If we could somehow penetrate the bureaucracy and get access:
An e-ripoff of the U.S.
http://articles.latimes.com/2011/aug/21/opinion/la-oe-sparrow-medicare-fraud-20110821
To me the solution isn’t to do away with the electronic payments, but to invest in serious validation before payment. Given the half billion in money already lost (and I am sure it is lost, I doubt they could recover a nickle on a dollar) pay someone like us to put in some fraud detection and instead of immediately sending a check, wait 24 hours and do some analysis first. I am sure that it takes very little cleverness to deduce what ‘normal’ behavior is (i.e., people who are legitimate) and put in filters to look for whatever isn’t matching, then do some deeper analysis on that. My understanding is that in almost all cases the fraud is the minority (though sadly, minority might still be 20-30%) so with effective ‘legitimate’ filters one could substantially reduce the load on the servers. However, as mentioned in the article, there are powerful forces on either side that won’t want to employ anyone like us to actually quantitate fraud (even if we are reducing it (perhaps especially if we are reducing it)). When I worked at FINRA (the brokerage cops), my boss wanted to mine the data to look for patterns that indicated fraud rather than waiting for the fraud to be reported and then develop queries to find them. No one was interested in doing so and I often wonder if he ever made any impact. However, given the dramatic drop in cost for hardware it would seem fairly simple to develop a system that could basically be plugged into their network to analyze the data and give some sort of score on the level of fraud and then take a percentage of the amount of fraud detected. Then, we ‘work for free’ which should make the bureaucrats very happy, that is until they see the checks they would be writing.
Oh well, it was just a thought…
