From a
notice posted by Social Security on
FedBizOpps.gov:
The Social Security Administration (SSA) is seeking providers of software solutions for natural language processing in a medical context. ... In support of SSA's disability process, we are pursuing several initiatives in health IT [Information Technology] directed toward interoperable exchange of claimant medical records to speed receipt and subsequent case decisions. A significant element of these efforts is the application of intelligent analysis to the medical evidence received to assess any potential matches with policy guidance and facilitate subsequent manual review.
Good for SSA.
ReplyDeleteA well-programmed computer can cull electronic medical records faster and better than a human employee.
This would be great for listings that can be met on the numbers. It could really improve the speed and accuracy of processing initial claims...
...to the detriment of reps, of course. (There, I said it.)
It is to the detriment to the whole system. Won't someone have to ensure that the machine is paying the correct claims. The quick review will lead to quicker denials which will then race to OHA where there are still not enough people to provide due process hearings. So the point of this boondoggle is?
ReplyDeleteThat of course is a trick question because there doesn't have to be a point to a boondoggle as there "appears" to be a point. Can anyone say "one book"?
Second Anonymous:
ReplyDeleteThe software won't be making the final decisions - the intent is to analyze the medical records as they come in and flag potential allowance cases for closer "human" review earlier in the process then they might otherwise be detected. It would even be good if the system could detect and flag potential obvious denials as well.
1st Anon agrees with 3rd Anon. Machines wouldn't be paying cases. They'd be culling.
ReplyDeleteThat could quickly focus the "human" input and judgment on the areas that matter most. Humans are doing the same thing every day, with far worse speed and accuracy.
I'd rather have a computer making an initial run through the file for detectable flags. Better that than a burned-out, underpaid state employee at DDS.
Could also help assuage some of the alleged disparities (quality, accuracy, etc.) across the states.
ReplyDeleteFrom Anon 2,
ReplyDeleteThe problem is not faster review at the DDS level. DDS moves those cases pretty good because they don't have to really explain their decision. The problem is they are penalized for paying cases. DQB sends back cases and that counts against the DDS/employee. (There I said it and IT IS TRUE.) The one good thing I can see about a computer system review is that maybe the DDS's wont be afraid to pay some cases because the computer "flagged" it for them and this is their fall back position.
But no matter how you look at it the cases will be moving faster at DDS, racing to OHA and unless OHA is given more people to do the job correctly. The sum result is a negative (except, I admit, for those claimants who may get paid earlier in the process which is a good thing). It appears to me to be a perfect train wreck for OHA and that maybe what someone wants.
My point is computers can only do so much even with very sophisticated predictive models. Just look at the OPM/ALJ hiring morass. Hire more people, actually train them (not have them watch those stupid video conferencing programs)and really focus on quality as much as quanity because the quantity-only approach is not working.
Asserted that "Nothing is better than winning your million dollar case." AND "Finding a dollar in the street is better than nothing".
ReplyDeleteTHEREFORE if A is better than B and B is better than C - then it follows that A is better than C.
C = "Winning your million dollar case"
B = "Nothing"
A = "Finding a dollar in the street"
THEREFORE = Finding a dollar in the street is better than winning your million dollar case.
Of course this logic fails because of the semantic shift in the meaning of the word "nothing" .... and of course there have been landmark Supreme Court decisions that interpret words like "permanant" and "disability" in a completely different way - Such as Cleveland vs. Policy Management, etc.