Way cool way useless
We truly believe when it comes to diabetes research the motto isn’t is this needed to be done rather hey let’s do this because it can be done. This is particularly true with all the new way cool whiz bang technology that is flooding diabetes. Here is a perfect example of what we are talking about per a post on the VentureBeat web site;
“Researchers have developed an artificially intelligent (AI) system that can predict future blood glucose levels before hyperglycemia (very high blood sugar) or hypoglycemia (very low blood sugar) occur.
In a paper published on the preprint server Arxiv.org (“Predicting Blood Glucose with an LSTM and Bi-LSTM Based Deep Neural Network“), they describe a recurrent neural network with a bi-directional long short-term memory (LSTM) that’s capable of learning long-term dependencies. Basically, the memory cells in LSTMs allow the neural network to combine its memory and inputs, improving prediction accuracy. And because it’s bidirectional, it can access context from both past and future directions, expediting training time.”
Now we must admit that sounds so damn cool. Yet like so many things that are way cool and whiz bang it doesn’t mean a whole hell of a lot if the patient doesn’t act on this information or can’t act on it. And like so many pieces of useless information even if this way cool whiz-bang neural network (boy that does sound so damn cool) works it’s not needed. Never mind that insulin dosing algorithms do the exact same thing which when combined with an insulin pump allows for the system to act or when combined with a smart insulin pen allows the patient to act. Nope let’s give the patient information that may or may not happen which they may or may not be able to act upon.
But let’s not stop there OneDrop is making a big deal out of the fact they are the only conventional glucose meter to send readings to the way cool whiz bang Apple Watch. Never mind that the Dexcom CGM does this already providing more and better data, nope the folks at OneDrop see this as revolutionary. Which isn’t surprising as OneDrop thinks everything they are doing is revolutionary.
Folks we hate to break the news to all these way cool whiz-bang high-tech companies, but diabetes management is not rocket science. The reason we don’t see better overall outcomes is not because there aren’t enough toys in the toy chest. The reason isn’t because the drugs available don’t work. The reason is patients are human and not computers.
Let’s say this way cool neural network and tells the patient a hypoglycemic event is coming -which isn’t unique by the way – what does a patient do? They get some sugar into their body which will likely cause the neural network to predict – hey guess what now a hyperglycemic event is coming – which unless the patient is on insulin they can’t do much about. And what happens if this prediction does not come true, what happens if the neural network is wrong? The last time we looked there is a huge difference between a “prediction” and a “fact”.
All the polls predicted that Hillary Clinton would be sitting in the White House today.
Listen we understand that everyone loves the toys in the toy chest. That artificial intelligence sounds so damn cool. We also understand why these mad scientists just love to tinker. But seriously folks this isn’t that complicated this is not about more or better information, this isn’t about more or better toys, this is all about transforming the how to into the want to. Patients are not computers who can be programmed, they are humans with a chronic disease they don’t understand, didn’t want in the first place and requires work to manage properly.
That’s the God’s honest truth but this being the wacky world of diabetes we can have none of that.
