Health knowledge made personal
Join this community!
› Share page:
Go
Search posts:

Medicare Low Safety Rankings At Harvard Teaching Hospitals Disputed and Flawed–Attack of the Killer Algorithms Chapter 16

Posted Feb 14 2012 1:02am

Yes it seems I am on a tear here and again we come back to business intelligence algorithms being at the heart of the issue.  Personally I think the rankings of doctors and hospitals in it’s current fashion is somewhat just there and not something I would put a whole lot of faith in if I were a consumer going down column by column to make a decision.  If I were having surgery, the overall is what makes sense and these are Harvard teaching hospitals in question here.

We had the same thing when they tried to rate doctors out here in California as when their patient mortality rates came out, those doctors who saw some of the most destitute patients with extraordinary needs and not doing well at all came out looking like they had a good chance with killing you, when in fact they were taking care of some of most ill patients around, so demographics messed that one up.  I do wonder how they calculate all of this and anyone can crunch numbers but are they relative?  Sure there are a lot of areas for measurement but not all are equal.  The mark for the expenses for ICD10 was way off and it is costing millions more than the government estimated. 

I do kind of wonder in the Health IT areas what goes on when I see HHS and their Facebook contest so all could communicate during a disaster, and I would riots would not be in the picture there. 

HHS Seems To Be Confused on Social Networks–Facebook for Disaster Support Contest, Give Me a Break Use Twitter Like Everyone Else Does

As a side note don’t look for HealthGrades to do any better here either, just different algorithms crunched in a different set of formulas and besides a couple years ago I found my former doctor who had been dead for 8 years still alive and well and seeing new patients and honoring Healthnet. 

I dug around in the data and actually that experience is what set me off on talking about “flawed data” as it was right in front of me.  I talked with doctors who were listed on staff at facilities where they had set foot, found one guy listed as a doctor who wasn’t one and so on.  A few months ago they were working on their algorithms too just before getting purchased by a PR firm so now they will be marketed better and I hope their data gets better and not just spun marketing from their end of the field.

The problem here is that payment is based off of these numbers and we do have folks that think an algorithm is everything and it kind of reminds me of when I went to school and I was so mad when they tried “grading on the curve” as I wanted my real grades and this seems to be what we have here in essence.  Listen to the radio broadcast below and let Professor Siefe from NYU educate you on how naïve we are and gullible with stats and numbers, and sadly it looks like the government here is sucked in here big time.  There are places where analytics do work but with current economic conditions folks are spinning those algos and create even more algos to substantiate the first ones they wrote. 

“Numbers Don’t Lie, But People Do”–Radio Interview from Charles Siefe–Journalists Take Note, He Addresses How Marketing And Bogus Statistics Are Sources of Problems That Mislead the Public & Government

We just don’t have enough trained people around that know how to work with analytics and mostly of all “flawed data” so this is the Attack of the Killer Algorithms” Chapter 16.   Below are a couple other Killer Algorithm posts and a summary of earlier chapters can be read here… Killer Algorithms Chapters 1 –12


Insurers to Provide User Friendly Summaries of What’s Covered–Good Luck As Constant Revisions Will Be Required as Business Intelligence Algorithms For Profit Change–Attack of the Killer Algorithms Chapter 14
Attack of the Killer Algorithms Part 13–Bank of America Style - Flawed Data With Credit Scoring Agencies–Dead Man Banking And Currently A Consumer Under Attack

Medicare’s first public effort to identify hospitals with patient safety problems has pinpointed many prestigious teaching hospitals in Boston and around the nation, raising concerns about quality at these places but also bolstering objections that the government’s measurements are skewed.

Massachusetts General Hospital and Beth Israel Deaconess Medical Center, both affiliated with Harvard Medical School; and Boston Medical Center, affiliated with Boston University, were among those having substantially more complications than the average hospital, according to data evaluated by the Medicare program.

But leaders of a number of the nation’s major teaching hospitals are questioning the accuracy and fairness of the data, saying they do not properly account for how severely ill their patients are. The numbers were compiled as part of a series of efforts by the government to judge - and ultimately pay - hospitals on the quality of their care.

The data show high rates of serious complications for elderly patients at about three out of 10 major teaching hospitals, including some of the biggest institutions in Boston, Philadelphia, Los Angeles, Cleveland, and Chicago. Overall, teaching hospitals were about 10 times as likely as other hospitals to have high complication rates, according to a Kaiser Health News analysis.

Dr. Evan Benjamin,cq Baystate’s senior vice president for quality, said “the use of the data to compare one institution to another is flawed.”


Post a comment
Write a comment: