How A Diabetic Uses Simple Analytics To Keep His A1C’s Below 6.0%

A little background. My employment history includes companies like SAS Institute (advanced statistics and analytics), Microsoft (personal computing), and Oracle (relational databases). I’ve also worked for Big Data companies that specialize in providing customers with elegant solutions to some of the most difficult data-related problems.

Because of that background, it should come as no surprise that I tend to look for technical solutions to any problem. As a type 1 (insulin dependent) diabetic since 1982, it’s not difficult to figure out what one of my most important problems would be.

The problem. As any diabetic will tell you (whether Type 1 or Type 2), our only goal as a diabetic is to maintain a blood glucose level within the normal range. To drive between the lines, so to speak. “Normal” is defined as being near 80 mg/dl, or what a non-diabetic would average over a long period of time. Effective insulin utilization is required to make that happen.

Insulin is the hormone that a non-diabetic’s pancreas produces in precise amounts so the body’s cells can process glucose for energy. For non-diabetics, the pancreas produces just enough insulin so that glucose levels remain in the normal range. Going above normal, also called hyperglycemia, can result in long-term health effects like blindness, nerve damage, kidney damage, coma, and amputations. Going below normal, also called hypoglycemia, is more dangerous short-term and can result in loss of consciousness and ultimately even death.

Because diabetics either don’t produce any insulin (Type 1) or their bodies do not properly utilize what it does produce (Type 2), it requires manual injections of insulin or oral medications which stimulate the pancreas to produce more (or both). Without these medications, diabetics would not live very long. Even with these medications, the cognitive cycles required to remain between the lines and within the normal glucose range are enormous and persistent.

Enormous because the amount of insulin that is required to utilize the resulting glucose from a digested meal can fluctuate based on a number of factors like current glucose level, amount of carbohydrates, types of carbohydrates, time of day, recent exercise duration and intensity, dehydration, caffeine, illness, and even stress.

Persistent because a diabetic is thinking about these things any time they eat, exercise, drive, go to sleep, or do pretty much anything. It’s a never-ending process of just staying alive. Calculating how much insulin to take can only be a simple formula if you eat the exact same thing and do the exact same thing every day. Unfortunately, that’s not realistic.

How would it be possible to calculate an accurate insulin dosage for each of my meals when all — or even some — of those factors are involved?

To figure out the answer, let’s address the problem one step at a time.

Step one requires me to calculate how much insulin (if any) is required to get us from our current glucose level of, for example, 150 mg/dl to my target of 80 mg/dl.

Step two requires me to calculate how much insulin is required to offset the meal I’m about to consume. At the very least, this will require me to count the number of carbohydrates by looking up the “nutrition facts” either on the product packaging or online, and it may even require the use of measuring cups.

Step three, the last and most important step, requires me to consider all of the factors which are in play for this particular meal. Have I exercised much today? Am I stressed? Am I sick?

Years ago, to get the answer, I would calculate an insulin dosage by doing a quick calculation based on the ratio of 1 unit of insulin lowering my glucose by, for example, 10 mg/dl. Therefore, to go from a glucose level of 150 to 80 would require 7 units. Then to determine how much insulin I would need for the meal, I’d count the number of carbs and do a similar ratio calculation. And I’d be wrong almost every time.

If the dosage was too much or too little, I’d rely on taking another glucose reading 2 hours after the meal so I could either give more insulin (to lower my glucose) or eat more food or glucose tablets (to compensate for taking too much insulin). While neither is healthy, it was the best I could do.

Today, I do something else, and I’m almost always accurate with my dosage and my A1C’s are consistently below 6.0%.

What do I do? I use the property of repetition to my advantage.

You know this Einstein quote? “Insanity: doing the same thing over and over and expecting different results.” The key to this is knowing that doing the same thing over and over will yield the same result, so long as you do it the same way every time.

In my meal scenario, if I’m about to have chicken alfredo from one of my favorite restaurants and I’ve exercised that day and I’m stressed about something at work (all factors that impact my glucose level), how do I calculate the most accurate insulin dosage? What do I need to know?

Asked a different way, if I’ve had that exact same chicken alfredo meal with those exact same factors six other times in the past, would it be easier to calculate an insulin dosage? As long as I have a log history for that particular meal, I can review my past dosages for that meal, calculate how accurate those dosages were by looking at my glucose readings a few hours after each meal, then come up with a dosage based on my current glucose level.

No, I don’t do it in my head or even use a spreadsheet. I use a tool that you can use as well.

I have logged those meals into a relational database and I let a few algorithms calculate a dosage for me. It takes less than ten seconds to calculate a new dosage and log the meal. The next time I have that same meal, my accuracy will improve even more because I’ve added another log for that meal to include in the algorithm.

It’s important to remember that each meal needs to be as similar as possible. For this to happen, we’re fortunate on two fronts. First, restaurant meals and frozen meals are typically the same portion sizes over and over again. In a world of mass production, this helps a lot. Second, when we’re serving ourselves at home, our tendency is to consume the same portion sizes each time. Because of this, all we need to do is track the different factors that are in play for each of our most common meals, and we’ll have a history to make more informed and accurate decisions moving forward. If you’re like me, you’ll find that you consume the same 15–20 meals over and over again.

Use your own repetitive tendencies and consistency to your advantage to make diabetes a little easier to manage, and take some time to evaluate the tool that I use everyday. It’s freely available to everyone at my site. See if it might work for you as well.