Is Uber Good (or Bad) for Your Health?

Eight years ago, David Slusky’s wife went into labor. It was the middle of the night. They needed to get to the hospital, but they lived in Philadelphia and didn’t have a car.  So they did what they’d do any other night they needed a ride… They called an Uber. 

David SLUSKY: And my wife said, just, just play it cool. Don’t tell him I’m in labor. Maybe he won’t figure it out. 

But when the driver took an unexpected route to the hospital, David became a little bit alarmed. 

SLUSKY: I had this moment of this is the worst night to be kidnapped by your Uber driver. The driver then goes past Spruce, which goes the wrong way. And then turns left on pine and then turned to us when stopped and said, I have three children at home. I understand your situation. Pine is a lot smoother than locust. And took us to the hospital and we had our daughter who’s now eight and a half. And so that got me thinking. Is this a common story? Are people using ride-sharing as part of our healthcare transport system? 

*      *      *

That night, David Slusky got a healthy new baby daughter —  and the seed of a new research project. 

A few years later, Slusky, by then an economist and professor at the University of Kansas, came back to those questions about how people get to the hospital when they’re in a rush. The answers show just how limited our thinking has been when it comes to the basic act of getting  medical care quickly. 

From the Freakonomics Radio Network, welcome to Freakonomics, M.D. 

I’m Bapu Jena. I’m an economist but I’m also a medical doctor. And in each episode, I dissect fascinating questions at the sweet spot between health and economics.  

Today: Calling an ambulance is often the first thing we think to do if someone is in need of immediate medical care. But  could rideshare companies offer some people a safe alternative for getting to the emergency room? We’ll look at a study that examined that very question, and some other research on the hidden benefits — and drawbacks — of Ubering.

David Slusky and a colleague, Leon Moskatel, uncovered a fascinating connection between Uber and ambulances. But before we unpack their findings, I just want to acknowledge the controversies surrounding rideshare companies. They’ve made headlines for labor issues, like how they treat and compensate their drivers. And there have been troubling allegations of sexual assault in the cars and sexual harassment at company headquarters. 

We could ask a ton of different economic questions to understand how these companies  impact people’s lives. But today we’re looking at just a couple of different ways they impact our health. 

So back to that ambulance study:

We all know the benefits of ambulances. Turn on those sirens and traffic moves out of their way. They can get you to the hospital quickly. Not to mention that they’re staffed with trained technicians and stocked with important medical equipment and supplies. Medical care can start on the way to the hospital — and a few precious minutes can save lives in an emergency.  

But ambulances have their drawbacks too. 

Loren ALDER: These are often a thousand dollar bill for a 10-mile ambulance ride, um, which can obviously put a very large strain on people. You know, even if the insurance company is paying four or $500, you know, you’re still left with a, you know, 500, $600 bill, um, that’s being sent to you as the balance. 

That’s Loren Adler. He’s the associate director of the U.S.C. Brookings Shaffer Initiative for Health Policy. He’s an expert on the practices and laws that sometimes lead to huge healthcare bills. Speaking of huge health care bills, reports of costly ambulance trips have made national headlines — $3,600 for a four-mile ride; $8,400 to go from one hospital to another. And when it comes to these bills, Loren explains that it’s partly because billing for ambulance rides is different than billing for say, a primary care doctor. Doctors have to have contracts with insurance companies if they want to see patients. But ambulances don’t have that same restriction. 

ALDER: They had this ability to say, Okay I’ll try to get as much money as I can out of the insurance company. And then on top of that, I can balance the bill with the patient and try to recuperate even more money that way. 

In a lot of states, there’s no cap on what ambulance companies can charge. Often it turns out there also isn’t a lot of competition. A single company can easily end up as the sole provider of ambulance rides in any given region and they could end up charging more to patients and insurers as a result. 

ALDER: I think fundamentally black and white, this is wrong. This is just a broken market that can never really work. By law, when someone calls 911, the ambulance has to pick them up, whether they have insurance, whether they can pay or not. And similarly, the insurer has basically no ability to steer patients to a preferred ambulance provider. 

Ambulance bills are often a surprise, not just because the health scare was unexpected but also because, unlike other medical care, there’s no way to find out beforehand what the ride will cost. The No Surprises Act, which takes effect in January, is supposed to eliminate a lot of practices like this. But ground ambulances aren’t actually a part of that law. You can see why patients might be tempted to bypass them, especially if their “emergency” didn’t require things like cardiac monitoring or extra oxygen. By one estimate, at least 11 percent of ambulance trips weren’t actually medically necessary. 

So… after economist David Slusky’s wife gave birth — he started to wonder how often people heading to the hospital did the same thing that he and his wife did — Calling an uber  when faced with a medical need that was too urgent for public transportation but not urgent enough for an ambulance. He and his colleague Leon Moskatel set out to find some answers.  

So, what did David and Leon do? Well, first, they gathered data on when Uber rolled out in different places across the country. Their goal was to figure out if people started using Uber instead of ambulances when Uber entered the market. They relied on the fact that Uber launched at different times in different places. They reasoned that, with no other trends triggering a shift away from ambulance use, Uber would be the only explanation.  

They used a national database containing information on what happens to patients who call 911, including whether they took an ambulance to the hospital.  If you listen to the pilot episode of this podcast, on marathons and mortality, it’s the same dataset that I used.

With that data in hand, they could see how ambulance use changed after Uber became available in certain areas. 

Oh, and by the way, they only looked at Uber data, not Lyft, the other popular rideshare company. That’s because Uber entered the market first in most major U.S. cities. And although Lyft did take some business away from Uber, the change in ambulance use was really tied to the start of any ridesharing at all, not the start of rideshare competition.

David and his colleague published their research in 2017, in the journal Health Economics.

SLUSKY: So we really found one thing, which is that the relative ambulance rate declined by about 6.7 percent when Uber became available. So the way I would think about this is that if you had a hundred thousand people living in an area and you had a hundred ambulance uses in a given time period, that after Uber became available, you would only see 93 ambulance uses for those hundred thousand people. Every quarter from Uber’s introduction, there’s a greater decline in the ambulance volume. And if you look at this figure in our paper, you’ll see that’s a pretty consistent negative sloped line.

So, most likely, as more people heard about Uber and got used to using it, the more popular it became as a way to get to the hospital. 

SLUSKY: So we put this paper together. And then people come out of the woodwork and say, that’s my story. “I did that.” I used Uber to go to, you know, to go to the hospital because they didn’t want to take an ambulance or because I couldn’t take an ambulance. Or I’m an Uber driver and I have taken people to the hospital lots of times. Or a friend of mine who was the chief medical officer of the emergency management system in Kansas city. And the first responders said, you know, we know this is happening. We see this happening all the time.

The effect of Uber on ambulance use is real, but not huge. But as David sees it, the main message of the data isn’t about ridesharing as a viable alternative to ambulances. It’s about what this shift says about emergency medical care. 

SLUSKY: So let’s start with the metaphor of using a cannon to open a door right? That, like an ambulance, is an incredible mobile emergency room that can bring people back from what looks like the dead. That’s an amazing thing. But that’s not a cheap thing. That is a really expensive, intricate tool that certainly can make all the difference in the world in some cases, but also can be far more than is needed in others. And so that’s the first piece. 

And the second piece is the high cost of calling an ambulance, as we talked about earlier.

SLUSKY: I think part of the solution is a much more continuous spectrum of emergency medical transportation. Right? Cause right now you have very few options. You can call and get an ambulance. You can drive yourself, maybe. But there’s nothing in the middle. What would it be like if you video call 911 from your smartphone? 911 Says, do you consent to the feed from your apple watch? And you say yes, and then the provider now immediately knows your heart rate, can know your other vitals. And then that dispatcher now has a whole range of options, right? Not just an ambulance but can say, you know, we’re going to send an Uber driver, regular driver, regular car.

You can’t drive yourself. Somebody should see you in the next hour, but you’re not bleeding. You’re not in any imminent, imminent danger. Or maybe the next one is. We’re going to send a regular car driven by an EMT with a defibrillator, with an oxygen tank, with a first aid kit, with stitches and maybe we can stabilize you there. Or, you know what, we’re going to send, you know, a car like a cop car with lights and sirens and driven by the EMT with all the things I just mentioned. So that car can run red lights and can speed but it’s not a mobile emergency room with all the bells and whistles and people in the back. Right, so you can imagine like actually trying to fill in that space in the middle so that we can use ambulances where we need them and not use them where we don’t. 

It’s an important point. Maybe the choice shouldn’t just be ambulance or Uber… maybe to fix the current broken system, we could develop some other option that draws on the strengths of each. 

But that’s not the only way that these companies have affected our safety — for better and worse. Coming up: Uber and alcohol. 

*      *      *

So we just talked about how rideshares are making a small dent in the emergency transport market, which I think is a good thing. But are there other ways that these companies are disrupting life in ways that affect our health? Now that rideshares have been around for several years, there’s enough data to see some patterns. 

Michael Anderson is a professor of economics at U.C. Berkeley. One day, a colleague pointed out to him a news story about how drunk driving deaths had decreased in several cities.

ANDERSON: And so we started thinking about what are the possible factors that underlie those declines? And the rapid growth of ridesharing was one of the ones that jumped to our minds immediately. And that’s how we became interested in the topic. 

Michael and his colleague, Lucas Davis, approached Uber directly to ask for data on the number of rides taken in cities across the country. At first, the company refused but eventually, they agreed to share how many Uber rides occurred, over time, in each area. The researchers then overlaid that data with traffic fatalities in the same places. 

ANDERSON: What we found is that the cities in which Uber activity rose the fastest had the highest declines in drunk driving fatalities. 

It’s important to note that this study was done before the pandemic. Obviously, rideshare use changed once Covid hit and a lot of us stopped going out. So the findings that Michael is talking about are really from the Before Times — and the world we’re hopefully getting back to.  

ANDERSON: Going into the pandemic, the average level of Uber activity in a city in the U.S. would reduce drunk driving fatalities by about 6 percent, and overall traffic fatalities by about 4 percent. So the number of lives saved per year would be on the order of a couple hundred. 

Sometimes it’s the driver whose life was saved, sometimes it’s the person they would have hit. Either way, it was a notable reduction. 

Of course, economists like to put dollar amounts to everything, and one of the measures that we often use is called the value of a statistical life. It’s a measure that helps quantify the value to society of avoiding a death. 

The value of a statistical life is often used by the federal government to gauge the benefit of environmental laws, road safety laws (like speed limits), or any other policy that could have an effect on our health. Typically, the economic value of lives saved by a law or regulation is compared to the cost of that policy — for example, a regulation that improves the quality of the air that we breathe, but is costly to some industries.

ANDERSON: The Department of Transportation I believe uses a value of statistical life of $10 to $11 million right now. So if we apply that to our results, we basically find, you know, quote-unquote, savings or — in terms of lives saved, on the order of about $2.5 to $5 billion per year. The value of the life-saving benefits was probably actually larger in economic magnitude than, uh, Uber’s market capitalization was if you just thought about the share of the market capitalization that’s due to U.S. ridesharing. 

Now, since I’m using some economic jargon, there’s another term that economists often use called externalities. An externality refers to any situation in which our actions have effects on others, like choosing whether or not to get vaccinated. Economics teaches us that people tend to do too much of those things that have negative externalities (like smoking in public places) and too little of those things that have positive externalities (like getting vaccinated). In the case of Uber, a drunk person may use Uber instead of driving under the influence. Uber then creates a positive externality because drunk passengers might’ve otherwise driven themselves and caused injury to others. Michael estimates that this positive externality of Uber, which keeps some drunk drivers off the road, could prevent hundreds of deaths a year.  

ANDERSON: That was sort of an unintended consequence that turned out to be quite significant in terms of both public health benefits and also economic gain in some sense, if we think about converting the lives to dollars. As we come out of the pandemic, I think the general expectation is that probably ridesharing will return to at least its previous levels and hopefully continue growing. And then over the longer horizon maybe we start to get more autonomous technologies that further reduce the cost of essentially outsourcing driving when you shouldn’t be driving. so I’m sort of cautiously optimistic that— that the trends we see in, reduced, drunk driving will— will continue. 

It’s a heartening side effect of ridesharing, especially in light of some upheaval triggered by these companies. But not all the studies of ridesharing are about improvements. 

Keith Teltser is an economist at Georgia State University. When he saw research on the link between ridesharing and drunk driving, he had questions. If people were using Ubers as a way to drink safely, then could they also be drinking more because of that? If you’re out with your friends and you already know you’re gonna call a car to get home, then why not have one more drink? That behavior is something that economists call “compensating behavior.” Epidemiologists sometimes call it “risk compensation.” The University of Chicago economist Sam Peltzman described this phenomenon more than 40 years ago in the context of automobile safety regulation. He showed that when it came to driving and auto safety, people tended to offset the benefits of safety features by just driving faster and less safe.  Keith wondered about whether a similar thing – what we now call the Peltzman effect – might be happening with Uber and alcohol use. So he tracked down some data from a couple of different sources.

TELTSER: One was — the behavioral risk factor surveillance system from the C.D.C., where survey respondents can report, “How many days did you consume alcohol over the past 30 days? How many drinks on average did you have per drinking day? Um, how many, uh, binge drinking instances were there?” then there’s another question asking, “What’s the maximum number of drinks you had in a single occasion?” So, this could also measure how drunk or the intensity of drinking that’s taking place. And then, we wanted to figure out to what extent this additional drinking was occurring outside of the home — you know, at bars and restaurants. And so we tracked down the Quarterly Census of Employment and Wages, and we looked at employment numbers and the earnings among employees at drinking establishments and full-service restaurants.

Keith compared all these measures in cities where rideshare companies had opened with cities where they hadn’t. 

TELTSER: So, it was really cool we found roughly a 3 to 5% increase, in all of these measures, following the introduction of UberX., and, we found little to no effect on the extensive margin, meaning,, people who reported drinking at all in the past 30 days. So, it seemed like it wasn’t doing much to induce people to go out and drink who otherwise wouldn’t have gone out to drink, but the people who are going out and drinking they’re drinking more. And then interestingly, we find that the increase was larger among men. It was larger among self-reported students and individuals who reported being aged 21 to 34. Binge drinking instances increased by — by 10% among people who are aged 21 to 34. 

So, no real increase in the number of people going out to drink instead of staying home, but a small increase in how much drinking people were doing when they did go out.  

TELTSER: Maybe each individual person is not drinking that much more, but in aggregate it’s a non-negligible increase.

And what did these extra drinks mean for bars and restaurants?

TELTSER: We find, roughly like 3 to 4 percent increase in employment at drinking establishments, and a similar increase in the total earnings of employees at drinking establishments, which we think is a nice finding and makes sense, right? Because earnings for drinking establishment employees are largely driven by, tips and tips are given in proportion to the amount of alcohol someone’s consuming. So it all seems to hang together nicely. 

The study didn’t turn up any strong evidence that this extra drinking was affecting people’s mental health, and it also didn’t show an increase in the kinds of risky behaviors that sometimes go along with excessive drinking. For example, they found no connection between ridesharing and unexpected pregnancies. 

TELTSER: I don’t think the — the effects have been large enough to cause widespread concern, but I think it’s also worth keeping an eye on to the extent that, it seems like the largest drinking effects are occurring among young adults. Maybe, this additional drinking now translates into negative health outcomes later in life, and we’re just not able to detect that with a short window of data.

Clearly, these rideshare companies are “disruptors,” much as I hate using such a cliche term. 

In fact, anytime my daughter has an idea, like going to get ice cream on a Saturday afternoon, she claims her idea was a disruptive innovation. [beat] Okay, not really, but you get the point. 

Look, in all seriousness, the interesting thing is just how many industries are changing because of Uber and Lyft and all the other rideshare companies. Companies that set out to change the way people hail a taxi are suddenly having these unexpected effects on people’s health. 

Of course, they aren’t the only ones.

There have been a lot of changes in how our health care gets delivered. Getting a vaccine at a grocery store. Having a doctor’s visit over the computer. Even urgent care centers. All of these, at some point, disrupted the status quo. They make it easier and often cheaper for us to get medical care. 

But the question is: which disruptors are worthwhile? And how do we know which ones are making life better? 

Unlike Ubering to the emergency room, disruptions in the healthcare system don’t always reduce how much we spend on health care. 

Let me give you an example: Let’s say you have a sore throat, a fever, you’re a little short of breath. Prior to covid, you wouldn’t go to the emergency room for that. You might not even go to your regular doctor, especially if you couldn’t get an appointment right away. But you might stop into the urgent care clinic right near your house, just for a check-up. And you might walk out with a prescription for antibiotics or a referral for a chest x-ray. The economist Janet Currie has shown that urgent care centers actually raise the overall amount we spend on healthcare for exactly the reason this little anecdote illustrates, but outcomes don’t improve. To put it simply, if the service weren’t there, you know, we just wouldn’t use it. Some people may “downgrade” their care, opting for a walk-in clinic instead of the ER. And that does save money. But many more people use these clinics only because they’re there — and they’re convenient. 

The crucial question is: does this care make us better off? And on that, I think the jury is still out. Healthcare spending may go up without saving any lives, but there’s also the reassurance that we get when we see a medical provider. That’s much harder to quantify.  As economists often see it, if a service is used, then it probably has value.  

Anyway, I hope you enjoyed our discussion today. That’s it for Freakonomics, M.D. this week. You can find links to all the studies we mentioned at freakonomics.com.

Thanks for listening. It would be great if you could give us a review on Apple Podcasts or wherever you’re listening. It helps new people discover the show.

And if you have any thoughts on the show, I’d really love to hear from you. You can email me at bapu@freakonomics.com. That’s B A P U at freakonomics dot com.

*      *      *

Freakonomics, M.D.  is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and People I (Mostly) Admire. This show is produced by Stitcher and Renbud Radio. You can find us on Twitter and Instagram at @drbapupod. This episode was produced by Jessica Wapner and mixed by Eleanor Osborne. Original music composed by Luis Guerra. The supervising producer was Tracey Samuelson. Our staff also includes Alison Craiglow, Greg Rippin, Emma Tyrrell, Lyric Bowditch, Jacob Clemente, and Stephen Dubner. If you like this show or any other show in the Freakonomics Radio Network, please recommend it to your family and friends. That’s the best way to support the podcasts you love. As always, thanks for listening.