News from the Tandem investor and analyst day Sept 25, 2018

There is a large slide deck at for some talks given at the Tandem investor day. If you want to listen to the talks they should be available for about a month via a link at

Here’s a few highlights.

They talked about business conditions and how fast they are growing (they have paid off their debt, and expect to be EBITDA-positive in 4Q18, and reach cash-flow break-even in 2H19.) There are lots of slides about the business, but I’ll focus on the technology and plans.

They discussed product plans for pumps, with a focus on the evolving systems for algorithms to use CGM to control insulin dosing. They liked the iCGM approval Dexcom got for the G6, and will go for a iPump designation for their pumps. The idea is to decouple the FDA approval of pump hardware and software from the FDA approval of closed-loop algorithms. This will make it much faster and less expensive to roll out new pumps that can use approved closed-loop algorithms. And I see that the t:sport pump is back on the roadmap, for 2H20. This is a quite small non-disposable pump that holds 200u. It uses a very short tube to a separate infusion set, and it takes dosing commands from the phone rather than the pump body, so it is somewhat like an omnipod. They like the short tube rather than a no-tube design in case an infusion set fails or some adhesive fails, and because the insurance companies like permanent pumps with disposable supplies. Supposedly with this tiny pump taped next to the infusion set there’s no snagging of a long tube, no unsightly tube hanging out anywhere, and the pump is naturally hidden from view. Oh, and it’s IPX-8 which you can think of as “ok to submerge in water to a depth of 2 meters for 1 hour,” but we’ll have to wait for the actual product spec to know the manufacturer’s guaranteed depth and time.

They talked about the study results for the recently-approved predicted-low suspend algorithm (“Basal-IQ”). In the study they experienced about 1/3 less time below 70mg/dL. And there was no increase in mean BG or time over 170 (i.e., they don’t cause rebound highs). Folks were suspended about 1 1/2 hours per day on average, and really liked the system. Not just for the BG control, but also for the ease of use, with no fingersticks and minimal alerts.

They talked about the hybrid closed-loop algorithm (“Control-IQ” nee “inControl”) they got from TypeZero. It runs a 5-minute cycle, and makes decisions for basal attenuation, basal increase, and boluses for hyperglycemia correction. They gave results from small published studies (17 over 85 days at Stanford summer camp, 14 over 6 months at home) that showed about 78% of the time between 70 and 180, and less than 2% of the time below 70. The ongoing Control-IQ pivotal trial has 168 patients. In addition to the clinical outcomes (time in range 70-180) they have 2 broad goals: the technology should be easy to use, and the patients should find it useful. Evidently the folks in the clinical trial are liking it.

In the roadmap after Control-IQ comes out (expected 2019) the next version is characterized by no carb counting and “personalized settings.” From reading elsewhere, I think “no carb counting” means you announce a meal as small, medium, or large, and the algorithm takes it from there. I’m hoping that “personalized settings” means we’ll be able to set our own target BG and high/low range, rather than the standard 70-180.

They showed this provocative slide comparing happiness with t:slim vs 670G.


@bkh, that is really exciting, thanks so much!


Real exciting the pump uses your programed basal rates and the automation only kicks in to prevent lows and highs.