October 16, 2023
I think I’ve always loved data. But until my daughter, Alma, was diagnosed with type 1 diabetes, I hadn’t really worked with any datasets that had this much personal meaning for me.
Alma was diagnosed about a year ago, when she was 5. It seemed completely out of the blue, and we knew almost nothing about it before that point. We quickly learned that type 1 diabetes is an autoimmune disease where your pancreas stops producing a hormone called insulin. In healthy people, insulin allows sugars in the bloodstream to transform into energy. Without insulin, that sugar doesn’t get converted, and it keeps building up in the bloodstream. This can lead to lots of serious symptoms and side effects.
Luckily, as most people know, people with type 1 diabetes can inject insulin to manage their blood sugar. But guys, it is TRICKY. It’s not one of those diagnoses where you just go home and take a pill every day and you’re good to go. The amount of insulin a person needs varies. A lot!
We are extremely fortunate in that we have health insurance and we have access to the best technology available to people with type 1 diabetes. She’s got a continuous glucose monitor and an insulin pump, and those two devices talk to each other.
These devices are AMAZING. So incredible. But it’s still hard to basically BE a pancreas for yourself or someone else. We still need to figure out the exact insulin-to-carb ratio for each meal and snack. We need to figure out the exact amount of background (“basal”) insulin she needs. We need to figure out how sensitive her body is to the insulin she gets - that also varies throughout the day. That’s where the data piece comes in. I want to look at the data from Alma’s different devices to try to understand the patterns and the nuances so that I can make the best treatment decisions for her.
I’ve been digging into this for a while, on and off, and I thought I’d start writing about the things I’m finding, the tools I’m using, and any plots I’m making that I think are helpful. If you’re interested in some of the earlier stuff I was doing, you can find a bunch of posts over on github: https://michellejw.github.io/sugar/. I won’t be updating that page going forward, but that repository does have some tools that I’m sure I’ll grab as I continue to dig in.