Medicaid’s Role in Rising Buprenorphine Use w/ Stephen Crystal

Subscribe to UnitedHealthcare's Community & State newsletter.

Health Affairs' Rob Lott interviews Stephen Crystal of Rutgers University about his recent paper exploring how states with substantial increases in buprenorphine uptake as an opioid use disorder treatment response grew alongside increased Medicaid prescribing from 2018–24.

Order the September 2025 issue of Health Affairs.

Currently, more than 70 percent of our content is freely available - and we'd like to keep it that way. With your support, we can continue to keep our digital publication Forefront and podcast

Subscribe to UnitedHealthcare's Community & State newsletter.

Episode Transcript

Rob Lott (00:00):
Hello, and welcome to a health podocy. I'm your host,

(00:04):
Rob Lott. Friends, as you mayknow, at health affairs, this
September is all about our brandnew theme issue dedicated to
understanding the opioid crisis,a relatively nebulous but
unquestionably dire policychallenge. Now, one of the

(00:28):
particularly distinctivecharacteristics of this crisis
is just how long it's been goingon. We are definitely in the
third decade of the strugglenow, and although that's been
just incredibly frustrating andsad, it also means that there
have been a number of phases orwaves of this crisis, an

(00:51):
evolution that we can study.
And it also means that therelatively slow gears of public
policy reform have neverthelesshad the opportunity to turn and
turn and turn in ways thatmaybe, just maybe might be able
to respond to the crisis byimproving the tools at our

(01:14):
disposal and removing theobstacles that have prevented
their optimal use. Well, here weare some twenty five plus years
since the crisis began, and thequestion remains, did our policy
responses to the opioid crisisactually achieve their goals?
That is, in a way, the subjectof today's health odyssey. I'm

(01:38):
here with Doctor. StevenCrystal, distinguished research
professor and director of theCenter for Health Services
Research at Rutgers University.
Together with coauthors, he hasa new research article published
in this month's issue of HealthAffairs. The paper's title is
also its main finding. Stateswith substantial increases in

(02:00):
buprenorphine uptake did so withincreased Medicaid prescribing
from 2018 to 2024. I cannot waitto learn more about this work
and what it means forpolicymakers today. Doctor.
Steven Crystal, welcome to AHealth Odyssey.

Stephen Crystal (02:19):
It's a pleasure.

Rob Lott (02:21):
All right, let's start with some background.
Buprenorphine is sort of thegold standard in terms of
medication assisted treatmentfor people with opioid use
disorder. Yet of the populationthat might benefit from
buprenorphine, it's a prettysmall proportion that actually
use it. Can you take us back to2017 before your study period

(02:46):
begins and describe some of thebarriers that were preventing
people from getting treatment ifthey needed it?

Stephen Crystal (02:54):
Yes. That was the year actually, that was the
year that was, that was the yearthat when the opioid overdose
crisis became sufficiently onthe radar to justify a
presidential declaration ofpublic health emergency. This is

(03:16):
a period when the overdose rateswere increasing very rapidly. In
fact, they increased about fortypercent in the two years of 2016
and 2017. So everybody wasfeeling really very worried
about that trend.

(03:38):
As you mentioned in your intro,this crisis had been sort of a
slow motion crisis for a longtime, starting in the 2000s
really at scale and evenearlier. But the policy response
was really, I've come to thinkof it as frozen in time, frozen

(04:04):
in time from a very differentera of the 1980s and 1990s when
these medications for opioid usedisorder were really pretty much
considered experimental and wesort of have to control them
very carefully and exercise avery high level of surveillance

(04:24):
both on the patients and on theproviders. So this was one of
the great ironies of policy atthat time is that you could
write as a typical physicianwith a DEA license, you could
write all the opioid analgesicsprescriptions that you wanted,
and that was assumed to be, youwere assumed to be competent to

(04:47):
manage that from your generaltraining, but somehow if you
wanted to prescribe a saferpartial opioid blocker,
buprenorphine, that had reallymuch lower potential for

(05:07):
problems like respiratorydepression, you were somehow
considered to be, you could onlydo this with a special license
from the federal government, aspecial X waiver, special
training requirements, and a lotof oversight by the DEA, which
itself was a real concern tomany prescribers, many primary

(05:27):
care doctors who might considerdoing this.
Nobody in primary care wants tomake themselves a target for DEA
audit. That combined with thestigma towards people with
opioid use disorder was reallykeeping treatment in that very

(05:48):
limited framework. The samething was happening in methadone
where when it was in Methadonewas initially introduced, we
wanna make sure people have tocome in person, they have to be
supervised every single day, andfor not very good reasons, that
model has more or lesspersisted, the highly surveilled

(06:09):
model, if you like. So that wasthe uptake prescribing was
almost always from physiciansand the uptake was very poor. So
so a large part of the problemwas simply unavailability of
providers in in many areas ofthe country, in many

(06:29):
neighborhoods, and those and andand minority neighborhoods were
particularly underserved in asystem where treatment pretty
much had to be in person andfrom physicians.
So that all changed quite a lotfrom 2017 to 2024, the period

(06:51):
that we looked at.

Rob Lott (06:53):
So tell us a little bit about that. Was that change
a sort of realization in thesort of policy and treatment
community that this isn'tworking, we've gotta change the
rules. What sort of inspiredthese policy shifts and then
what what exactly were theshifts that took place?

Stephen Crystal (07:12):
This is sort of like somebody once said it takes
seventeen years for a newevidence based practice to sort
of become the standard. So in away, it was reflective of the
very inertia bound system thatwe have and the very slow pace

(07:35):
of change. And you multiply thatby the stigmatized nature of
this population, the fact thatmany providers would prefer not
to be for that not to be thepopulation in their waiting
room, and the sort ofconservatism of policy. And so I

(08:00):
think it represented a belatedresponse to the evidence that we
have, and the evidence hasreally been accumulating for a
long time, since the 1990s ofthe effectiveness of
buprenorphine. When you thinkabout medical treatments that we
have available, it's hard tothink of a lot of other

(08:22):
treatments that actually have asgood a response, we have
evidence that overdose deathscan be reduced anywhere from one
third to two thirds for peoplewhen they have day supply of

(08:45):
buprenorphine, and that evidencehas been accumulating.
And we were starting toaccumulate health services
research evidence about theinadequate uptake of MOUD. So
one of the things that that ledto, and this was actually sort

(09:05):
of, a sense, a consumer driveninitiative to get rid of the X
waiver and some of the peoplewho were influential in that
movement were people who hadlost family members.

Rob Lott (09:18):
So just sorry to interrupt, to clarify, the X
waiver was basically the limiton who was allowed to prescribe
buprenorphine. Is that accurate?

Stephen Crystal (09:28):
X waiver was a special endorsement to your DEA
license with an X literally atthe end of your DEA number that
said that you had gone throughthe training, you had been
certified and registered withthe government as a
buprenorphine provider. And butthe sequence of events was that

(09:54):
there were variousliberalizations of that
requirement. One of the earliestthings that started happening in
that period was opening upbuprenorphine prescribing to
nurse practitioners andphysician assistants. And that
was a very substantial changebecause now the advanced

(10:16):
practitioners actuallyprescribed more buprenorphine
than primary care physicians.They've become the largest
group.
So that represents an opening upof the provider shortage that
was one of the big limitingfactors in the earlier period.
So that was sort of happening in2017 through 2019. In 2020, the

(10:41):
second big change that tookplace was the opening up to
telehealth strategies forbuprenorphine in with the onset
of the pandemic, which hascontinued. And there was a lot
of advocacy to continue thoseemergency that turned out to be
an experiment, a pandemic eraexperiment that taught us a lot

(11:04):
and taught us that you didn'tneed to be entirely in person.
So that has also represented anopening up of provider supply.
And then the third big changethat happened in 2022, there
were a couple of modificationsof that ex waiver requirement

(11:24):
earlier, but then in 2022, theMAT Act was passed eliminating
that requirement. So the hopewas among people concerned with
opioid use disorder that thosethings in combination would
really lead to a dramaticreduction in this gap and the

(11:46):
gap between people with opioiduse disorder and even people
with severe problems likeoverdoses anywhere from twenty
to thirty percent or so werebeing treated and the rest were
not with MOUD. And the hope was,and the expectation was among

(12:07):
many people was that it wouldlead to very considerable
increases in buprenorphineuptake. And the bottom line from
a national point of view wasthat that didn't happen, And
that was when you look at thenational figures. So that was
very disappointing to many ofus.
So we decided to dig in a littlebit more closely to how that

(12:30):
played out on a state to statelevel.

Rob Lott (12:32):
Got it. Okay. So going into the study, you had a sense
that there was a disappointmentthere when you looked at
national numbers. You wanted tosee what was sort of driving
those lackluster results, and soyou looked at a a state by state
basis. Can you tell us a littlebit about sort of the outcomes
you studied, how you measuredthem, and how you approach this

(12:56):
question more generally?

Stephen Crystal (12:57):
One of the things that is particularly
concerning in this area is ourlack of underlying data on need,
and it turns out that the bestdata that we have on a
population basis really don'ttell us very much about the true

(13:20):
underlying rates of opioid usedisorder across states. That's
for a variety of reasons thatI'm not going to take too much
time to get into, but the datathat we have are not very good.
So after looking, wanting to getsome sense of variations in use
relative to need, basically, youlook at data, the data that we

(13:47):
do have from sources like theNational Study of Drug Use and
Health, you see about a two toone ratio across states of
identifiable opioid usedisorder, as opposed to more
like a 30 to one variation intreatment. But what we chose to

(14:13):
do for this paper, because ofthe limitations of those data,
is focus on variations inbuprenorphine use per thousand
state population and per fataloverdose as another measure of
relative need. And by all ofthose measures, you see

(14:35):
persisting enormous variationacross states, both in the rate
of treatment and in the trendsin treatment.
So when you look at a flatnational trend, it's made up of
improving states andretrogressing states. And each

(14:56):
state has a very complex storyof its own, but what it tells us
is that in our system, when itcomes to Medicaid, as they say,
when you've seen one Medicaidprogram, you've seen one
Medicaid program. So it'senormously variable, and it
became increasingly clear howmuch Medicaid was the driver of

(15:21):
overall prescribing rates acrossthe country. And the changes in
federal policy that we talkedabout, you could sort of see as
enabling or necessary but notsufficient for broader uptake.

(15:42):
And the the the the way thoseflexibilities played out
depended enormously on statelevel factors.
And and there's a whole range ofstate level factors ranging from
from supply of providers to toprior authorization requirements

(16:07):
and programs like Medicaid tolimitations on despite this
national increase in nursepractitioner prescribing,
there's a lot of variation instates and what they're allowed
to do. State rules abouttelehealth across state lines,
many, many other barriers thatare really the result of state

(16:32):
policy. So we saw that in fact,proof of concept, there were
states that responded veryrobustly during that period of
time with improving treatment,but others that actually
retrogressed.

Rob Lott (16:47):
Well, I wanna hear a little more about those states
and some of that variation, butfirst let's take a quick break.

(17:48):
And we're back. I'm here withdoctor Steven Crystal talking
about, variation in, uptake ofbuprenorphine prescribing from
state to state from between 2018and 2024. Just a moment ago,
said, you know, even while somestates were sort of making the
most of these policy changes,there were other states that

(18:10):
didn't see the same success.What do you think was behind
that variation?

Stephen Crystal (18:16):
So one of the things that came out very
strongly in this was the role ofa Medicaid expansion status,
which makes intuitive sense whenwe think about the fact that
people with opioid use disorderare disproportionately single

(18:36):
adults. They're in thisexpansion population, so the
states that have expandedMedicaid to low income single
adults have experienced, havebrought a lot more people with
opioid use disorder into theprogram, and that has had the
effect of creating a greaterfocus on addressing opioid use

(18:59):
disorder within the Medicaidprogram. We saw this very
close-up in New Jersey, whichmade a lot of initiatives within
its Medicaid program goingbeyond just the financial
eligibility. They improvedreimbursement for providers.
They eliminated the priorauthorization, and we saw from

(19:24):
those changes that they improveduptake considerably.
The original goal of thatinitiative was to make MOUD
truly a primary care, make itpart of primary care. When you
have a patient, your patient mayhave diabetes and opioid use

(19:44):
disorder and a couple of othermedical problems, and the opioid
use disorder can be managed inprimary care along with those
other conditions, and that givesus a broader range of providers
available, and that was part ofthe thinking behind eliminating

(20:09):
the X waiver. That turned out tobe a little bit optimistic in
that primary care providers areoftentimes, even with the X
waiver eliminated, not thateager or ready to start
embracing MRUD, but a lot ofother providers emerged,

(20:33):
particularly during the pandemichere, many hybrid providers and
telehealth providers. So thecombination of all those things
worked very well in some states,but for example, in the states
that never expanded, we actuallysaw over the entire twenty

(20:54):
eighteen to twenty twenty fourperiod, a slight reduction in
population level and Medicaid,and the larger reduction in
Medicaid paid prescribing. Soone of the things that was most
interesting was the ratherrobust overall improvement that

(21:19):
you tended to see in the lateexpanding states because we had
that wave of states thatexpanded in 2014 or soon
thereafter.
And then you had another wave ofstates, often sort of by popular
demand through referenda orother means that expanded
between 2018 and 2021. And thosestates on average saw quite a

(21:47):
bit of improvement. And theinteresting thing was that when
states improved their Medicaidbuprenorphine prescribing, they
also improved their all payerprescribing. Now that might seem
intuitive, but it was actually acontroversial issue in health

(22:10):
services research because someof the earlier literature found
from the early expansions thatMedicaid expansion didn't lead,
it led to more Medicaidprescribing, but it didn't lead
to much more all payerprescribing, so there was this
idea of substitution, thatprescriptions would be picked up

(22:36):
from other sources. Well,there's a few problems with
that.
One of them is that when itshifted to other sources, one of
those sources was self pay, andself pay is always going to be a
barrier to people staying onmedication, low income people.
But the interesting thing, andwe're pursuing this further with

(22:58):
more elaborate statisticalanalysis, is that the later
expansions occurred in adifferent environment where
there were fewer barriers toproviders entering the field,
fewer barriers to, you hadaccess to telehealth. So in that

(23:18):
context, when expansion tookplace, you saw increases in the
population level or payerprescribing that were not simply
substituting for for other otherother payers, but real

(23:39):
population level levelimprovement. So that's kind of
depressing story that you getfrom lack of national average
improvement is actually a muchmore complicated story, and it
tells you that that a lot ofimprovement is possible at the
state level, but it requiresstate level action.

Rob Lott (24:02):
Okay. Well, that that's a great sort of segue.
I'm wondering if you can can Iguess that what I I'm curious.
Can you think of another policyor intervention where the rates
and trends have varied sodramatically from state to
state? And I guess in thatcontext, what might those
scenarios have in common?

(24:23):
What can we learn?

Stephen Crystal (24:25):
It's interesting because in general,
most medical treatments, thestandard of care becomes a
little bit more consistentacross states. So what's
different about opioid usedisorder is first and foremost,
the amount of stigma that'sassociated at the patient level,

(24:48):
at the provider level, andthere's still this lingering
feeling that extends to somepatients and extends to many
providers that a certain amountof discomfort with medications

(25:10):
for opioid use disorder becauseof the perception that you're
substituting one drug foranother. And this has been very
prevalent in the 12 Stepmovement, for example, and it's
resulted in MOUD not beingembraced as much, for example,

(25:31):
in residential care where peoplego to a so called rehab and they
come out, now they're at morerisk than ever because they've
now lost their physicaltolerance, they're at greater
risk for an overdose, and theyhave not been started on MOUD.
So the fact that OUD is stillseen oftentimes as much of a

(25:54):
moral problem as a medicalproblem. It has really meant
that the uptake has been highlyvaried across states.
And if you were to think ofother interventions that have
been so varied, the other onethat comes to mind is really

(26:17):
another opioid use disorderrelated treatment, which is harm
reduction. And harm reductionprograms may actually have had
more to do with the recentreductions in fatal overdoses
than we've seen than treatment,since the treatment hasn't

(26:38):
expanded that much.

Rob Lott (26:40):
And when you say harm reduction, you're talking about
like naloxone or syringeservices. Syringe services,

Stephen Crystal (26:48):
safe injection sites. Well, thought is that
that's still another area whereenormous amount of work needs to
be done. And one of the thingsthat I think we've left this
study left us with is atremendous amount of concern
about whether we were going tobe able to sustain this

(27:09):
reduction in fatal overdoses.Some aspects of it are rather
mysterious. Some of it has to dowith depletion of susceptibles,
if you wanna use a technicalterm, which means that some of
the some of the people withopioid use disorder have already

(27:34):
experienced their fataloverdoses.
Some of it does have to do, wethink a lot of it has to do with
naloxone distribution, but allof this is very susceptible to
retrogression, and we've seenretrogression already in some of
the states during the unwinding.Because one of the other things
that happened during thepandemic was the suspension of

(27:57):
Medicaid disenrollments. And nowthat that's ended in 2023 and
2024, we saw retrogression atthe population level and the
Medicaid level in buprenorphinein a number of states. So when
you think about the potentialimpact of the new Medicaid

(28:20):
changes, the work requirements,which theoretically there's
exemption for people with opioiduse disorder, but how do they
apply for and get thatexemption? How do people deal
with all of this reporting forwork requirements?
And those changes were intendedto reduce Medicaid roles, and

(28:42):
they will. And we think there'stremendous potential for
retrogression on opioid usedisorder treatment in in as as
as states come under these thesenew financial pressures on their
Medicaid programs.

Rob Lott (29:04):
Okay. Wow. Well, that's probably a great place
for us to wrap up. A starkpicture of the future, but, I'm
glad that we have a better senseof the variation taking place,
and, perhaps that informationcan be used to inform policy
decisions going forward. DoctorSteven Crystal, thanks so much

(29:27):
for taking the time to chat withus.
I had a great time.

Stephen Crystal (29:30):
Thank you. It was a pleasure.

Rob Lott (29:32):
To our listeners, thanks so much for tuning in. If
you enjoyed this episode, pleaseleave a review, recommend it to
a friend, and, of course, tunein next week. Thanks, everyone.
Thanks for listening. If youenjoyed today's episode, I hope
you'll tell a friend about ahealth policy.