The Benefit Whisperer

The episode delivers a solid volume of non-obvious, data-backed claims about prior authorization dysfunction - the 3% agreement rate across insurers, the 88% non-appeal figure, the administrative cost comparisons - but is diluted by repetitive affirmations, meandering tangents (ACA exchanges, California governor's debate), and some throat-clearing that slows the pace.

“the majority of things that require prior authorization require prior authorization from only one insurer and not the other three”

“there was something like 5,700 procedures total that required prior off. And the number per insurer also differed significantly. I mean, one of the insurers had just over 1,000 total, and another insurer was over, I think it was almost 4,000”

Several genuinely fresh framings appear - the 'digitizing the black box' critique of AI acceleration, the Fannie/Freddie mortgage standardization analogy, and the AI arms-race 'Dr. Seuss stars on/stars off machine' metaphor - but the broader thesis (insurance opacity is bad, administrative waste is enormous) is well-trodden territory in health policy circles.

“we have a little cartoon for the Dr. Seuss stars on and stars off machine, which is exactly what we're seeing. Like the people in the middle, Mr. McBean is making tons of money working for both sides of the house here”

“when you pull out your insurance card, you're paying about 30% transaction costs. When you pull out your credit card, you're paying 2% or 3%”

Don Berwick (former CMS Administrator), Kevin Shulman (Stanford health economist with active published research on prior auth), and David Scheinker (co-author of the specific prior auth database study being discussed) are all genuine practitioners with direct, first-hand evidence - rare for a benefits podcast; the caliber materially elevates the conversation.

“Dr. Don Berwick, former head of CMS, Dr. Kevin Shulman at Stanford, and David Scheinker, a PhD”

“famously at Intermountain Healthcare had, beyond decades ago, we were famous for using clinical protocols, clinical guidelines”

The episode is unusually concrete for a benefits podcast: named institutions (Cleveland Clinic, Duke, Intermountain), specific numbers (5,700 prior-auth codes, 3% cross-insurer agreement, $1.7T administrative spend, Netherlands 1/10 transaction costs, Swiss hospital with 6 finance staff), and published research referenced throughout, though some figures lack explicit sourcing.

“the Netherlands has a multi-payer system and their transaction costs are one-tenth of ours”

“we only have six people in our finance office and they're really busy. for a thousand-bet hospital”

The host asks useful clarifying follow-ups (pressing David on the exact procedure counts, raising the fiduciary duty question to Berwick directly) and structures the conversation around an employer lens, but frequently validates rather than challenges guests and lets imprecise claims (e.g., 'Hoffman effect' instead of Hawthorne) pass unchecked; the session also loses focus toward the end.

“Why don't employers, number one, why don't they know that they can demand better? And is that because the consultants don't know or because they're incentivized not to tell them?”

“David, you said something about something that requires prior auth by one carrier doesn't buy three or four others. Can you expand on that a little bit?”