Extreme Heat’s Impact on Healthcare Use and Spending

Health Affairs Publishing’s Rob Lott speaks to Jeff Romine of Carelon Research about his recent paper exploring new research on how extreme heat affects health care use and costs, finding consistent increases in emergency department visits and some hospitalizations, but little change in outpatient care.

Order the May 2026 issue of Health Affairs.

Sign up for our free Health Affairs newsletters to stay up to date on health policy news and analysis.

Rob Lott: As listeners know, the month of May marked the release

of a full theme issue of Health Affairs Journal dedicated

entirely to research and analyses at the intersection of

climate, health, and equity. Now, at this point, it's pretty

clear that climate change has significant and serious impacts

on people's health. We also know that different populations are

affected differently with the most vulnerable populations

paying the highest price. And yet there's still so much we

don't know, including how this variation and these costs

manifest when controlling for factors like insurance status,

geography, and age. That information sure would be good

to know, especially as we think about potential policy

interventions aimed at mitigating the health

consequences of climate change.

These are some of the questions we're going to consider on the

podcast today. I'm here with Doctor. Jeff Romine, a research

scientist with Carillon Research. And together with

coauthors, he has a new article in the May issue of Health

Affairs describing, quote, Extreme heat, healthcare use and

costs, Evidence from Commercial Insurance, Medicaid and Medicare

Advantage. This is such an interesting paper, a lot of

really illuminating findings, I can't wait to hear how they did

it.

Doctor. Romine, welcome to our Humble podcast.

Jeff Romine: Yeah. Thanks for having me. Very excited to, be

here and always love talking about research.

Rob Lott: Awesome. Well, you're in the right spot for it. Why

don't we, maybe start with the existing literature, in this

space? And I'm hoping maybe you can tell us a little bit about

what we know about the impact of extreme heat on people's health,

especially things like emergency department visits and mortality.

Jeff Romine: Yeah. That's a great question, and my answer

will not cover the totality of the body of literature. There's

a very strong body of evidence showing that extreme heat is

definitely associated or depending on you're if an

economist like me or not, would say, causes, increased

utilization when it's hot. So, as you had in the question, ED

utilization is where there's some strongest evidence. There's

really good data on ED utilization that, academics can

use, really good public access data there.

So very strong evidence from many different papers showing

that all cause utilization increases, heat related illness

increases, and just very many different types of conditions.

So, like, illness, chronic kidney disease, mental health.

Mental health is especially impacted by extreme heat, so

very detrimental to mental health. Can definitely increase

the utilization of four severe mental health disorders. So very

strong evidence of many different great authors doing

great work showing that, yes, there's there's a strong

association between extreme heat and ED utilization.

Rob Lott: Okay. So we've got great data on emergency

department visits. How about mortality?

Jeff Romine: Yeah. There's, very strong evidence showing that

mortality increases when it is hotter, especially among the

more elderly populations and especially more, among

individuals that have, higher chronic disease load. So there's

also for mortality, there's also very strong evidence that

mortality increases when it's cold. So, not just extreme heat

impacts mortality, but also, extreme cold. So on both ends of

the temperature spectrum having impacts.

Rob Lott: So we know less about outpatient use. Is that right?

And I'm curious if you can tell us a little bit why there's less

of an evidence base in that space.

Jeff Romine: Yeah, yeah. So ED utilization has very strong body

of data to be able to be used, and there's just not this data

to be available for studying outpatient use. It really

requires an administrative claims like insurance database.

You'd have to go to CMS or like a pay all payer database, people

would have this sort of outpatient utilization data. So

much more rare data to be available.

And there's also some suggestions that the signal

could just be smaller for outpatient utilization. So

there's just not as much research being done there as

it's not the focus.

Rob Lott: So you said the signal might be not as strong. I guess

what you're sort of getting at is that if there's an impact, it

might not be as sort of pronounced as sort of a increase

in emergency department visits after a heat wave, that kind of

thing. Is that sort of what you're getting at?

Jeff Romine: Right, right. So it could be that there has been

research in this field that just hasn't been, could've or might

not have been published because it's just a weak signal or might

just be that the methods that were being used couldn't detect,

any changes in ED or in outpatient utilization

associated with heat. So it's one where it's almost hard to

guess at because there is just such a lack of evidence in

outpatient utilization. The little bit of outpatient use,

and its association with heat is from a few different papers

showing that there's cancellations of outpatient

visits, like planned outpatient visits when it's hot, so almost

an avoidance behavior there. And there's a few papers from China

showing that, cause specific utilization increases, so

respiratory and other conditions like that, but nothing looking

at the totality of outpatient use.

It's really and especially totality across an entire

country. It's really just challenging data to have.

Rob Lott: Got it. Well, you attempted to, fill some of that

gap, with your paper in the May issue. You looked at the

relationship between extreme heat and ED inpatient and

outpatient use among people with different kinds of coverage, a

large national insurer with commercial insurance, Medicaid

managed care, and Medicare Advantage plan. You also looked

at costs. What were some of your top line findings?

Jeff Romine: So really the top line findings are that we find

similarities between our paper and other work showing that when

it's hot, that ED utilization increases. We also see some

increases in hospitalization or what we call in our paper in

inpatient utilization, and that's been documented in some

papers, but not nearly to the extent of ED utilization or

mortality. And then our final place of service that we looked

at is base unsurprisingly based on the questions we've talked

about so far is outpatient use. And because of the dataset that

we're using, we actually outpatient utilization for

around forty million people in our sample. So we find that

there really was not much of an association between outpatient

use and hotter temperatures.

So could be many reasons why this would be, but really just

going into more detail on some of these, we find similar

impacts on ED utilization as other papers, but what we add to

the table or bring to the table is that we actually have direct

cost data. So by having direct cost data, we're not we don't

have to approximate a change in cost based on how assuming that

the change in cost is the same as the change in utilization. So

we're actually modeling cost directly, and we actually find

that there's actually similarities between ED cost and

ED utilization. It's about the same percent change in ED cost

and ED utilization, so that is something new that we bring to

the table there. For inpatient utilization, see that there are

a couple different groups, ones that we might think of as more

vulnerable, such as individuals who have Medicare Advantage

coverage or children or some insurance types of children, as

well as some member or some individuals on Medicaid.

So see that they actually have increases in inpatient

utilization and hospitalization, which not totally new, but a

little bit new that we divided up by age group and insurance

type. And then finally, for the outpatient, we, as I said

earlier, really don't see too much there, which is new and

also a bit surprising.

Rob Lott: Great. Well, I want to ask you a little bit about that

surprise, if you will. But first, let's take a quick break.

And we're back. I'm here with Doctor.

Romine, a research scientist with Caroline Research, talking

about, evidence from commercial insurance, Medicaid, and

Medicare Advantage of the effect of extreme heat on healthcare

use and costs. And you just a moment ago said that there was a

bit of a surprise in your findings. Can you dig into that

a little bit? What was unexpected?

Jeff Romine: Yeah. So what we found is that outpatient

utilization, unlike ED utilization, hospitalization,

mortality really didn't have a clear trend across the

temperature distribution. The other outcomes, we'll see

decrease when it's colder, and as the temperature increases,

we'll see an increase relative to a temperate temperature. So

everything we do in our paper is relative to essentially 60 to 70

degrees. What we find is that over a 100 degrees, there's not

more outpatient utilization than between 60 to 70 degrees, which

seeing as all of our other other outcomes, we actually ran

outpatient last.

It was just our final outcome. We had seen everything up to

that point. We have this clear trend. It utilization goes up

when it's hotter, and we don't see that for outpatient

utilization. And there's a couple different things that we

hypothesized in the paper and that based on some papers that

look at planned outpatient utilization, it could be that

it's just an avoidance behavior.

People are people are canceling planned outpatient surgeries or

outpatient visits because they want to avoid really hot

temperatures, which is totally reasonable, mitigation behavior.

Don't wanna go out when it's hot, but it also, beyond the

scope of this paper, mean that there's just missed

appointments, missed care. So even though there isn't an

increase, it could be because not for good reasons. So we

don't it's hard for us to get that in our paper, but that's

based on other papers. The other unexpected finding that we had

is when we were first starting this project, we hypothesized

that cost would actually increase more than utilization,

and that would kind of indicate that there's actually increase

in the average cost of visits when it's hotter, but we really

don't see too much of a difference in percent change in

cost or the percent change in utilization.

So we're actually kind of surprised by that aspect as

well.

Rob Lott: Got it. And can you say a little more about sort of

the time frame where you were looking at these these costs and

the the utilization in relation to these extreme heat events?

And I guess what I'm getting at is, is it possible that the the

increased costs or increased utilization might might be

taking place further downstream than than what you're looking

at.

Jeff Romine: Yeah. A fantastic question. So, just to to help

the listeners get context of just our the years we're looking

at, we were looking at 2016 to 2023. And then for our actual

analysis of how is heat associated with utilization and

cost, our main specification, our main model is looking at in

the same week. So we allow for how We look at the number of

days in different temperature bins within the same week.

So we know how many days are above 100 degrees, how many days

are between 80 to 90, and so on, and we run a regression with a

bunch of controls to account for seasonality, different factors,

and that's really looking at how does same week temperature, how

is that associated with cost utilization in that same week?

So to kind of test whether there are lagged effects, so does heat

now increase utilization later, we actually allow for three

weeks, the following three weeks to be included as well. We can

understand how does heat now affect utilization three weeks

later. And it changes our results number we're getting,

but doesn't change the magnitude or the relative significance of

our results. So we actually see most impacts contemporaneously,

so within the week, and that really matches other papers in

the literature that really show that most of the impact is

happening in the same week, the same day even, and then just the

next day.

Rob Lott: Got it, got it, okay. So one could envision a world

where someone, you know, is stuck at home on an extremely

hot day, they, you know, they are, you know, suffering and

affects their health and maybe it exacerbates a chronic

condition and, you know, they're still struggling with that a

week down the road or two weeks down the road or even, you know,

months and beyond. So I can imagine that sort of timeframe

could have an effect on how, act I don't want to say how

accurate. So the timeframe could sort of affect the downstream

outcome that you're measuring. Is that a fair take?

Jeff Romine: Yeah, yeah. So there will be sort of off set as

it's delayed care, have utilization later, but it also

could be that heat is actually pushing care that would have

needed to happen later earlier. So it could exacerbate

conditions sooner. So there's this trade off between delay but

also delaying more possibly less acute care but also pushing more

acute care sooner. So there's this trade off and this tension

between these two.

So that's why we account for this month long lag as well to

ensure that there's not these longer, what you call medium

term effects that would change our results. But it definitely

could be that we're we are not including years long. So we're

not looking at how it heat may increase the progression of

chronic kidney disease or something like that. So we're

not looking at progression of disease, but we are seeing that

most of the effect is contemporaneous, is within the

same week. We're not seeing too much of a difference when we

account for this more medium month long impacts.

Rob Lott: Got it. Well, as with most of the papers we publish

here, it's not so simple as you're suggesting there's a lot

of trade offs and a lot of overlapping factors that have

different impacts. I guess in that context, you know, on one

hand, the fact that different populations feel different

impacts from extreme events is not very surprising. On the

other hand, this complicates any effort to find the kind of

interventions that might be most effective in the face of these

events. In other words, like one size fits all solutions may not

be sufficiently targeted.

And I'm curious, how, you think we should be thinking about this

tension.

Jeff Romine: There are differences in how heat is going

to impact different populations. What we find in our paper is

that almost all or all populations have an increase in

need of utilization, but only some have an increase in

inpatient utilization. So we see that targeting older individuals

may provide an opportunity to have a more focused and per

person higher impact and have a higher improvement in health per

person. So there's always gonna be limitations in what can be

done, so there is evidence that some individuals are gonna be

more impacted, more susceptible to heat, and allows for

targeting there. And there's many other papers showing that

there are specific chronic conditions where individuals are

more impacted.

The paper that we did previously to this one looked at only the

chronic kidney population. We saw they were actually impacted

more by heat than just the average individual in this

paper. So having cuts of the population and understanding how

heat differentially impacts groups allows for, in the time

being, there to be more targeted interventions and allows for,

just as a from a research perspective, the possibility to

find, like, statistically significant impacts of these

interventions. Are these interventions effective? Are

they decreasing the utilization and cost that is associated with

heat?

So by targeting different populations, it allows for

larger improvements in health per person on these targeted

pilot interventions or so on, and really allows for

intervention to be most helpful by doing this targeting.

Rob Lott: Got it. Well, a great framework for us to think about

future research as well. Doctor. Romine, thanks so much for

taking the time to talk with us today. I had a really good time.

Jeff Romine: Thanks for having me.

Rob Lott: Well, everyone, thanks for tuning in. If you enjoyed

this episode, please recommend it to a friend, leave a review,

and of course, tune in next week. Thanks, everyone.

Jeff Romine: Thanks for listening. If you enjoyed

today's episode, I hope you'll tell a friend about A Health

Podyssey.