Is More Data Always Better?

I call Jade my little girl, but the truth is that at 10 years old, she’s almost as tall as me, and I think that she’d probably win a fight.  Anyway, last year, Jade was diagnosed with type 1 Diabetes, which means that she has to take four insulin injections each day to substitute for the fact that her body no longer produces any.

Diabetes can be life threatening as the body fails to control sugar levels effectively over time, but provided that you manage the condition well, the risks are minimised, and for most people, aside from inconvenience, there is little impact on quality of life.

There are to ways to manage diabetes, the simple way, in which you take a standard dose of insulin at regular periods throughout the day, and eat appropriately, or the more complex way in which you set the insulin dosage in line with activity and the calorie count to maintain a stable level of blood sugar.

Because Jade is only 10, she manages her condition by taking a standard dose ahead of meals, and then eating normally.  She measures her blood sugar 4 or more times a day, and (because I’m a geek), i put these into Excel, and built some graphs.  This is what hers looks like over 4 randomly chosen days:

Insulin and Blood Sugar Levels

What’s immediately clear about the graph is that there is a lot of variance in the blood sugar levels – this is not unexpected, as diabetes is an inability of the body to control blood sugar.  It seems to me though, that this is a poor management system, and it would be better to use variable doses of insulin and maintain the blood sugar levels within tighter bounds.

The problem is, that a dynamic management system is inherently more difficult to achieve, as it requires more data – you need to understand more variables, however there is a major benefit in using a dynamic management system, and that is that you control the condition fully and can live an entirely unrestrained life, rather than living to the limitations of the management system.

So, how would you set up a dynamic system to manage the condition?

It is absolutely essential that you measure everything over a period of time.  You need to understand personal responses to specific conditions, and be able to understand the exact modes of cause and effect that your body has.

The key indicator is blood sugar level, as that is what you are managing, but it is also important to recognise the factors both external and internal that impose on the system.

Things you need to measure:

  • Blood sugar
  • Calories obtained from food
  • Calories spent through exercise

Ideally, you would want to measure the blood sugar levels both before and after exercise, and also before and after food, but it would also be advisable to measure them more frequently so that you can see what kind of responses your body has to food – how long it takes for the food to be digested and enter the blood stream, and how quickly sugar levels fall for a given dose of insulin compared to a given calorific intake.

When you have data about a system, you can start to draw conclusions and use those conclusions to predict future behaviour.  For example, if you know that it takes a dose of 5 insulin to reduce a blood sugar level from 17 to 6, and you know that a meat pie will move your blood sugar level from 6 to 17, you know that you require a dose of 5 to control the meal.

Likewise, if you know that a 30 minute run will reduce require 350 calories and it takes 15 minutes for your digestive system to release calories from food, you will know that you need to eat a snack 15-20 minutes before you run, and tak an appropriate dose of insulin to manage the take up of blood sugar when you are running.

When you are able to measure data accurately, it is much simpler to understand underlying trends within the system and make forecasts about future behaviour.  You control the blood sugar levels with insulin doses based on comprehensive knowledge about what your calorific requirements are to sustain a known level of activity, and a acquired data about the levels of insulin that your body requires to control the sugar.

The point?

Well, the point is that with any complex system, whether blood sugar levels, or website traffic, the more data you have about the way in which the system behaves, the more you can do to control it.

Charity Plug

Science costs money, and Diabetes research is expensive.  If you’re planning to make any donations to charity, or are looking for someone to be the recipient of sponsorship money, I’d recommend Diabetes UK.