A.I., Health Care, Google, and Elon Musk

“Now, if you’ll excuse me, I’m going to go home and have a heart attack.” – Pulp Fiction

I wanted to get a doctor appointment to treat my invisibility, but he said he couldn’t see me right now.

A computer can predict who will die using medical data better than a doctor.  As of today, like science has no answer as to how California copes with the landfill requirements of Kardashian body hair, science has no understanding of how the computer is doing it.

A gentleman by the name of Dr. Brian Formwalt led a study where approximately 1,770,000 electrocardiogram records were fed into a computer.  An electrocardiogram is also known as an ECG, for obvious reasons.  For less obvious reasons, it’s also known as an EKG.  EKG stands for elektrokardiographie, which is exactly the same thing as an electrocardiogram, but in German.  If your doctor calls it an EKG, he just might be thinking about expanding his living room.

Always be careful when Germans research expanding anything.

But back to the study.  So, there were 1,770,000 records, but only 400,000 people in the study, so the average person had more than four records.  Obviously, these weren’t all healthy people, since I have had (I believe) exactly one ECG in my life, and it was for a pre-employment physical as an astronaut for Wal-Mart®.  At least the recruiter told me Wal-Mart© needed astronauts, before Wal-Mart™ cancelled the program when China accidentally delivered 50,000 small space shuttle toys rather than one life-size one.  I guess that’s what happens when you buy space shuttles by the pound.

But what is an ECG?

Electrocardiograms are the little machine light that makes the beep sound every time your heart beats.  The beat is measured by injecting elves into your body that send radio signals to the machine every time that your heart muscle squeezes them.  Okay, the technical side might be a bit off, but it doesn’t really matter if you or I know exactly how the machine gets the data.  It’s just the device that goes beep-beep-beep-beeeeeeeeeeeeeeeeeep to let you know that John Wick’s® dog died.

Cardiac surgeons are the guys you want to see for a change of heart.

Okay, so now you know everything that you might need to know about technology invented in 1895.  But it now produces an electronic file rather than the old method, where the heart rhythm was tattooed on the backs of ill-tempered Chihuahuas.  The 1,770,000 records were then fed into a computer that had been previously taught to read ECGs.  The simple question was asked – which of these patients will be dead in a year?  I mean it used to make me feel better when my doctor told me, “that’s normal for your age,” but then I realized that at some point being dead will be normal for my age.

Since all of the records were over a year old, it was known which of the patients were alive and which were dead.  Essentially, the doctors were (with very little data) asking the computer to predict the future.  It did.  And it did it better than human doctors.  Some of the ECGs looked absolutely fine to human doctors – they detected no abnormality, yet the computer was able to see something that accurately allowed it to predict the death of the patient.

Then the next doctor told me it looked like I was pregnant.  I said, “But I’m a guy.”  He replied, “But it looks like you’re pregnant.”

It doesn’t surprise me.  Computers are powerful tools that are great at taking lots of data and being able to compare it quickly.  The reason that they can do this is they:

• Have 100% focus, and if they get a sore throat you can give them Robo-tussin®.
• Don’t need to make payments on second wife’s Mercedes® and third wife’s Lexusâ„¢.
• Can retain every previous ECG reading ever seen and instantly recall the pattern if needed, much like I can retain the plot of every one of the episodes of Gilligan’s Island.
• Don’t get distracted by how healthy a patient looks or how much kale he eats.

These are great advantages.  In the future, machines will be able to do things where we may never understand how they made a correlation, or, as in this case, even what the correlation is.  Arthur C. Clarke Third Law states that “any sufficiently advanced technology is indistinguishable from magic”, and he’s right.  Health care generates amazing amounts of data, and also outcomes.  It’s only a matter of time until some big corporation gets evil . . .

Oh, yeah, Google®.  It bought Fitbit®.  Now it knows what you’re searching for, and it also has a treasure-trove of heartbeat and fitness data.

Google® is female.  It won’t let me finish a sentence without giving suggestions.

Well, I guess that’s kind of scary.  But at least Google© doesn’t have access to medical records.  Oh, Google™ has patient names, diagnoses, prescription data, and records from 2,600 hospitals.  Millions, perhaps tens of millions of patients?  In (possibly) all of these states:  Minnesota, Wisconsin, Michigan, New York, Rhode Island, Maryland, D.C., Illinois, Indiana, Kansas, Missouri, Kentucky, Tennessee, Virginia, Oklahoma, Arkansas, Tennessee, Texas, Mississippi, Louisiana, Alabama, Georgia, and Florida?

Nah, that should work out fine.  There isn’t a record of Google® ruthlessly monetizing every corner of the Internet not already inhabited by Facebook™, Amazon® and Microsoft©.

I think the case is clear for someone to go through this data.  With only a few records and outcomes fed into it, a computer is better at predicting medical outcomes than a very good doctor.  If all of the data could be available?  I think we’d have a legitimate revolution in health care.

Frankly, if we don’t descend into civil chaos, I think that this health care revolution is certain.

But Google®?  Google™ has proven itself untrustworthy.

I’d suggest that we give control of the initiative to a leader that’s more trustworthy than Google®, like Bernie Madoff, but he seems to be otherwise, um, detained.  And I’m sure that Jeffery Epstein has better morals, but, um, he seems to have accepted a unique opportunity with the Clinton Foundation.

Heck, let’s give the job to Elon Musk.