Web Log


Beware of Buzzwords. Software Companies Are Selling You a Bill Of Goods

9/22/2016 3:35 PM

I was reading along yesterday when I ran across this from my buddy Anne Zieger over at EMR and HIPAA crowing about how Machine Learning might tame healthcare's Big Data.  First, there is no such thing as Big Data.  Depending on whom you talk to, the term big data was either coined by Dr. Francis Diebold at UPenn in a paper called A Personal Perspective on the Origin(s) and Development of \Big Data": The Phenomenon, the Term, and the Discipline or Roger Magoulas from Ultimately, big data is more about attitude than tools; data-driven organizations look at big data as a solution, not a problem.  Either way, Big Data is a sales tool to get you to buy that worthless piece of nothing called Hadoop.  Beware of Trojans bearing gifts and beware of people trying to sell you things.  Hadoop and Map Reduce are combined to give a programmer the ability to sift through huge amounts of unstructured data.  The problem with that is that they use Java to loop through this massive data store to look for things. If you've been following along with our blog, you already know why Java is particularly unsuited for this kind of work.  Plus, we already have a tool for that and we've had it since 1979.  Its called a database.  No, databases can't handle unstructured data, but there is no such thing as unstructured data, there is only data that hasn't been structured yet.  At Sentia, our solution is to structure the data, shove it into a database and run queries against it.  That requires no new technology and the first query will take about as long as doing it the dumb Hadoop way.  The second Hadoop query will take just as long as the first, because it works exactly the same way as the first, but the second database query will take thousands of times less, because the data is already structured.  If your database server can't handle the volume of data that your Hadoop servers can, format the boxes that Hadoop is installed on and add them to your database cluster.  That should solve the problem.

Second (did you forget there was a first already?) just like Big Data, there is no such thing as Machine Learning, at least not yet.  Margaret Rouse posted this definition on TechTarget, where she states that machine learning is used by Facebook "... to personalize each member's feed. If a member frequently stops scrolling in order to read or "like" a particular friend's posts, the News Feed will start to show more of that friend's activity earlier in the feed."  

We here at Sentia call that a query (yes, on a database) and have been doing it for decades.  Here's an example of how we do it.  In the Electronic Medical Records (EMR) application we wrote for our sister company Sentia Health, we query the database to find the top twenty diagnoses by doctor in a rolling six week average and rank them by number of occurrences.  This allows us to show the doctor a list of the most common things he or she is diagnosing that changes over time.  for example, in the fall and winter influenza (the flu) rises to the top of the list and when the doctor selects it, he or she can duplicate the first instance of that diagnosis and copy it over to the new patient, literally completing the patient documentation with one click.  We wrote that query in 2002, before there was "Big Data," Hadoop or "Machine Learning."  Heck, that was before there even was a Facebook and long before it was open to everyone and FAR before they started using "Machine Learning" to update your newsfeed.

Why do we mention all this to you?  We told you above: Beware of Trojans bearing gifts.  The people who came up with all this are trying to sell you something.  The reporters, in this case bloggers, don't have the faintest idea what they are talking about.  Their job is to publish or perish, not to tell it to you straight.  While Walter Cronkite didn't have to know what he was talking about to report the news, he just told you what happened and he tried not to put a spin on it.  That isn't the case today.  The sales people, and everyone hates a sales person, don't care what you think or what the product is as long as they get your cash.  They are paid to come up with new verbiage to part a fool from his money.  If you are falling for this, yes, you are a fool.  If you don't have the tools necessary to winnow the wheat from the chaff in your software purchases, you are an idiot and you deserve what you get.  that's fine as long as you are happy with it, but this affects everyone.  It is most of the reason that healthcare is claiming 20% of the Gross Domestic Product in the United States.  Idiot administrators are buying buzzwords from dishonest salespeople.

Here's another example: Athena Health claims they are a cloud based EMR.  Athena is practice management and uses VPN to log in to a desktop remotely.  Yes, you log into a windows machine with an installed, windows based application on it.  That means every one of their clients has its own installation of their product.  Talk about an expensive administration nightmare.  That doesn't even count setting up all those machines with the application installed on them.  "Cloud" is their buzzword.  They want you to think their product is SaaS (Software as a Service) like our sister company Sentia Health provides, but the can't because they don't know how to build applications that run in a browser.
We've talked about buzzwords before.  Buzzwords are designed to obfuscate facts and to put lipstick on the pig that a sales person is trying to foist off on you.  Yes, YOU, dear reader.  If you go to a doctor (or any other business) that buys into this, and all of you do, you end up paying for that polished porcine piece of prevarication out of your own personal pocket.

If you want real software, built without using mind numbing buzzwords, you know who to call.  Don't make out with the red lipped porker, or buy a pig in a poke.

Date Written Comment By Comment