venture capital, acquired 11 companies, and served over 50% of the Fortune 500. One Drop’s previous expertise in analytics/Big Data and future ambitions gives us high confidence that DDII would be an excellent fit.
    • Medtronic: In addition to the recent Glooko investment, Medtronic has perhaps the most advanced clinical decision support in CareLink 3.0; however, this has not been updated in several years. Medtronic recently partnered with IBM’s Watson Health to improve diabetes care through cognitive computing, analytics, and Big Data. The aims sound very similar to what HCT hopes for with DDII, and we wonder if the Northridge-based company will apply.
    • Dexcom: Dexcom has long talked about pattern recognition and more advanced clinical decision support, and we imagine this will be even more possible once Gen 5 is approved (expected by end of year). The company has been very open and forward thinking on the data and digital health fronts, with partners that include Tidepool, Diasend, Glooko, Databetes, Training Peaks, and Apple. The company’s Bluetooth-enabled Share Receiver already posts data to Apple’s HealthKit, enabling a pretty seamless flow of data to other apps (i.e., thus far, we’ve gotten CGM data into mySugr, OneDrop, and MealMemory). Dexcom also released an Apple Watch app in tandem with the launch on April 24. HCT previously funded Dexcom on Gen 6, and we imagine several CGM analytics approaches could fall under DDII.
    • WellDoc: The company’s BlueStar app is made for type 2 diabetes and incorporates some of the elements DDII is looking for; we imagine a type 1 version could get funding under this initiative as well. A range of patients and clinicians and researchers has hoped that WellDoc would create a product for type 1.
    • Livongo: The startup’s cloud-connected meter has some limited pattern recognition and contextual feedback, and DDII could certainly help take it further.
    • Telcare: Along with a cloud-connected meter, Telcare has an analytics and coaching platform that could be a fit for DDII. We are not sure what capabilities this has and have not personally tried it. Telcare raised an impressive $32.5 million in October 2014.
    • Artificial pancreas groups: Many artificial pancreas algorithms do pattern recognition and have adaptive components; groups like UVa and DREAM have talked about applying their algorithms to optimize insulin pump settings.
    • Insulin manufacturers: We see insulin dose titration as a major need for patients and providers, and the insulin companies do this routinely as part of clinical trials. We could imagine titration software being funded under DDII – Sanofi might be a logical choice, given its iBGStar meter business and long-whispered-about Lantus calculator.
    • Big Tech companies: Google, Amazon, Apple, Facebook, Microsoft, and Samsung have all expressed interest in healthcare, and their collective software experience could meaningfully contribute to the field of type 1 diabetes. We hope they hear about the proposal and bring some novel ideas to the table.
    • Nightscout and #DIYPS: We wonder if the Nightscout community and #DIYPS (entrepreneurs Dana Lewis and Scott Leibrand of #WeAreNotWaiting fame) might apply to this project, given their patient perspective and diabetes-specific software development. If so, we imagine they would partner on additional elements of their business model, including the regulatory side in particular.  

Why did HCT choose not to fund any grants under the 2013 Automated Insulin Delivery Initiative?

  • This may sound familiar to HCT’s major $30 million Automated Insulin Delivery Initiative in December 2013. Ultimately, no proposals were funded. In a recent interview, the HCT team candidly shared that they did not see compelling evidence that the applicants had a clear pathway to getting a product to market – in short, the funding would not have been impactful. Some proposals were academic endeavors, while others had an eye on commercialization but didn’t bring PMA experience, didn’t have approval from the Board of Directors, etc. We were surprised to hear this and are happy to see that HCT is being very upfront about what is likely to be successful here.
  • HCT is applying similar criteria to this RFP – proposals MUST seek to commercialize a product with high potential for near-term impact on people with type 1 diabetes. With the DDII RFP, HCT is not interested in research endeavors, and wants to get solutions into the hands of people with diabetes as soon as possible. Proposals really need to cover all bases, especially a go-to market strategy.

Close Concerns Questions

Q: Who will apply for DDII? What organization is best equipped to carry a product in this area to market? Will any non-diabetes players be interested? 

Q: What will the regulatory pathway look like for meaningful decision support software?

Q: What will a successful business model and go-to market approach look like? 

Q: Will any company be able to solve the interoperability challenges, particularly on the electronic medical record (EMR) side? 

 

-- by Adam Brown and Kelly Close