If you think about it, the life of a Type 1 diabetic (T1D) abounds with data. For example, it is recommended that most Type 1 diabetics test their blood glucose (BG) at least 5-6 times a day. So that’s 5-6 data points each day.
I’ve been a T1D for over thirty-three years. I can’t say that I’ve always done what was recommended, but, if I averaged 3 finger sticks per day, that would be 36,135 total drops of blood (and 36,135 test strips). Ouch! But more importantly, there are more than 36,135 data points with those 36,135 drops of blood and those 36,135 measurements of blood glucose.
There are a lot of other factors involved in managing your diabetes on a day to day basis. Basal rate, insulin sensitivity, carbohydrate to insulin ratio, and carbohydrate intake are just a few of the factors that can be specified and/or measured. And there’s a plethora of other factors that are less measurable – stress, hormones, activity, illness, sleep etc.
You can imagine the difficulty of trying to correlate all of these dimensions, analyze them, and then make informed decisions based on them. My head spins sometimes just thinking about it (even though I have to do this every day). Here are some examples of the data I am collecting and sharing with my endocrinologist in between visits.
Even with all of this data at my fingertips (pun intended), managing my diabetes is not easy and it is far from predictable. I could do the exact same thing, eat the exact same thing at the very same time and treat it with the exact same amount of insulin – and I can just about guarantee you that I will not get the same results. There are times when my blood glucose levels truly surprise me, especially the very lows – and those are the dangerous ones.
IBM Watson and diabetes
IBM is partnering with medical device leader Medtronic on a predictive diabetes management solution. In a pilot to see how Watson can help people with diabetes, Medtronic and IBM took 600 past patient cases and applied cognitive analytics to the data from Medtronic insulin pumps and glucose monitors. Watson was able to predict hypoglycemia – extreme low blood sugar – up to three hours in advance of onset — early enough so a person could take action to prevent a potentially dangerous health event.
WOW – three hours in advance so that I could do something about it. Wouldn’t this be sweet (pun intended again)?
IBM and Medtronic unveiled this partnership at CES 2016 earlier this year:
This is one of the most exciting developments in diabetes management that I can remember. (In full disclosure, I use a Medtronic insulin pump).
When I moved from 6 shots a day to my pump, I kept asking myself – why didn’t I do this sooner? And when I added a continuous glucose monitor to the mix, I had a similar revelation. All of this technology is amazing, don’t get me wrong. But a lot of the data correlation is still pretty manual (on my part and on my doctor’s part), and there is surprisingly little insight gleaned from this data (and there is so much data).
I can picture the dialog now (although note that this is only my vision of how this solution would work and not a depiction of the eventual solution):
Me:“Watson, I’m going to mow the lawn.”
Watson: (understanding how I’ve reacted to this type of strenuous exercise in the South Florida heat, what my current blood glucose and active insulin levels are, what I’ve eaten and probably so much more).
“That’s great, Kimberlee. I’ve set a temporary basal for the next hour and a half – and I really think you should eat 30 grams of carbohydrates before you get started. And don’t forget to hydrate.”
Me: “You got it.”
I, of course, would happily do what Watson recommended because I want to keep my blood glucose levels in range and get on with my life in as “normal” a fashion as I can.
And let me tell you: Such insight and advice would truly transform this Type 1 diabetics life. Talk to me below on how you think Watson could change the way you manage your T1D.
Several weeks ago, I wrote a blog in which I discussed the history of condition monitoring and the differences between condition monitoring and predictive maintenance. You’ll recall that we left off introducing the next evolution of maintenance: condition monitoring 4.0. The most current advancements in predictive maintenance require automation of the analysis process using AI models. ...read more
Did you know that 178 work-related accidents happen every 15 seconds? It’s a staggering number with potentially devastating effects on employees in asset-intensive industries. Businesses pay a high price for accidents, too. In the United States, estimated annual costs associated with occupational injuries and illnesses total USD $170 billion. Despite extensive regulation and strong commitment to ...read more
People in charge of managing the performance of physical assets in industries such as energy and utilities, mining, and real estate are now reducing overall maintenance costs by 5-10 percent. They’re also increasing equipment uptime and availability by 10-20 percent. If results like these don’t make their competitors nervous, I don’t know what will. These ...read more