How to lie with statistics – Iowa Medicaid bids

Today, The Des Moines Sunday Register reported that the bids of the insurance companies that were chosen to manage Iowa’s Medicaid program contained misleading and unverifiable data.  (See link below.)
One example you cited was from UnitedHealthcare’s bid, where they stated that their Kansas affiliate reduced premature births by 23 percent from 2013 to 2014.  Although it is possible that the efforts by UnitedHealthcare caused this result, it is also possible that the result were caused by random fluctuation.
If you measure 100 health indicators, and the results follow a statistically normal distribution, then a few of the indicators will show significant improvement and a few will show significant worsening, even though there is no real change in long term trends.
It’s just like if you asked 100 investors to each pick a group of stocks by throwing darts at a stock listing.  A few would make amazing gains and a few would suffer tremendous losses – none of which could be attributed to the ability of the investor.

In the bids by the insurance companies, all they had to do was “cherry pick” the best statistics after the fact to make themselves look good.  So, while the statistics my be true, they are clearly misleading.


Please Comment Here

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s