Occasionally I step over into a bit of the financial side of things and the reason for this post is that back in August of 2009 I made a post about the US needing a Department of Algorithms…hmmmm…not far off now is it, call it what you like but audit functions for code should be upon us about now, we need it. You can read the post at the link below which I created due to the Madoff situation and the Harry Markopolos testimonies.
Tuesday, August 25, 2009
“In healthcare we have an agency that every likes to complain about called CCHIT that certified the safety, accuracy, and interoperability of healthcare records, just wonder why this process might not be explored in other areas where transactional data exchanges are taking place. We want our medical records to be accurate and concise, so why not the same with our investing? This is just a question I pose.”
Also we have tons of algorithms running like crazy in healthcare with insurers and other financial institutions that could also use the same auditing. We go to great lengths to certify the software and algorithms contained within for medical records, but do we hold any of the accounting algorithms of the insurers accountable…no. This has gone on a long time and not the first time I have suggested such a process here. Algorithms should be written to give accurate results and many are written for desired results which are not always on in the same with the word accurate.
FINRA reports to the SEC and focuses requests on flawed codes to understand how they are constructed. This has moved up a notch from the original requests as more information is now being requested. FINRA has also beefed up their staff with more individuals WHO UNDERSTAND CODE. I still think back to the Goldman case and I have often wondered how the entire case went down with having a full understanding of what actually happened.
The former coder stated that he accidentally moved some code, and I have done that, but again I’m not writing anything as sophisticated as stock market investing software, so that part could have been true and again I wrote my opinion up when all of this was current, and stated you do need to understand the “value” of the code so as not to incriminate someone using a partial open source module for example. We all do that and build additional code on top of perhaps several open source models put together and have them fully function to fit the request of the software. Also, a couple of other relative posts from the Medical Quack in 2009 below:
This is really a good time to start asking questions on CODE as more personal data and profile information is sold on the internet and while it may be good for gathering statistics, we have those who narrow some of these queries down to an individual level and you end up with flawed data and algorithms that are sold beyond what they can really deliver in the way of business intelligence. In healthcare I am always happy to see HP or Microsoft provide their services as they do not have a stake with another subsidiary who can profit with algorithm designs of a subsidiary when it comes to reimbursement or revenue cycling.
Again, I stand on my suggestion for a Department of Algorithms that I wrote about 2 years ago all the way around for financials and healthcare, as we want accurate and honest code that will not fleece consumers as we have had our fill of that and the economy and the redistribution of the money speaks for itself. Call it whatever you want but get some auditing done with some REAL CODERS who can dig in and find both the flaws and the accuracies.
In the healthcare/insurance side of regulation we have actually had one state legislature begin to question the same thing with insurers and ask for their CODE. Back in October of 2010 I wrote about their inquiry here:
New Jersey Legislature Getting Smart– Bill to Modify Claim Procedures to Include Asking For Insurance Company Algorithms-Bill A3334
If you are lost in the math in this article, pick up a copy of this book, “Proofiness, the Dark Side of Mathematical Deception”…more at the link below and that will explain enough to perhaps bring a basic understanding of how math and algorithms if not done accurately can be far from reality at times.
“Proofiness–The Dark Side of Mathematical Deception”–Created by Those Algorithms–New Book Coming Out Soon
One other bit of back information, some insurers get sued over their algorithms too.
We all remember the AMA legal case against United which was 15 years of paying lower than customary fees for doctors and patients who required out of network services and lawsuits are still going on that one as so many of the insurers subscribed to the data base used by the subsidiary of United. There are outstanding cases against Blue Shield, Aetna and more who licensed use of that data base. It was just about a two years ago that HealthNet finally agreed to pay claims at a level of 14% above what the data base called for. It all comes back around to those algorithms.
So after reading this far, again it seems fair to link the algorithm audit processes here to both the financial and healthcare sector as they are all transactions where someone gets paid and we want accurate and correct formulas used. One more mis match that has bothered me of late is the credit agency FICO claiming they can sell analytics with free information from the web combined with a consumer credit score to sell this to insurers and pharmaceutical companies as predictive analytics to determine by a score if a patient will take their meds. You don’t have be a coder to figure out that while this might make a good report for overall statistics if you want to query the information for relevance, but to take it down to a one on one bases with flawed data, it’s wrong and it’s mis matched false intelligence algos being sold. Don’t be fleeced on this one, no matter how good the marketing looks. My write up in FICO was carried over by a reader to the Daily Kos.
So while FINRA is opening the door to audit code, we could certainly use some of that in healthcare transaction audits as well. We had the case last year with Blue Cross running the “breast cancer” algorithm that was reported in Reuters that had to be corrected.
And one one last closing note here, we have so much of our personal data being sold and million and billions made from all of this, and there’s no way you can stop the algorithms so next step, license those folks and tax them, healthcare could use a bit of this money coming back to help doctors and hospitals on the brink of bankruptcy with reduced contracts signed with insurers created by algorithmic business models. Big business uses them all the time and laws just can’t keep up.
Privacy Wanted–So Let’s Require Those Who Sell Web Data to Register and Tax the Transactions and Publicly Disclose Who They Sell To With a Federal Registry
There’s nothing wrong with using algorithms and software today for analytics, and we need and depend on them to run so many areas of our lives and business, but if they are not accurate and are only created for “desired” results where monetarily someone has to lose, well it’s time for that Department of Algorithms to certify that they are in fact fair and creating fair game for all.
Further more we need lawmakers to also make the effort to use technologies that are up to date and allow for better decision making to stop some of the intellectual fleecing that we have seen for the last number of years. We have lawmakers who want to put data out for public use that has tons of flaws and errors and have no clue on the cost. HHS and enforcement agencies by all means need access but again with what is proposed here and the ROI on the expense of this is just not there. We have for profit MD and hospital rating services all over the web with flawed data to include dead doctors and other information thieves can mine for fraud and correcting that information certain supersedes this project by all means.
Digital Illiteracy Is Killing Us With New Bill Wanting MD Medicare Compensation Put Online-No Clue on the Cost, Time and Truckloads of Errors to Audit and Correct
So let’s beef up our IT infrastructure and get lawmakers and the audit enforcement agencies the intelligence they need to investigate and ensure that the “secret sauce” algorithms used today are all on the up and up. If lawmakers used better technology to study and query, we would also have better laws for the agencies to work with from the beginning, it’s a technology team effort here. BD
IBM Watson Capabilities Being Pitched to Financial Industry-Congress Must Not Have Felt They Needed This So Further Behind We Fall With Effective Intelligent Lawmaking
(Reuters) - U.S. securities regulators have taken the unprecedented step of asking high-frequency trading firms to hand over the details of their trading strategies, and in some cases, their secret computer codes.
The requests for proprietary code and algorithm parameters by the Financial Industry Regulatory Authority (FINRA), a Wall Street brokerage regulator, are part of investigations into suspicious market activity, said Tom Gira, executive vice president of FINRA's market regulation unit.
"It's not a fishing expedition or educational exercise. It's because there's something that's troubling us in the marketplace," he said in an interview.
Most of the algo-related requests, he said, have been made to hedge funds that use quantitative trading strategies.