(skip to content) jana e. beck

HFCT #1: Motivations

HFCT is the acronym I’m using for a high-frequency cholesterol testing study I’m conducting on myself (along with a group of QS-ers conducting similar studies) in the fall of 2017. This is my first post about the project!

High-frequency what?!

For six weeks spanning the end of September—I started last week!—through the beginning of November, I’m participating in a high-frequency cholesterol testing study along with a small group of other self-tracking enthusiasts. The study is being organized and guided by Quantified Self and sponsored by Amgen. To the best of my knowledge, it’s the first study of its kind, and a rather oxymoronic-sounding kind it is: a group n=1 study. I can’t say it any better than this introductory post about it on the Quantified Self site:

At the 2017 Quantified Self Global Conference, we met to discuss a collaborative QS project that we’re calling, “Blood Testers.” Our immediate goal is to learn more about ourselves from high frequency self-testing of our blood lipids (i.e., cholesterol and triglycerides). Our long-term goal is to advance progress in self-directed research by better understanding what makes these types of projects succeed or fail. In most Quantified Self projects, one person does almost all of the work, perhaps with a bit of advice from friends and online feedback. But what if you could work in a group of people with varied skills to explore questions you developed through conversation and collaboration? Everyone can pose questions and determine for themselves what data they want to collect, but can also benefit from others’ unique skills, compare results, and team up to tackle challenges like device validation and data analysis. The idea isn’t to take control away from the individual, but to provide resources that connect the community through developing shared methods.

This post describes what’s motivating me to engage in this study. Keep an eye out for more posts soon about my experimental plan and protocols and what I’ve learned so far, after going through the device validation and first few days of my experiment.


I have type 1 diabetes. Just in case: type 1 diabetes is an auto-immune disease that results in the destruction of the insulin-producing cells in the pancreas, requiring those with the condition to inject or infuse (via an insulin pump) exogenous insulin. Insulin is a potent substance; to put it very briefly and bluntly, too little insulin kills you slowly (or makes you go blind, suffer kidney failure, and/or require the amputation of extremities due to peripheral nerve damage, to name just the most common complications), but too much insulin can kill you very quickly: within hours, if not minutes. So with type 1 diabetes you have to fucking get it right, and you don’t get a break. Ever. Constant data collection, review, and adjustment (of real-time, hour-by-hour insulin dosing and medium-term insulin pump settings, mainly) is required for me to have even a halfway decent control of my blood glucose and thereby stay healthy and energetic with minimal diabetes-related interruptions in my life1.

Because of how necessary and compulsory diabetes self-tracking is in my life, I’ve always had a somewhat conflicted relationship with the Quantified Self movement. To the extent that I’m also a data nerd and would love to have data about all sorts of aspects of my life to play with2, I feel right at home with QS-ers. But to the extent that I don’t have a choice about self-tracking when it concerns my diabetes, I don’t feel as strong of a connection with “hobbyist” QS-ers.

All of this is a long-winded way of saying that what’s attracted me to participating in ‘Blood Testers’ is the novel—for me—experience of putting a lot of time and effort and discipline into tracking something that is not diabetes, at least not directly. Only four days in, I’m already annoyed with the food logging involved in my experiment—I fucking hate food logging, but I also want the data that comes out of my experiment to be worth analyzing, and that’s what’s helping me through the annoyance of food logging. (I’ve also designed a food logging protocol that gets around the things I hate most about it; see post #3 for details.) It’s interesting to be fully in control of what I’m investigating, including what data I’m collecting and how I’m conducting my experimental protocols—including how strict these are, with the obvious tradeoff of lower data quality under looser protocols. In talking to others about my project, I’m finding myself using the phrasing “I have to…” (e.g., “I have to bring my own food,” “I have to go test my cholesterol now,” &c), but it doesn’t feel quite accurate. I’m choosing to do all these things for the sake of an experiment—and, more importantly, an eventual dataset—I want to be worthwhile. Having and recognizing that choice is rather interesting, even compelling.

Some secondary motivations for my participating in this study are that, first, I have a near-stratospheric HDL. On every fasting lipid panel test I can remember of my adult life, my HDL has been over 100 mg/dL 🚀. (For context, the bar for a “desirable” level of HDL (according to the Mayo Clinic) is anything above 60 mg/dL, and a value over 100 mg/dL places me well into the 95th percentile according to this set of statistics from the CDC for women, although the data was gathered across 1988-1994.) One primary care practitioner a while back told me that high HDL can be a genetic thing, and sure enough, I asked around in my family, and it looks like my father’s side of the family has a fairly strong genetic predisposition to high HDL.

Genetically high HDL hasn’t been studied much. So one of the things I’m curious about in this study is to see how variable my HDL is, especially as related to my diet, and I expect I’ll be doing a bit more research into what is and isn’t known about high HDL. Just this evening, I’ve come across a paper published in Science in March of 2016 that connects one particular genetic cause of high HDL with an increased risk of coronary heart disease (CHD), against the general trend that correlates higher HDL with reduced risk of CHD—the reason for HDL commonly being called the “good” cholesterol. I haven’t done enough research yet to fully absorb this paper, but I found this snippet from the conclusion relevant to the n=1 experiment I’m embarking on:

Our results are consistent with a growing theme in HDL biology that steady-state concentrations of HDL-C are not causally protective against CHD and that HDL function and cholesterol flux may be more important than absolute levels.

If I’m understanding what’s meant here by “cholesterol flux” correctly—and I’m not sure I am; I’m in the early days of learning the intricacies of the lipid system—then my study that includes four days of lipid testing at two-hour intervals could be really interesting to see if I have any “flux” at all in my HDL and/or other lipids. I don’t know if tracking the kind of intra-day change in lipid levels that I’ll be doing (at two-hour intervals on four days across my experiment) gives evidence of the kind of “cholesterol flux” the authors are talking about, but perhaps it will.

Finally, I’m piggy-backing on the food logging I’m doing—have I mentioned yet how much I fucking hate 😡 food logging?—to study the relationship between my blood ketones and diet. I’ve done a lot of reading over the last…almost a decade now about low-carb and ketogenic diets, and I have evidence that following a low-carb diet is beneficial to me in that it (statistically and quantitatively) significantly improves my blood glucose control. But keeping to a ketogenic diet (as opposed to a lowish carb diet) all of the time is not something I want to do. So I’m very curious to learn about my ketone (and blood glucose, naturally) response while going from a designed-to-be-guaranteed-ketogenic diet to a more moderate, but still quite low carb diet—the “lowish” carb I’d more-or-less like to maintain long-term that I’m not certain is low enough to result in enough of the benefits that a strictly low-carb and ketogenic diet does.


The next post is all about my experimental plan.

  1. Such interruptions include, for example, sudden and extreme drops in blood glucose that require immediate handling and may sideline me for a half hour or more from whatever I’m doing. 

  2. I don’t often succeed with extra-diabetes self-tracking projects because I get fatigued with the tracking pretty quickly.