Policy wonks often offer greater transparency as the solution to problems like out-of-control prior authorization and predatory pricing. It's a start, not a solution.
Your point about the futility of "appeal every denial" as a patient-level strategy is exactly right. The aggregate solution is not individual appeal volume; it is plan-level visibility and accountability. And the plan sponsor lever for that is contractual, not legislative.
Self-funded employers cover more than 100 million Americans through ERISA-preempted plans. Most contract with a carrier as TPA and a PBM as the pharmacy administrator. Most of those contracts do not require denial reporting by service category, do not require appeal and overturn rate reporting, and do not specify independent medical review thresholds. Those are contract terms an employer can negotiate at renewal today, without waiting for Congress.
Your independent-board proposal is structurally sound, and worth noting that the contract version of it already exists in some Centers of Excellence arrangements where high-cost specialty decisions go to a salaried external clinical reviewer the plan controls. It is rare. It does not have to be.
The contract lever does not solve PA at the system level. But it removes the "we are powerless" frame for the largest single bloc of insured Americans. Worth naming.
There are only two reasons to deny services. First is when it is not a covered benefit. Second is when the reason for authorizing a service is not supported by evidence of efficacy.
In the first instance, it appears to be a black and white case: the services being requested are not on the list of benefits in the Evidence of Coverage. I believe that there are times when a more nuanced assessment should be permitted if legal to do so. An example of this is where (in the old days when I worked at a Southern California health plan for seniors) we used to approve the purchase and installation of an air conditioner in Palm Springs for COPD patients without A/C were frequently in the emergency room with heat exacerbated crises. Or when we approved the installation of a hand rail, a ramp to the front door, or prescription glasses when these non-covered services would avert repeated falls for a frail member.
In the second case, the reason for requesting coverage of a benefit that is often high cost / low value is not something the health plan deems persuasive. I find that these discernments are most often overturned upon appeal. Perhaps the overturn is due to giving more credit to the reasons cited by the requesting physician than the health plan would grant. In these cases, it seems to me that it would be important to check in after the finally approved service is rendered to determine whether the service was actually effective as intended. This kind of feedback loop should be incorporated into the prospective process to avoid such initial denials in the first place.
I want to make the case that benefit coverage is not well understood by most patients. I recall a case related to me by an enroller, who when a new enrollee handed him her paperwork asked how soon she could order her refrigerator. She mis-interpreted the "medical appliances" in the benefits schedule. How many people know whether chiropractor services are covered in their plan? Often times it is either wholly non-covered or only under certain circumstances. What are the circumstances and how much confusion does that create?
In medicine, it is well known that a lot of high cost / low value healthcare is rendered on a regular basis. It is a waste of resources and contributes to the inflation of health care costs. We also know that a lot of antibiotics are prescribed because patients demand them when they are really not indicated, for example. I think these issues require a lot more patient education in a way that they can understand, not in a legalese or medical jargon that fails to educate. These steps could help avoid the whole denial cycle in the first place.
A bit off today's topic, but I get a very guilty feeling if there is a health-insurance-wonky substack, and I don't plop down the latest news on the effect of the lapse, on Jan 1, 2026, of the ACA expanded subsidies. (Which I presume many people who read these want to know about.)
Thus, I'll plop down the latest status on the effect of the ACA expanded subsidies, based mainly on Charles's Gaba's substack posts. (I subscribe to his substack. He seems to catch every bit of new information.)
Based only on the initially-available data, including just what happened through the end of Open Enrollment in January, these are the findings:
OPEN-ENROLLMENT RESTRICTED, NOT SPLITTING BY INCOME GROUP
About 4.9% of the 24 million people did not re-enroll. This is across all income groups considered together, and not splitting people down by income.
This available-now open-enrollment-only data has limitations that make it tend to overstate the number keeping coverage, especially because about half of the people considered re-enrolled were automatically re-enrolled, and are expected to not actually follow through and pay for coverage. Additional others are expected to drop coverage later this year or eventually, finding the premiums work out now to be too high to successfully manage.
OPEN-ENROLLMENT RESTRICTED, RESTRICTING TO OVER THE RETURNED 400% FPL SUBSIDY CLIFF
This zone of income is of special interest to me, because it's where the really humongo cost increases arise.
Thus, below 400% of FPL, people would have experienced a rise in the net premium they have to pay of up to 2% to 4.5% of their income, which would be for the same plan as without the lapsed expanded subsidies. (Many of those might have been able to bring down the increase in what they have pay by choosing a cheaper plan, which, as we know from previously mentioned data details, did actually happen.)
Above 400% of FPL, where subsidies stopped entirely with the lapse of the expanded subsidies, it is not hard to find increases of 40% of income on the net premium, with, thrown in, a jump from a mid-level plan down to the cheapest available plan in that case. Thus, in Wyoming, a 62 year old couple with an income of $88,000 a year has premium jump from about $6,000 a year to about $40,000 a year, going from a mid-range plan to the cheapest available plan, as well.
In some other states, it's not 40% of income. Older people might still have 25% of income jumps, and in younger ones, closer to 30, it might be 5% to 20% of income. (Very roughly--I have not researched this precisely.)
Thus, one expects people over the returned 400% of FPL "subsidy cliff" might have really been thrown under the bus here, and both my look at the data:
show that, with similar numbers, though some of the exact computations we went after differ a little bit.
Thus, overall, we have numbers like around 38% of people in the just-over-the-cliff 400% to 500% of FPL group dropping out of being covered from 2025 to 2026, with it being even more of
a serious drop in some of the highest-premium states, like WV and WY.
However, with these complexities:
Charles and I both observed about 250,000 to 300,000 people seem to have moved themselves into the just below-the-cliff 300% to 400% of FPL income group on the exchange since 2025. Obviously, those people have either decided to work less and make less money, saving that humongo premium jump over-the-cliff, or work the same but keep profits in a corporation that they own that employs them--keeping the money outside of ACA-subsidy use. O, they were just putting in a false low number as estimated 2026 income on the exchange. (In the latter case, they will have to pay back all of the savings at tax filing time for 2026, which ends a year from today! Those are the rules, and those people will not be pleased!)
Anyway, where I or Charles has found say that 38% dropping, it is expected that some of the people in the 38% actually didn't drop, but took the work-less or try-cheating approach.
There is another slight complexity coming from that a proportion of people who appear as missing from 400% to 500% FPL group in 2026, say that 38% again, may have gotten omitted because, seeing they were not going to get a subsidy, they avoided entering an income on the exchange. (This not entering an income, if they kept coverage, would tally them as covered in 2026, but in a different group, the "unknown income" group, rather than the 400% to 500% FPL group in the released data from CMS. ) I tend to believe this number moving to unreported income because no subsidy now is small, because, all through open-enrollment, there was the possibility of a partial removal of the 400% FPL subsidy cliff in legislation that even a few Republicans in Congress were supporting. (So, entering an income would be the way to make that possible much lower premium possible automatically, without lots of work later.)
Being very conservative, though, I did a pooling method of all the over the cliff income groups with the no-income-reported group, and still found about a 16% coverage drop in that combined group. (It's a certain group, probably of mostly over the cliff people, dropping at an 16% rate. Still much higher than the overall 4.9% dropping rate. )
My intention is, eventually, to make, and post, a table where you can kind of see, in all of these over the cliff cases, how many actually dropped coverage, how many may have just lowered their income to below the cliff, with it also clear in my table which numbers are subject to possible smallish overestimates of proportion dropping from people to switching to not reporting an income on the exchange.)
(Post continued as a reply to the comment--comment length limit too short)
As long at there are insurance companies to profit off our health care without providing any health benefit then in my mind that is rent-seeking and should be banned. We need to move to a single payer system which eventually needs to move away from a fee for service reimbursement model. We're bleeding out here in the USA.
As time goes on, there is the expectation that many more people will drop than shown in the prior numbers from just open enrollment. Many, indeed, were auto-re-enrolled, and may never pay the premium.
Such numbers, being released at the Federal Level, are at least many months away. I am also not sure those numbers will include by-income numbers, allowing us to isolate the over-the-cliff cases.
Charles Gaba is always looking for early data, from single states with their own exchange, where we get an earlier idea.
He already has one post on Colorado, https://charlesgaba.substack.com/p/colorado-q1-effectuated-enrollment , where already he has a 5.1% higher drop rate than at the same time last year. (Further, Colorado has some partial state-financed mitigations of the subsidy increases, so it would be a state that tends to understate the eventual coverage losses across all states, most of which did not mitigate at all.)
As well, from the KFF survey I cited earlier, we have, about 17% reporting they are not confident they can afford their premium all year.
Charles Gaba (here: https://charlesgaba.substack.com/p/this-sounds-familiar-wakely-report ) also caught a very significant report by a Wakely consulting group, which does actuarial work for insurers. They apparently have access to lots of insurer data on who followed through after open enrollment and actually paid, which is especially important because about half of people counted as continuing coverage at open enrollment from 2025 into 2026 are people who were auto-renewed by their exchange, and may have seen the much higher amounts they would have to pay, and never paid the bill.
The report from Wakely that Charles cites is this one:
"While plan selections declined by roughly 5%, actual enrollment is projected to fall between 17% and 26% on average when accounting for unpaid premiums and ongoing attrition. Some states may see reductions higher than 26%, with higher reductions skewing toward states operating under a Federally Facilitated Exchange."
When you look at people who actually paid for coverage ("effectuated enrollment" seems to be the official term), in Georgia, the number is 28% less than last year, same period.
"There’s one other important data point missing from the January CMS Medicaid/CHIP enrollment snapshot:
Until now, the summary report also included a brief mention of total effectuated ACA marketplace enrollment in Qualified Health Plans (QHPs), rounded off to the nearest 100,000. As of December effectuated enrollment was ~21.8 million people.
However, starting this month, this data point is missing...and that’s not by accident; it includes this footnote:
As of the January 2026 data, Marketplace enrollment data are no longer included in this report but will be available separately soon."
(Not to impugn Dr. Oz unjustly, but he has a bad record around propaganda and deception:
Around making an ACA change that the Republicans are trying to get through having the EXACT OPPOSITE of the effect it has
Your point about the futility of "appeal every denial" as a patient-level strategy is exactly right. The aggregate solution is not individual appeal volume; it is plan-level visibility and accountability. And the plan sponsor lever for that is contractual, not legislative.
Self-funded employers cover more than 100 million Americans through ERISA-preempted plans. Most contract with a carrier as TPA and a PBM as the pharmacy administrator. Most of those contracts do not require denial reporting by service category, do not require appeal and overturn rate reporting, and do not specify independent medical review thresholds. Those are contract terms an employer can negotiate at renewal today, without waiting for Congress.
Your independent-board proposal is structurally sound, and worth noting that the contract version of it already exists in some Centers of Excellence arrangements where high-cost specialty decisions go to a salaried external clinical reviewer the plan controls. It is rare. It does not have to be.
The contract lever does not solve PA at the system level. But it removes the "we are powerless" frame for the largest single bloc of insured Americans. Worth naming.
Let me begin with a disclaimer: I am not a fan of United Healthcare. However, to give them their due, I noticed a positive posting about this topic of referrals today in FierceHealth that deserves mention. https://www.fiercehealthcare.com/payers/unitedhealthcare-reduce-prior-auth-requirements-30?utm_medium=email&utm_source=nl&utm_campaign=HC-NL-FierceHealthPayer&oly_enc_id=5345J6883012C3J
There are only two reasons to deny services. First is when it is not a covered benefit. Second is when the reason for authorizing a service is not supported by evidence of efficacy.
In the first instance, it appears to be a black and white case: the services being requested are not on the list of benefits in the Evidence of Coverage. I believe that there are times when a more nuanced assessment should be permitted if legal to do so. An example of this is where (in the old days when I worked at a Southern California health plan for seniors) we used to approve the purchase and installation of an air conditioner in Palm Springs for COPD patients without A/C were frequently in the emergency room with heat exacerbated crises. Or when we approved the installation of a hand rail, a ramp to the front door, or prescription glasses when these non-covered services would avert repeated falls for a frail member.
In the second case, the reason for requesting coverage of a benefit that is often high cost / low value is not something the health plan deems persuasive. I find that these discernments are most often overturned upon appeal. Perhaps the overturn is due to giving more credit to the reasons cited by the requesting physician than the health plan would grant. In these cases, it seems to me that it would be important to check in after the finally approved service is rendered to determine whether the service was actually effective as intended. This kind of feedback loop should be incorporated into the prospective process to avoid such initial denials in the first place.
I want to make the case that benefit coverage is not well understood by most patients. I recall a case related to me by an enroller, who when a new enrollee handed him her paperwork asked how soon she could order her refrigerator. She mis-interpreted the "medical appliances" in the benefits schedule. How many people know whether chiropractor services are covered in their plan? Often times it is either wholly non-covered or only under certain circumstances. What are the circumstances and how much confusion does that create?
In medicine, it is well known that a lot of high cost / low value healthcare is rendered on a regular basis. It is a waste of resources and contributes to the inflation of health care costs. We also know that a lot of antibiotics are prescribed because patients demand them when they are really not indicated, for example. I think these issues require a lot more patient education in a way that they can understand, not in a legalese or medical jargon that fails to educate. These steps could help avoid the whole denial cycle in the first place.
A bit off today's topic, but I get a very guilty feeling if there is a health-insurance-wonky substack, and I don't plop down the latest news on the effect of the lapse, on Jan 1, 2026, of the ACA expanded subsidies. (Which I presume many people who read these want to know about.)
Thus, I'll plop down the latest status on the effect of the ACA expanded subsidies, based mainly on Charles's Gaba's substack posts. (I subscribe to his substack. He seems to catch every bit of new information.)
Based only on the initially-available data, including just what happened through the end of Open Enrollment in January, these are the findings:
OPEN-ENROLLMENT RESTRICTED, NOT SPLITTING BY INCOME GROUP
About 4.9% of the 24 million people did not re-enroll. This is across all income groups considered together, and not splitting people down by income.
This available-now open-enrollment-only data has limitations that make it tend to overstate the number keeping coverage, especially because about half of the people considered re-enrolled were automatically re-enrolled, and are expected to not actually follow through and pay for coverage. Additional others are expected to drop coverage later this year or eventually, finding the premiums work out now to be too high to successfully manage.
Charles Gaba, in the data, recorded a high proportion of people keeping coverage, but downgrading from say silver to a bronze plan, (thus taking on potential excessive copays if they do get sick, with larger deductibles and out-of-pocket maxes). Like 10% or so of people moving silver to bronze from Gaba's computations: here: https://charlesgaba.substack.com/p/final-2026-open-enrollment-report-d4f?utm_source=profile&utm_medium=reader2 .
KFF has, by an alternate method, it's own survey, about 28% downgrading coverage. (here: https://www.kff.org/public-opinion/a-follow-up-survey-of-aca-marketplace-enrollees/ ).
OPEN-ENROLLMENT RESTRICTED, RESTRICTING TO OVER THE RETURNED 400% FPL SUBSIDY CLIFF
This zone of income is of special interest to me, because it's where the really humongo cost increases arise.
Thus, below 400% of FPL, people would have experienced a rise in the net premium they have to pay of up to 2% to 4.5% of their income, which would be for the same plan as without the lapsed expanded subsidies. (Many of those might have been able to bring down the increase in what they have pay by choosing a cheaper plan, which, as we know from previously mentioned data details, did actually happen.)
Above 400% of FPL, where subsidies stopped entirely with the lapse of the expanded subsidies, it is not hard to find increases of 40% of income on the net premium, with, thrown in, a jump from a mid-level plan down to the cheapest available plan in that case. Thus, in Wyoming, a 62 year old couple with an income of $88,000 a year has premium jump from about $6,000 a year to about $40,000 a year, going from a mid-range plan to the cheapest available plan, as well.
In some other states, it's not 40% of income. Older people might still have 25% of income jumps, and in younger ones, closer to 30, it might be 5% to 20% of income. (Very roughly--I have not researched this precisely.)
Thus, one expects people over the returned 400% of FPL "subsidy cliff" might have really been thrown under the bus here, and both my look at the data:
https://normspier828307.substack.com/p/aca-2026-enrollment-after-expanded
and Charles Gaba's:
https://charlesgaba.substack.com/p/final-2026-open-enrollment-report-b83?utm_source=profile&utm_medium=reader2
show that, with similar numbers, though some of the exact computations we went after differ a little bit.
Thus, overall, we have numbers like around 38% of people in the just-over-the-cliff 400% to 500% of FPL group dropping out of being covered from 2025 to 2026, with it being even more of
a serious drop in some of the highest-premium states, like WV and WY.
However, with these complexities:
Charles and I both observed about 250,000 to 300,000 people seem to have moved themselves into the just below-the-cliff 300% to 400% of FPL income group on the exchange since 2025. Obviously, those people have either decided to work less and make less money, saving that humongo premium jump over-the-cliff, or work the same but keep profits in a corporation that they own that employs them--keeping the money outside of ACA-subsidy use. O, they were just putting in a false low number as estimated 2026 income on the exchange. (In the latter case, they will have to pay back all of the savings at tax filing time for 2026, which ends a year from today! Those are the rules, and those people will not be pleased!)
Anyway, where I or Charles has found say that 38% dropping, it is expected that some of the people in the 38% actually didn't drop, but took the work-less or try-cheating approach.
There is another slight complexity coming from that a proportion of people who appear as missing from 400% to 500% FPL group in 2026, say that 38% again, may have gotten omitted because, seeing they were not going to get a subsidy, they avoided entering an income on the exchange. (This not entering an income, if they kept coverage, would tally them as covered in 2026, but in a different group, the "unknown income" group, rather than the 400% to 500% FPL group in the released data from CMS. ) I tend to believe this number moving to unreported income because no subsidy now is small, because, all through open-enrollment, there was the possibility of a partial removal of the 400% FPL subsidy cliff in legislation that even a few Republicans in Congress were supporting. (So, entering an income would be the way to make that possible much lower premium possible automatically, without lots of work later.)
Being very conservative, though, I did a pooling method of all the over the cliff income groups with the no-income-reported group, and still found about a 16% coverage drop in that combined group. (It's a certain group, probably of mostly over the cliff people, dropping at an 16% rate. Still much higher than the overall 4.9% dropping rate. )
My intention is, eventually, to make, and post, a table where you can kind of see, in all of these over the cliff cases, how many actually dropped coverage, how many may have just lowered their income to below the cliff, with it also clear in my table which numbers are subject to possible smallish overestimates of proportion dropping from people to switching to not reporting an income on the exchange.)
(Post continued as a reply to the comment--comment length limit too short)
As long at there are insurance companies to profit off our health care without providing any health benefit then in my mind that is rent-seeking and should be banned. We need to move to a single payer system which eventually needs to move away from a fee for service reimbursement model. We're bleeding out here in the USA.
BEYOND OPEN ENROLLMENT:
As time goes on, there is the expectation that many more people will drop than shown in the prior numbers from just open enrollment. Many, indeed, were auto-re-enrolled, and may never pay the premium.
Such numbers, being released at the Federal Level, are at least many months away. I am also not sure those numbers will include by-income numbers, allowing us to isolate the over-the-cliff cases.
Charles Gaba is always looking for early data, from single states with their own exchange, where we get an earlier idea.
He already has one post on Colorado, https://charlesgaba.substack.com/p/colorado-q1-effectuated-enrollment , where already he has a 5.1% higher drop rate than at the same time last year. (Further, Colorado has some partial state-financed mitigations of the subsidy increases, so it would be a state that tends to understate the eventual coverage losses across all states, most of which did not mitigate at all.)
As well, from the KFF survey I cited earlier, we have, about 17% reporting they are not confident they can afford their premium all year.
Charles Gaba (here: https://charlesgaba.substack.com/p/this-sounds-familiar-wakely-report ) also caught a very significant report by a Wakely consulting group, which does actuarial work for insurers. They apparently have access to lots of insurer data on who followed through after open enrollment and actually paid, which is especially important because about half of people counted as continuing coverage at open enrollment from 2025 into 2026 are people who were auto-renewed by their exchange, and may have seen the much higher amounts they would have to pay, and never paid the bill.
The report from Wakely that Charles cites is this one:
https://www.prnewswire.com/news-releases/wakely-analysis-signals-significant-enrollment-shifts-in-aca-individual-market-as-2026-unfolds-302743570.html .
From it, I quote:
"While plan selections declined by roughly 5%, actual enrollment is projected to fall between 17% and 26% on average when accounting for unpaid premiums and ongoing attrition. Some states may see reductions higher than 26%, with higher reductions skewing toward states operating under a Federally Facilitated Exchange."
Late important news from Gaba:
here (https://substack.com/home/post/p-194964291 )
When you look at people who actually paid for coverage ("effectuated enrollment" seems to be the official term), in Georgia, the number is 28% less than last year, same period.
(Wow! That's a big drop!)
--
Otherwise, just today, Charles caught something, possible information suppression. From https://charlesgaba.substack.com/p/cms-posts-january-2026-medicaid-chip , I quote:
"There’s one other important data point missing from the January CMS Medicaid/CHIP enrollment snapshot:
Until now, the summary report also included a brief mention of total effectuated ACA marketplace enrollment in Qualified Health Plans (QHPs), rounded off to the nearest 100,000. As of December effectuated enrollment was ~21.8 million people.
However, starting this month, this data point is missing...and that’s not by accident; it includes this footnote:
As of the January 2026 data, Marketplace enrollment data are no longer included in this report but will be available separately soon."
(Not to impugn Dr. Oz unjustly, but he has a bad record around propaganda and deception:
Around making an ACA change that the Republicans are trying to get through having the EXACT OPPOSITE of the effect it has
(here: https://normspier828307.substack.com/p/cost-sharing-reductions-silver-loading ; initial discovery of the Oz untruth is by Charles Gaba) ,
as well as this propagandistic little gem:
https://www.youtube.com/watch?v=0z4GhR94KRI
clearly geared at those with miserable critical thinking skills, who are innumerate, as well.)
So, Dr. Oz, I suspect, is up to no good, and trying to delay important statistics!
(Dr. Oz. J'accuse!)