Type 4 Fun!

In Part 1 of Drugs, Technology & Goops I wrote a bit about my experience with CGMs to monitor glucose levels continually. It was a limited study in only what I had personal experience with but it teased a few other topics. I’m still building larger data sets to expand on my study of how exercise impacts my glucose levels – there are some interesting things to report there – as well as increase the sample size of my G7 Libre 3 MARD comparison. There have been some developments to touch on and I’ll briefly describe some of the other ways technology can impact my health outcomes.

First, a correction and point of clarification. I stated that Medtronic did not have stand alone CGM that wasn’t a part of their AID (automated insulin delivery) product line. This is not totally accurate, while it is being discontinued this year Medtronic has offered the Guardian Connect since 2018. It is being replaced by the Simplera CGM as a part of their Smart MDI (multiple daily injections) system which includes a smart insulin pen, which is interesting. The Simplera may suffer from accuracy issues when compared with other CGMs, however.

Secondly, a quick note about data. There is a lot of it out there, and not all of it is good or effectual. One of my complaints with apps and software meant to collate metrics is that they don’t provide access to the raw data. The Libre 3 app doesn’t give you the option to download anything let alone parse individual data points. Similarly, the Dexcom companion app Clarity only allows you to download reports but not create your own.

Additionally, I recently learned about the non-profit Tidepool. A software company with tools for clinicians and people living with diabetes manage their data and even provide automated insulin dosing. It’s fascinating and encouraging to see a third-party try to develop additional tools and options. I’m excited to see what if it can offer me the access that I am hoping for.

Finally, I mentioned that I would hope to have highlights from the recent Advanced Technologies and Treatments for Diabetes conference! I have those now!

Some presentations of note came from Medtronic on maximizing outcomes for people with diabetes using their MiniMed 780G insulin delivery system in exercise. They have updated their guidance to users and healthcare providers and found significant TIR (time in range) improvements using a fixed insulin dosing regimen. This update to guidance will be helpful for pump users to plan for exercise and adapt to the needs of unplanned exercise.

And since artificial intelligence is being applied in every imaginable scenario or function, another interesting presentation came from the University of Virginia’s Center for Diabetes Technology. They introduced the first outpatient studies involving a neural-net artificial pancreas. The particularly relevant significance of these studies is the possibility of leveraging AI to develop systems capable of adapting to frequent physiological changes that have prevented current AID systems from generally achieving higher than 70% TIR. This kind of machine learning could have deep implications for athletes as fitness and fueling demands ebb and flow throughout a training block or competition season.

There were other discussions of AID in exercise, as hypoglycemia is a frequent concern for people with T1D and the fear of which is cited as a principle reason not to exercise. Real world evidence found around 20% of physical activity results in hypoglycemic events, 74% of which occurred during aerobic exercise. Typical guidance for users of AID is to set a temporary glucose target for the duration of a planned activity, this was used in 75% of reviewed cases but in at least half of them the target was set within a hour of the start of the activity which is considered to be too late.

Part… 3? I’ve lost track. It doesn’t matter, this part will be brief.

Blood sugar is not the only health metric that I (and many others, perhaps unnecessarily) track as an amateur athlete. Here are some of the others and a quick thought about them.

1. Heart Rate Monitoring

This comes in many forms each with differing degrees of precision. This article from Cleveland Clinic does a really great job of outlining the differences in optical and electrical heart monitoring and even compares the various wearable methods of detection.

I personally wear a Suunto Vertical Titanium Solar and love it for it many reasons but I find the HR data accuracy to vary sharply based on levels or types of activity. This anecdotal data seems to coincide with the conventional understanding of photoplethysmography versus electrocardiography.

I’ve only ever worn one other wrist based, optical heart rate monitor. It was also a Sunnto watch, the Suunto 9 Baro. I would say any difference between the two sensors is negligible in my experience but I look forward to seeing how the technology advances.

2. Sleep Tracking

For a deep dive into the weeds of sleep tracking, I recommend this study but in terms of the reliability of commercially available wearable devices it is all somewhat dubious. Consumer devices rely on heart rate and/or body movement to assess levels of sleep or wakefulness and that can be algorithmically misleading. Brain activity is a key indicator that can’t be measured by an optical sensor on your wrist or the Withing Sleep pad that I have under my mattress which uses pneumatic sensors to pick up on my body movement, respiration rate, and heart rate. That the pad can pick up a heart beat through a mattress is a pretty wild thing, but personally I’m not going make any health decisions based off it. I’m not sure I’ve ever look at the Withing app for sleep data and felt like it accurately represented the night before.

3. Body Composition

This is another area where I am hardly an expert. Frankly, it baffles me that electrical signals measured by a bathroom scale can delineate bone, water, fat and muscle weight. I have another Withings product here the Body Cardio model of scale which I’ve had for a number of years. One of the interesting things about the Withings scale that I appreciate is that Withings appears to acknowledge its limitations. When I first purchased the scale, one of the metrics that it claimed to be able to measure was Heart Rate Variability, a few months after I started using the scale, a firmware update removed HRV.

HRV is an important because high or low variability reflects on the body’s ability to adapt to changing situations. a high HRV may indicate lower stress levels while a lower HRV may indicate current or future health issues (diabetes being among these)

Athletes can therefore find HRV data to be very valuable. The only problem is that it is also very difficult to measure. In clinical settings, it is captured by an electrocardiogram machine in a lab or a Holter monitor that uses multiple electrodes affixed to your chest that you would wear for a day or two. Outside of medicine, there chest straps and other pulse oximeter like devices that can be used but to claim a wrist worn device or a scale that you stand on for a few moments is going to be able to accurately capture HRV data is highly suspect.

---

I’m very skeptical of products like the WHOOP band or the Oura ring primarily for this reason. Being aware of the challenges with optical heart rate sensing, it feels rather bold to make the kinds of claims that these health trackers make. I’d love to be proven wrong, I just think that for the kinds of prices these products go for and how expensive it would be to compare, it is cost prohibitive to experiment with.

Living with diabetes and taking the medications that I need impacts my cardiovascular health in a way that is potentially substantial enough based on studies that I would love to be able to see my own granular metrics. Perhaps, one day reliable lab quality data will have common everyday accessibility.