Well this appears to have been a straight linear calculation that went south and even it if were to be done in a linear basis, oh my gosh, trying to find relationships with such varied data would require some real stretching of a formulas for sure.  Everyone was looking for that “magical algorithm” to allow this predictive analytics calculating to take place and it has a fail.  You can’t relate everything in data.  We have folks still chomping over at the Heritage Group sponsoring that $3 million carrot and it looks more like a “gaming site” and I though the project was nuts to begin with but former HHS director Leavitt got sucked in to this one too thinking the magical algorithms would some how just nail all this information.  I wrote back in August of 2011 about it really not going anywhere other than a few code modules that might be produced of some use. 

So What’s Going on Over at the Heritage Health Prize Competition to Predict Who Ends Up Being Re-Admitted to the Hospital–A Lot of Mathematical Algorithms For One…

There have been others like the WellPoint publicity stunt for the $10 million dollar prize for solving healthcare and here we are still not solved at anyone who works with data knew that was a farce, even back then.  With the way the formula is set up and in a research letter published by the Journal of the AMA, the teaching hospitals and safety nets are the ones at big risk who need the money are the ones to be hit, so what are we doing here? It was also quite entertaining to see HHS put on that Facebook contest which went nowhere but when you have the markets on Wall Street putting out press on the value of the stock to drive it up, it rolls over into other places with inflated value which we all know after the IPO, showed its face along with rogue algorithms that cost investors some money.  Folks that are not into the mechanics of data just kind of grab on to this stuff in the media not knowing if is matters or not. 

To quote Charlie Siefe in one of his lectures you get something along this line, “well the formula has a square root in it, and the story has been all over the news in the last few weeks, so it must be good”…yup seen that a lot and it still happens out there.  You can watch that video on the left hand side of my blog in the education area and he will tell you about real math and formulas and how it’s marketed and skewed.  My favorite one he mentions is the clinical study about the fish and neurological abstract with measuring the brain in the fish with responses…the fish was dead (grin). 

Hospital leaders have called for a redo of this and I don’t blame them and I wonder myself if they will even be able to come up with one unless they do one that is personalized for each hospital with goals as that’s the only way it would come close.  I spent many years in sales and I didn’t have the same quotas as my co-workers, shoot mine were always very high but I was a bit of an over achiever too.  Here of late too I have also called the ratings sites bunk on giving grades to hospitals, same reason applies with not comparing apples to oranges. 

How can UCLA that does procedures that maybe only 2-3 hospitals do in the country and see some of the sickest patients get a “C” and frankly I don’t blame the Harvard hospitals for disputing those ratings either as some of them are in the same boat.  As things in technology change some things don’t work and go by the way side and those are one that nobody would miss.  Sure keeping a listing like a Yellow Pages is fine and allow comments but cut out the high school grading and get more accurate information out there with fewer dead doctors for one. 

Physician Rating Sites Are Seeing Fewer Reviews These Days–Flawed Data Catching Up Along With Loss of Value As Consumers and Doctors Lose Interest?

Well time will tell on what happens here with this flawed system as it’s not going to make any friends and when anyone gets judged on flawed analytics that are not accurate, nobody will be happy.  We just need some hybrids in running some executive departments today, some with a little computer science or IT backgrounds as attorneys running everything is going to sink us and you can see where we are today as they duped, and see that same square root there making it look good:)  We really need executives who can walk and chew a little code or IT as to have to rely on staff for everything and not be able to make sense out of data mechanics is like working with one hand tied behind your back. 

On the same topic, I have had analysts ask me questions about code.  They are the experts on doing their spreadsheets, not me,  and drill down and use the software that developers and code makers create.  When I wrote my medical records system, the doctor knew it better than me as far as using it and that’s the way it goes as I was writing code and not seeing patients:)  Point made here is that having some kind of hybrid in these executive jobs would really be nice and save time and offer a new look at leadership with not ignoring technology’s impact and that means more than just having an executive who can use an IPad.  The people who use the software are better at using it than most developers by a long shot. 

If you feel you have been duped or want to learn more about it, watch the 5 videos on the left and it will give you a new point of view here on how this really works, it’s all code and IT infrastructure running on servers 24/7 handling the the results, and humans put in the parameters.  Here’s the same videos as featured on my Algo Duped page.   One more comment here and a link is to a post and video about the value of data.  This is a good video on how companies represented like Ford, NASA are all in a forum talking about big data and how to work with it and their hopes to find the “miracle data scientist) which is an Oxymoron as this is not splitting atoms, to help them find the answer.  

When people guess and the need for greed with flat out creating analytics one can pitch and sell, it gets out of context and this is the biggest issue facing consumers today.  The gal from T-Mobile in the video is not proud in when she comments on their analytic and says some of what we are doing is “silly”.  So her issue is getting rid of silly and you will hear them talk about company quants the folks who create the mathematical business models) and about half way through it gets rolling. We can’t use nonsense queries and models created to make money (we learned this from the banks) only  that will end up hurting consumers and too much is already happening and we come back to the formula with the square root that dupes everyone to think it has value when it does not.  

CMS is not trying to make money but they get duped too on what works and what does not with math and formulas.   All formulas are not created equal and if you watch the Quant video documentary at the left some former quants explain it to you and how you get sucked in.

Big Data/Analytics If Used Out of Context and Without True Values Stand To Be A Huge Discriminatory Practice Against Consumers–More Honest Data Scientists Needed to Formulate Accuracy/Value To Keep Algo Duping For Profit Out of the Game

In April, doctors Karen Joynt and Ashish Jha wrote in the New England Journal of Medicine that the federal policy of penalizing hospitals where patients are more often rehospitalized within 30 days was misguided, unfairly targeting those in poor areas with large minority populations.

In a research letter published Tuesday by the Journal of the American Medical Association, the researchers, who are from the Harvard School of Public Health, added some heft to their argument.

They looked at penalties assessed against 2,189 hospitals and found that the largest hospitals and those that are teaching hospitals or safety net hospitals, meaning they care for the highest portion of poor patients, were most likely to have their payments docked.

The likelihood of a person returning to the hospital after a heart attack depends a lot on housing, financial stability, and other socioeconomic factors, she said.



Post a Comment

Google Analytics Alternative