At yesterday's House Social Security Subcommittee hearing all of the members were struggling under the illusion that agency management could seriously address agency backlogs. I have my own ideas for what the agency could do but I'm not under the illusion that they would seriously address backlogs. More money is the only thing that will work; "brute force" as former Commissioner Astrue described it when the agency was in a backlog situation not as serious as the current one.
Oct 27, 2023
Brute Force Needed
Subscribe to:
Post Comments (Atom)
11 comments:
Backlogs are going to get worse beginning November 17 when the MAGA GOP in the House shuts down the government again.
A consistent thing I've noticed over the past several years of watching SSA executives testify before Congress is that these executives seem severely under-prepared to testify before Congress. Many times they have appeared to be caught like a deer in headlights. They are flat-footed. They do not know answers. They are not empowered to frankly admit to Congress that years and years of no confirmed COSS has hamstrung the Agency. Instead, we get mealy mouthed assurances that they have a plan or are looking into it or are committed to something. I still remember Michael Astrue boldly going to Congress to describe how the backlog can be reduced. I remember Carolyn Colvin when she was the DCOSS appearing before Congress to deliver the message that "we will do less with less." Oh for those glory days of leadership.
SSA needs a budget that will fund the IT enhancements and humans that are needed to serve the public. And SSA needs a confirmed commissioner now!
Surely the answer isn't just to throw money at the problem. There's a long list of things the Agency could do more effectively to perform better.
The local office here does pretty well except they do too much on the phone and some clients don't maintain good phone service. The problems or backlogs are with the state Disability Determination Services.
@1:00
No, the answer is not just to throw money at the problem, the answer is to throw money at the problem and require that the money be spent on things (i.e. technology, new hires, etc.) that would actually help the backlog and not require it be spent on things that won't help the backlog (i.e. new storage facility, CDRs, etc.)
Someone needs to be bold enough to tell congress the issue is funding. We want 1.6 billion more, if I remember correctly. 1.6. We are about to send 65 times that amount to foreign allies. Why is it so hard to give less than 3 percent of that to an agency that touches nearly every American every year. The house offered us a reduction of $300 million.
The backlog is the priority. It comes up at HQ every week, whether it’s from systems, or OC, or OPSOS or ARO, or OCE and everyone agrees the money for the hires and the tech have to come now if we want an improvement a year from now.
It’s a hopeless fight without funding. It’s that simple. We need authority to pay DDS more since states won’t do it and make that job worth it. If we had the money and paid DDS claims examiners at a 9/11/12 ladder the backlog would go away. If we had money to hire more folks to work PI workloads and let the seasoned examiners transfer to IC decisions and helped the DPB the backlog woold go away. But we can only print Social Security cards here, not money.
Money should be targeted for the agency to contract with consultants like BoozAllen on AI technology to automate disability decision making, which is the future whether we like it or not. The current system is extremely flawed and inefficient. Automation will be an improvement in efficiency and accuracy.
Automation is not the solution. Boooz Allen will import a flawed tech and auto deny cases
The agency has been using AL/ML in analysis of medical records for a while now. It is deployed everywhere and used to assist examiners in quickly identifying supporting information.
@1:52 SSA has been using AI for over a decade. In the disability process, SSA already uses two AI systems to aid in the disability determination process, one for using predictive modeling and text analysis in medical records (IMAGINE system), and one for selecting cases with the highest probability of allowance (you have seen these as the QDD/CAL cases). It’s difficult to fully automate the decision process. Even OHO uses AI for targeted reviews, as does DQB, but none of them can be fully automated yet because of the nature and complexity of the work involved.
Any AI in this context would be disastrous for claimants unless it really gave weight to the docs in D. It would fly in the face of SSR 16-3p and further entice decision makers to flip the standard on its head and make claimants provide sufficient medical consistent with their subjective reports to get over the bar.
If you're a regular, you've seen my past rants about 16-3p and how so many ALJs don't give nearly due consideration to their actual adjudication standard and by extension the claimants subjective reports. AI used to make these determinations is only going to result in more people denied because they didn't have sufficient medical substantiating their allegations (which, yet again, is not the standard!)
Post a Comment