Bigger is not necessarily better, according to a recent study of hospital admission rates and outpatient practice size.
Researchers looked at about 1,000 U.S. primary care practices with fewer than 20 doctors to compare their rates of potentially preventable hospital admissions of Medicare patients. They found that the admissions increased along with the number of doctors in the practice. Solo or partner practices had 33% fewer admissions than the groups of 10 to 19, and the middle-size practices (3 to 9 doctors) had a 27% lower rate than the biggest groups.
The study, which was published in the September 2014 Health Affairs, also looked at the practices' participation in medical home models and payment incentives, as well as whether they were owned by physicians or hospitals. ACP Internist spoke with lead author Lawrence P. Casalino, MD, PhD, chief of the division of health policy and economics at Weill Cornell Medical College in New York, about the results.
Q: Does your study suggest that small practices are better at preventing admissions?
A: The results do suggest this. But it's important to remember that this is an observational study. It's not a randomized, controlled trial. We certainly found an association between smaller practices and fewer ambulatory care-sensitive admissions, and also physician-owned practices compared to hospital-owned practices had fewer of these preventable admissions. But we don't know for sure whether there's a causal relationship. Whether it's because they're smaller or because they're physician-owned, or for some other reason we couldn't measure, we can't say for sure.
The other major caveat is that it's easy to take a quick look at this and say really small practices do better than everybody. We were only going up to 19 physicians. It's not like we compared 2-physician practices to Kaiser or the Mayo Clinic or Geisinger. That said, the ambulatory care-sensitive admission rates [in the smallest practices], if you compare them to the national averages, were pretty low. It looks like their performance is good, no matter who you compare them to.
Q: Were you surprised by the results?
A: Yes and no. The general belief, I think, is that larger groups are better, [but] I spent 20 years in a small practice. We started with 2 physicians, ended up with 9, and it was a good practice. On the other hand, I saw some pretty bad small practices. I could see reasons why small practices might do better, but I didn't know if, on average, they would.
Q: What are some of the possible reasons?
A: If it's really a causal relationship, if the results are better because the small practices are doing better, I would think it's because of the very close relationships and mutual knowledge that there can be in good small practices among the physician, the staff, and the patients.
In our practice, we each had our own medical assistant. We spent more time with our medical assistant than we spent with our spouses. In 20 years, I had only 3 medical assistants. We were very tight, we knew each other very well, and we could speak shorthand to each other. The patients knew the medical assistant very well and knew me very well. Pretty much everybody was on a first-name basis. The knowledge helped us assess whether there really might be something wrong with the patient, assess what's the best way to explain things to them, also understand their values and what's important to them. The patient, on the other hand, got to know us and got to see how much they could trust us.
[In my practice] I might have been running between exam rooms and my medical assistant Susan might have grabbed me as I went past and said, “[Patient X] just called.” I might have rolled my eyes, and she'd have said, “I know, I know, he calls almost every day, but today I think there's really something wrong with him. I just don't like the way he sounds.”
Some of the bigger organizations will have people, who may or may not be nurses, [who] have computerized protocols that they'll go through, asking, “Do you have this symptom? Do you have that symptom?,” and they'll give you advice about where you should be seen, when you should be seen. There are strengths and weaknesses to that. One of the weaknesses is you can't say, “I know him; I just don't like the way he sounds. I think we ought to get him in.” [In my small practice,] we'd get him right in and maybe we'd prevent an admission that way as opposed to saying, “Yeah, we'll give you an appointment next Monday.”
Q: Is that typical of a small practice in your experience?
A: Again, we were a good small practice. One of the things about small practices is there's no one really checking on what they're doing, so you can have some awful small practices as well. I'm sure there are small practices in which a physician is seeing 40 patients a day and spends very little time with them and never talks to anybody on the phone, and if they get a call from a patient who feels like they're sick, they send them to the emergency department. And I wouldn't expect those practices to have such a good track record of keeping people healthy enough to avoid admissions. But it seems like on average there weren't so many of those.
Q: The practice pattern you describe sounds similar to the goals of the patient-centered medical home (PCMH), yet your study didn't find the PCMH model to be associated with fewer admissions. Why?
A: In theory, you can do a patient-centered medical home in a practice of any size. In practice, if you're trying to get recognized by the NCQA [National Committee for Quality Assurance] as a medical home, there's actually a lot of paperwork. It tends to happen more in larger practices that have nonphysician staff who can do that kind of work.
The NCQA's recognition for the medical home has been criticized for being very much about things that are relatively easy to measure and check off boxes for—structural things—though NCQA is trying to improve things. A real medical home would consist of the old-fashioned, really intense, personalized care that good physicians have always done for their patients, as well as organized processes, such as those necessary for NCQA recognition, to systematically improve care for the practice's population of patients.
The studies of medical homes so far are equivocal. There isn't consistently strong data to say that they really improve things that much. That's not to say that they never will. In our study, where we had a partial measure of medical home-ness, so to speak, other things [being] equal, practices that had higher scores on our medical home index did not have lower rates of ambulatory care-sensitive admissions. Ours was one of the group of studies, of which there are quite a few, that didn't show any benefit from medical homes, but it wasn't really perfectly set up to test that. I believe that medical homes, if properly done, will reduce ambulatory care-sensitive admissions, but they can't just be about checking off boxes. They have to have the close personal relationships that the good small practices have.
Q: You also looked at the association between payment incentives and admission rates. What conclusions can you draw?
A: We didn't find that payment incentives had an effect. Similarly, hospital-owned groups tended to have higher medical-home scores, because they have more resources to fill out the paperwork and get the [information technology] systems in shape, but hospital-owned systems did worse, not better, in our study. We didn't find any effect, but do I think payment incentives can't be effective? No, I think they could. We had relatively crude data on payment incentives. We didn't know too much about how intense the incentives were or exactly what the incentives were for. So it's slightly disappointing that we didn't find an effect of them, but I wouldn't say, therefore, we know payment incentives don't have an effect. The literature on the whole is pretty mixed about this. The incentive programs that there are so far, many of them aren't potent enough to change outcomes very much.
Q: Can you draw conclusions from the study about whether smaller or larger practices should be particularly encouraged?
A: We now understand that it's important for practices to put systematic processes in place to improve the health of all of their patients, not just the ones who happen to come in and not just while they're in the office—making sure that people do come in that ought to and giving everybody what they need in between visits, whether from nurse care managers or e-mail reminders. It is hard for small practices to put those processes in place, for all kinds of reasons. Ultimately, I don't think that small practices that don't have any such processes will be able to match larger practices that do, and at the same time, have really good physicians.
It is possible that small practices will be able to share resources, like nurse care managers for patients with congestive heart failure, for example, that no one small practice could afford to have by itself. By sharing those resources, they might be able to have the best of both worlds—the small practice setting, but the resources to deliver all these organized processes to improve population health. Some independent practice associations are trying to provide that. Then there are quite a few [accountable care organizations] that are based in a hospital-owned group or even a really large medical group but are based on looser networks of doctors. We'll have to see how they perform.
I think some large organizations understand that there are benefits to human scale, and they try within their huge physician group or organization to make it so patients and physicians and staff can have more human-scale relationships with each other, but it's harder to do that in a big organization.
Q: What lessons should individual physicians take from your findings?
A: If you like being in a small practice, this gives you some ammunition to say, “Hey, we can do a good job.” If you're a physician in training who is considering where to work, you can at least think about small practices.