Memorandum
In another addition to the rapidly expanding insulin dose titration competitive landscape, Insulin Algorithms, based in Los Angeles, CA, recently announced FDA 510(k) clearance of its comprehensive insulin titration software. The clinician-facing decision support software is currently the only FDA-cleared product compatible with “all” insulin regimens (presumably all major basal and bolus insulins); the growing list of cleared products in the US typically support basal alone. In a study following 110 patients for one year, use of the system caused mean A1c to drop 3.8% from a high 11.0% baseline. (The data are published on the company’s website, but not available in a journal from what we can tell) Another study highlighted on the website (n=30) demonstrated A1c reductions of 0.7% (baseline: 8.4%) in five months. The company has partnerships with major glucose monitoring companies (including Abbott, Roche, and J&J), allowing users to check blood glucose with current meters and easily upload data to the company’s cloud server in the clinic. The server then automatically generates patient-specific dosing recommendations in 15 seconds, which the provider reviews before implementation. Details beyond regulatory clearance – launch timing, business model, pricing, algorithm details and configurability, etc. – are unclear at this point. This latest clearance by the FDA follows previous CE and Health Canada certification – clinicians in Canada were slated to begin pilot studies in early 2017, but so far, we have not seen any results. The company has an impressive list of advisors, both internally and externally: Renowned endocrinologist and former ADA President Dr. Mayer Davidson, who recently received the ADA’s 2016 Outstanding Physician-Clinician in Diabetes award, is the company’s Chief Medical Officer, and is supported by an impressive bulletin of medical advisors, including UNC’s Drs. John Buse and Richard Kahn, Tulane’s Dr. Vivian Fonseca, and USC’s Dr. Anne Peters – wow! Further, the company’s website sports evidence of interest from prominent hospitals Kaiser Permanente, Cedars-Sinai, and Clinica Sierra Vista.
- There is no doubt that both clinician- and patient-facing insulin titration software will be incredibly helpful, and we’ll be fascinated to see how the market evolves: patient-facing software is presumably harder to implement (more risk mitigation needed), but also more scalable (e.g., an HCP sets it up and then it runs on its own). On the other hand, HCP-approved software titration like Insulin Algorithms may be more likely to be followed, but it may not scale as readily (i.e., an HCP is still in the loop with every titration).
- Insulin Algorithms has come a long way since its inception as an online consultation system, which we last covered in 2014. At the time, it provided clinicians with second opinions on insulin dose adjustments through Dr. Davidson’s UCLA practice. We’re glad to see this new format, which is obviously far more scalable – embedding the titration into software that clinics can use.
Close Concerns Questions
Q: What is the launch timeline in the US?
Q: What is the go-to-market model? Is Insulin Algorithms sold to a clinic/health system as a subscription service or one-time fee? Is it a flat fee or per-head? How much does it cost?
Q: Is the software compatible with basal-only and basal-bolus and for all insulins currently on the market?
Q: With Insulin Algorithms’ glucose meter partners, is the integration only in the clinic, or can it also be done at home? Are there plans to partner with CGM companies and passive dose capture technologies?
Q: Can patients view the recommendation report through a patient portal?
Q: Can adjustments also be made remotely in between clinic visits (assuming glucose data can be accessed remotely with the current integration)?
Q: How many clinics have adopted Insulin Algorithms in Canada and EU? How many patients have been treated with it? Have there been any real-life data reflecting uptake, efficacy, cost, provider/patient satisfaction, or burdens on providers/patients?
Q: Are there any plans to develop a patient-facing version of the software in the future, which would be set up by the HCP but then run on its own?
Q: How much provider time does Insulin Algorithms require per patient?
Q: In the studies published on the Insulin Algorithms website, have other outcomes been addressed (e.g., hypoglycemia, psychosocial, time in range, glucose variability)?
Q: Are there any plans to move beyond insulin, like titrating orals, or GLP-1/basal insulin combos?
-- by Maeve Serino, Brian Levine, Adam Brown, and Kelly Close