Quality may actually improve in some areas, his research has shown, but any cost savings may be offset by greater utilization. Ateev Mehrotra is an associate professor at Harvard Medical School and a member of Managed Care’s editorial advisory board.
You wrote a piece with Jacob West called the “The Future Ecology of Care” that was published in Annals of Internal Medicine earlier this year. It refers back to the publication of a piece 50 years ago called “The Ecology of Medical Care.” You and West sketch a picture of how much health care has changed from simply going to the doctor or not and all the choices we have now.
Ateev Mehrotra, MD
Yes, there is a plethora of choices we have now in terms of where we can get care. A lot of my work has been on low-acuity conditions, such as sinusitis and UTIs, and how many different places—how many different modalities—you can go to get care, and which ones are focused on trying to get you in and out and being as convenient as possible. Urgent care centers, retail clinics, direct-to-consumer telemedicine, e-consults, kiosks. It’s really interesting to see all these options emerge.
Why has it happened?
In our interviews with patients who have used these care sites, the number 1 reason, the number 2 reason, number 3 reason, is always convenience. Patients describe difficulty getting in to see their primary care physician. “I call my doc, it’s Monday, they said the next appointment is Thursday; I wanted to get this taken care of right away” and “I work days, it’s hard for me to get off work.” That kind of language. And with retail clinics, people say, “I can get in here, and then, you know, get my care taken care of right away, and get the prescription at the pharmacy. It’s just a lot faster.”
And there are some general trends. We’re seeing greater difficulty getting into primary care providers. And I do think there is probably a societal shift in terms of what we think are reasonable wait times. A patient wait time of a day or two is probably just fine from a clinical perspective in many cases. But in an era of 24-hour, 7-days-a-week grocery stores and banking, of shopping whenever we want on Amazon, what patients have come to expect, I think, is that health care will similarly fit their needs.
It’s also a bit of a neglected market. If you are a CEO of a large health care system and you’re thinking, “Where am going to get my revenue growth,” this kind of convenience care is not something that you are going to think of. It’s a very low-margin business, and it may not be what you’re particularly good at.
It means entrepreneurs can get into this.
Right. As they say in business school, the incumbents have sort of neglected this area, so new entrants can come in and really take advantage of that.
And I don’t want to deny one other potential thing that’s driving these options, and it’s the dollars—specifically, the dollars from the perspective of patients. With higher deductibles, people are looking for different options, and these convenience-care options can be less than three quarters of the cost of going to the doctor’s office.
Do you think these convenience-oriented options are at odds with another push, which is the patient-centered medical home? It seems like the convenience-oriented options are like going to a 7-Eleven, whereas the patient-centered medical home conjures up the corner grocery store where people know you. Do you see these two as being at odds with one another?
Yes and no. I’ll start with the no part.
The patient-centered medical home is very much driven—one of the key tenets—in trying to make it easier for patients to come in and get care. Things such as open-access scheduling. So, convenience is part of the patient-centered medical home.
Another tenet of the patient-centered medical home is team-based care. Why don’t we have other providers—nurse practitioners, physician assistants, others—provide much of this care?
You can see why they share some fundamental ideas.
And how do they differ?
The convenience-care options really are a different model from the way we have been delivering health care, not just the patient-centered medical home.
The patient-centered medical home is more like big department stores, where you go for everything, and the convenience-care options are like the rest of the shops in the mall where you go for your candles, underwear—everything is sort of specialized.
What’s the point of having the one place where you get all your care? And I think there are three things that are useful to kind of disentangle.
The first is relationship continuity—the idea that this is the place where, kind of like Cheers, everyone knows your name and understands who you are as a person. And I think that patients value that idea, and I think providers value it. Why do we care about that? Well, we care about that partly because we believe that if the proverbial something hits the fan, having that baseline relationship will build the trust. Therefore, when we navigate things that are really tough, there is that underlying trust.
I think there are reasons to wonder whether this relationship continuity is overemphasized. In the majority of situations when someone has a new problem, an acute problem, they don’t see their primary care physician. If they go to a primary care practice, they’re seeing some other doc or some other NP. They’re going to an urgent care center and all these other sites. Or to a specialist. So I wonder if we’re viewing the current health care system through rose-colored glasses.
The second reason we value care in one place is information continuity. This is the idea that, if I’m taking care of a patient, it’s useful for me to know what other doctors and providers have diagnosed and treated—the medical history, which the patient might forget to tell me about.
The informational continuity I think is important, and the idea is that if you go to a patient-centered medical home, you’re more likely to get that.
But if we think about it really, is that informational continuity only available in a patient-centered medical home? Take, for example, the Cleveland Clinic, which has a partnership with CVS’ MinuteClinics and, through a partnership with American Well, is providing telemedicine care—not by Cleveland Clinic providers but by other providers.
But from what I understand, those records go right into the electronic health record for Cleveland Clinic. So I don’t know whether a patient-centered medical home is really necessary for informational continuity or if it is just easier than overcoming the interoperability issues.
The last point is that if I’m taking care of a patient with a lot of chronic illnesses, I need all that information. But for the types of problems we’re talking about here—for, say, a generally healthy 45-year-old woman who might come to me with a urinary tract infection—she can tell me most of what I need to know. Might she miss some allergies and so forth? Yes. I think there is that concern. But most of the time, with low-acuity problems, information continuity may not be as critical.
Some of the primary care physician groups are fighting or raising lots of questions about convenience medicine. Do you think those concerns are valid, or are they just protecting turf?
For the last 10 years, I’ve been studying these concerns or criticisms. First, I’ll tell you about what I think has been disproven and then move to what I think has more merit.
What we have found is that the retail clinics provide equal quality care to physician offices and, in some cases, even a little better care.
The first [criticism] is that the quality of care is bad. They say, “Hey, those nurse practitioners at those retail clinics don’t know what they’re doing; they’re undertrained; they’re just going to provide antibiotics unnecessarily, and they’re going to misdiagnose, and patients are going to go get care anyways from another doc, etcetera.”
What we have found doing research using charts, using claims, is that the retail clinics provide equal quality care to physician offices, and in some cases, even a little bit better.
A little bit better at what?
Providing a good patient experience and concordance with guidelines.
They usually are a little more consumer-friendly, and they take more time and that might be the greatest driver of the patient’s experience.
If patients have a cold, the retail clinics very rarely prescribe antibiotics compared with doctors.
The most common reason people visit a retail clinic is for acute respiratory illnesses—sinusitis, bronchitis, ear infections, etcetera. For some diagnoses, antibiotics are necessary and may be helpful, and for others it’s clearly not indicated. The retail clinic providers are much more consistent with the guidelines. If patients have a cold, the retail clinics very rarely prescribe antibiotics compared with doctors. And they’re more likely to use rapid testing for strep, which is a good way of limiting antibiotic prescribing. So, those are some examples of where the quality is better.
You also hear—it was more evident five to 10 years ago—that “retail clinics are just going to drive up health care spending because the patients are going to be misdiagnosed, and they’re just going to end up going to another provider and therefore, you’re going to have double utilization.”
But we find that no, retail clinics on a per-episode basis are about 30% to 40% cheaper than a doctor’s office, and the rates of follow-up visits are about the same.
What about your Health Affairs article in March that said that for low-acuity conditions, utilization went up? And I think you found spending also went up?
This is an issue that hadn’t been raised by the medical community—but I was really concerned about—which is that retail clinics per episode are cheaper, but maybe there are more episodes. That’s what we found in that Health Affairs study, which was that because it was more convenient, people got care more often. This is maybe one of those obvious findings when my father says, “And they pay you money to study this?”
I agree with my father that it may not be surprising, but it was useful to highlight.
The last area where I think there are concerns about retail clinics is the issue of continuity of care. And we already talked a little bit about that. But we find that patients who go to retail clinics are less likely to go to doctor’s offices after that. They like the experience, and they keep going back to the retail clinic. So, if we measure continuity of care measures, they are worse among retail clinic users. Now, the question is: So what?
That’s what I was thinking.
Right. So is continuity itself important? Some people argue it is, because if you do big, national studies in patients who have higher continuity, they have better outcomes and lower spending. But you have to look at the causal chain. Is continuity important because doctors will make sure that people will get preventive care? What we found is that retail clinics users were just as likely to get preventive care.
Well, maybe continuity is important because if you go to your primary care physician’s office and you have diabetes and you go for strep, they’ll take care of the strep but say, “Hey, wait a minute. John, what’s going on with your diabetes? Let’s make sure you’ve got X, Y and Z tests” and so forth.
This is an area where what we know is very limited because when we looked at claims for retail clinics, we didn’t have a lot of diabetic patients. But we didn’t find a problem among the limited numbers of patients for whom we did have claims data.
I think this is something that really does merit future research. Does getting care at a retail clinic have a negative impact on those patients? But the majority of patients who go to retail clinics do not have a chronic illness. From what we have seen, they haven’t been a key part of the population that the retail clinics serve.
Telehealth is also about convenience. You are coauthor of a recent study in JAMA that found that telehealth meant better access to dermatology care for Medicaid beneficiaries. But there was another study, a secret shopper study, published in JAMA Dermatology at about the same time that found that telehealth for dermatology resulted in missed diagnoses of important conditions like nodular melanoma. Some of the analysis that we’ve been going over for retail clinics—have you applied that to telehealth? And do you see things playing out a little differently for telehealth?
We’ve done many studies of telehealth, including several studies on Teladoc.
We have looked at quality in the same way we have looked at retail clinics. The signal hasn’t been as clear-cut.
The direct-to-consumer telemedicine companies had a similar rate of follow-up visits as primary care practices and similar rates of overall antibiotic prescribing.
But there were a couple areas of concern. They had a slightly higher rate of broad spectrum antibiotic prescribing.
And there was an issue that we’ve never had the problem in our health care system before, which is undertesting. We’re always talking about overtesting in the emergency department or the doctor’s office. Here, we found the opposite. And it makes complete sense. You have a video conference, or you’re on the phone, and you’re two states away and you don’t even know how to get a test for a patient. Therefore, physicians were less likely to order tests.
In some cases, that could be a good thing. For patients with back pain, you could say consumer telemedicine companies are doing great because they do not overuse CTs and MRIs for back pain. But in other areas, such as strep or UTIs, we would consider undertesting bad care.
I was involved in a different secret shopper study, which was led by folks at UC–San Francisco. And we found, very similar to a recent secret shopper on online dermatology companies, a number of misdiagnoses. Both of these studies have found in certain cases egregious quality issues. Obviously, if I was a patient, I’d be like, “Whoa, should I really go get care from these places?”
On the other hand—and I think it’s missed in this debate—is the work that has been done that shows really egregious quality issues for in-person visits, too. So, we are left with this really awkward situation, which is, “Oh, my God, the care may not be good here.” But is it worse, or better, than what we see in an in-person office?
Earlier in our conversation, I mentioned that the antibiotic prescribing rates for visits to retail clinics are similar to those for in-person visits. And you might say that’s great, and reassuring to the average American. But, you could also flip that around and say, “That’s horrible.” Because at in-person visits to a doctor, we believe that over half the antibiotics prescribed are unnecessary. Differences in how you frame it that might help explain why people are interpreting the data very differently.
So the issues with telemedicine are sort of the inverse of those of in-person, nonretail health—overtreatment versus undertreatment. That’s what I’m gathering.
As far as testing is concerned, I would agree with that. On the physician and emergency sides, it’s overtesting, and on the telemedicine side, it’s undertesting. It’s just so darn easy in the emergency department to order a CT scan. In a telemedicine visit, it’s really hard.
Is there a fix to the quality issues with telehealth?
The business model is typically built on the idea that we want to have one interaction, take care of the problem, and move on. It’s very transactional, if you want to call it that. And testing makes things more complicated because then you’ve got to get the test results, and then act upon it, and it creates a whole new layer of logistics.
But I feel like they need to develop methodologies or systems to actually be able to address and make it a lot easier for their docs to order tests. That’s going to mean better care. Because in modern medicine, tests do matter, and are critical in certain circumstances.
There’s also a trend toward bringing back the house call for very sick people. Is that part of this convenience notion? Do you see it fitting into this broader trend of convenience?
These convenience-care options are well suited for the large fraction of our population that’s overall relatively healthy. For those patients, I don’t really see the value of the house call. And the reason is that the business model is very problematic. You see companies that are starting to emerge in this market, and if you pay for a doctor to schlep in a car to a patient’s home, that’s extremely expensive.
Telemedicine and retail clinics are efficient either because the salaries of the employees are, on a per-hourly basis, a lot lower, or you’re being a lot more efficient with a physician’s time. That allows you to charge a lower price, which is critical in this era of high-deductible health plans.
I think it is a very different story for our sickest populations who have multiple comorbidities, and we’re trying to do whatever we can to really help them stay out of the hospital and take their meds. And in that context, getting a doc to go visit the patient, even though it’s a lot of money, may be worth it, because they get to see the patient in their home and really help them out.
Have you gone to a retail clinic yourself or used a telemedicine service?
I have not used a telemedicine site, but I have gone and encouraged my family members to go to a retail clinic.
And what was your experience like?
It was convenient. I had to get in quickly, and it was very easy for me to get in. And I appreciated that.
Why did you go?
One time, I needed a vaccine, and I needed it quickly. Another time my wife was worried about sinusitis, and she went in for a quick visit.
Editor’s Note: Transcript of interview was edited for length and clarity.