When it comes to antibiotic stewardship and clinical care, Brad Spellberg, MD, just wants to see the data.

Spellberg, chief medical officer at the Los Angeles Medical Center, has been a leading proponent for the responsible use of antibiotics for well over a decade. During that time, one of his primary efforts has been to convince healthcare providers that the antibiotic courses they prescribe for many of the indications they commonly see are far too long. He’s also been an advocate and chair of the steering committee for WikiGuidelines, a movement by healthcare professionals to construct clinical care guidelines based on reproducible, high-quality evidence.

Spellberg, an internal medicine and infectious disease (ID) physician, advocates for these ideas in part because, in his view, too much of medical practice isn’t based on high-quality data. He relays a story about when he was a resident and there were national guidelines, based on observational studies, recommending that post-menopausal women should be treated with hormone-replacement therapy (HRT) to reduce breast cancer and heart disease risk. Spellberg followed those guidelines, but in 2002, a randomized controlled trial by the Women’s Health Initiative showed that HRT increased the risk of breast cancer and cardiac events.

That was the first of several indications, Spellberg says, that something wasn’t right with the guidelines process.

“How much harm did we cause?” he asks. “You start to realize the guidelines process is not based on data. It’s based on people’s opinions, and they can be highly flawed.”

CIDRAP News recently spoke with Spellberg about how following the data has shaped his opinions on antibiotic stewardship, clinical care guidelines, and the antibiotic pipeline. The following excerpts have been lightly edited for length and clarity.

CIDRAP News: I’d like to start this conversation with the concept of “shorter is better,” because this is something that you’ve been a longtime advocate of. You’ve written about it, you talk about it a lot on social media, and you keep a running list of randomized clinical trials that support this idea on your website. Why is the concept of “shorter is better” so important to you? And is this something that’s always been obvious to you since you started as an ID physician? 

Brad Spellberg: No, actually. “Shorter is better” was branded in an editorial I wrote [in JAMA Internal Medicine] in 2016, but the concept actually dates back to 2008. Actually, it was Lou Rice, who at the time was the chief of medicine at the Cleveland VA Medical Center (he’s now chair of medicine at Brown), who first gave a keynote lecture on the topic at a national meeting and published an article of the lecture in 2008. He said in that lecture that it’s really hard to get doctors not to prescribe antibiotics. You don’t win a lot of those arguments.

Rice said that maybe we could be more effective in stewardship if we focused on duration of therapy, as opposed to not giving the drugs at all. And that was a brand new concept. I was not familiar with it at all when Lou sort of called that out. And the more I practiced after encountering Lou’s lecture, the more I realized he’s right. It’s really hard to get doctors not to prescribe, but actually, it isn’t that hard to say “Hey, did you know that there are randomized controlled trials showing you only need 5 days for pneumonia? Oh, really? Okay, yeah, we can do 5 days.” You win more of those battles. 

And the bottom line is you end up prescribing fewer antibiotics. So it’s a way to be more successful at stewardship than historical methods have been. 

CIDRAP News: More successful, say, than just telling patients they shouldn’t get antibiotics because they have a virus?

Brad Spellberg: That’s correct. And it’s more successful for two reasons. One, as we’ve said, you don’t win a lot of those arguments. The patients will just say “yeah, but how do you know it’s a virus?” It also doesn’t help. What about the patient who actually did need the antibiotic? They actually had a bacterial pneumonia? You weren’t ever going to tell them they didn’t need antibiotics, and they were going to get 14 days. 

If you focus on duration of therapy, you help the patient who needed the antibiotic but didn’t need 14 days. And yes, for the patient who didn’t need an antibiotic, it would have been ideal if they didn’t get any, but they’re still better off getting 3 days than 14 days. So the way I phrase this is that if you’re going to give antibiotics to people who don’t need them, at least do us the courtesy of doing it for a brief period of time.

CIDRAP News: In that 2016 JAMA Internal Medicine editorial, you delved into the history of the traditional antibiotic duration. I’m hoping you can expound on that for people who haven’t read that editorial. Why have antibiotic courses traditionally been 7 to 14 days?

Brad Spellberg: That was the editorial that first used the phrase “shorter is better.” I looked at references, I looked at national guidelines, and—with tongue in cheek—what I said was that I traced those references back in history to the fact that in 321 AD, Constantine the Great decreed there would be 7 days in a week, and that’s why we give 7- to 14-day courses of antibiotics. If good old Constantine created a 4-day week, we’d be giving 4 to 8 days of antibiotics, right? And we’d also only have 2 workdays before the weekend. And that would have been really nice. But that’s not what he said. He said 7 days a week. So we give 7 to 14 days’ worth of antibiotics. 

Now, to be fair though, not everybody uses Constantine units. There are 10-day [antibiotic prescription] believers, which is neither 7 nor 14. It’s not based on Constantine. Where does 10 days come from? The 10 days turns out to be based on the number of metacarpal bones that evolved in the hominid hand, which has led some of us to speculate. We wrote in an article in Open Forum Infectious Diseases that the world would have been a better place if we had instead evolved as three-toed sloths. 

And I gave that lecture at the Massachusetts ID society 2 months ago, and a cheeky physician who was in attendance decided to mail me a gift. So I now have a three-toed sloth on my desk. And he says his name is Steven. And I should tell people that Steven says that you should give 3 to 6 days of antibiotics instead of 5 to 10 days.

CIDRAP News: I haven’t checked your chart lately, but how many indications are we talking about? How many RCTs are out there that show shorter is better?

Brad Spellberg: Twenty-three diseases, and I think 150 or so randomized controlled trials.

CIDRAP News: But shorter isn’t necessarily better for everything, right?

Brad Spellberg: We should first say that shorter is as effective, with fewer side effects and less selection for resistance. And that’s what makes it better. It’s not that shorter is more effective. It’s the same efficacy, but safer and with less selection for resistance. 

There are several exceptions that are clearly established based on large randomized controlled trials. For example, prosthetic joint infections where you retain the device. In the DATIPO trial, 12 weeks was superior to 6 weeks—definitively superior. There is also a large randomized controlled trial of otitis media (middle ear infections) in children under the age of 2 where 10 days was superior to 5 days. Note that in other trials in children over the age of 2, that is not the case. 

But it’s established in children under the age of 2. If you give patients with strep throat oral penicillin three times per day, 10 days is superior to 5. However, if you use oral cephalosporins or if you use oral penicillin four times per day, it’s not clear that 10 days is superior to 5. So there are some specific examples where it’s been established that shorter is not as effective.

CIDRAP News: In another piece in Annals of Internal Medicine in 2019, you noted that short-course therapy remains underused by clinicians. Now, that was 6 years ago. Is that still the case, and if so, why?

Brad Spellberg: It’s getting much better. The American College of Physicians put out a position paper on short-course therapy in 2023, I think, so it’s 2 or 3 years old now. And they said it’s standard-of-care for several indications, like cellulitis, kidney infections, and other stuff. So it is now becoming much more commonplace. 

The way I phrase this is that if you’re going to give antibiotics to people who don’t need them, at least do us the courtesy of doing it for a brief period of time.

I just got off medicine [rounds] last week and we were talking about durations of therapy, and the residents on the team were shocked to learn that we used to treat pneumonia for 14 days. They’re like, “Why would you do that? Five days is adequate.” So we’ve turned the corner. They don’t even know about the longer durations of therapy anymore. It’s going to take time for it to reach full penetration, but we’re past the tipping point. 

CIDRAP News: This gets to a broader question. How would you characterize the current state of antibiotic stewardship in the United States? Is the US healthcare system doing a better job of promoting judicious use of antibiotics?

Brad Spellberg: We are doing a better job of promoting the concept. For example, Medicare has a performance measure that you have to have a stewardship process in your hospital to meet acute-hospital status. So promotion, yes; the word is out more effectively. 

But we’re not doing it better. We’re not actually using fewer antibiotics. And that’s because the way we’ve gone about this is incorrect. We keep acting as if it’s an education problem. We just make people aware and teach them. It’s not. It’s a fear problem. You don’t combat fear with rational knowledge. Fear has to be combated with psychology. One of the reasons I like “shorter is better” is its psychological approach. It recognizes that you’re going to lose a lot of those arguments, so you’re going to need to find a different kind of argument you can win. 

We need other approaches that people like John Bartlett (before he passed) were talking about for years that we just have never been able to get the IDSA [Infectious Diseases Society of America] to get serious about lobbying for. For example, how did we get hospital-acquired infection rates to actually decline in the United States? I mean, they were going up and up and up, and suddenly they started turning the corner and coming down. 

It’s very simple. You put a pay-for-performance measure into place. You tell hospitals, “We’re going to financially penalize you if your C diff rates go up. We’re going to financially penalize you for ventilator-associated pneumonia.” And all of a sudden hospital CEOs start going, “Whoa, whoa whoa—we need to figure this out.” And then they tell their surgeons and their internists that they better get these rates down. And they actually put teeth into their infection control strategies. The exact same principle should be applied to antibiotic stewardship. 

We keep talking to department chairs and division chiefs. That is not who controls medicine. You need to get to the business of medicine and align the business of medicine with the outcome you want to achieve. And if you do that, you’ll not just promote the concept, you’ll actually make a difference.

CIDRAP News: I want to talk about WikiGuidelines. You write on your website that wiki guidelines are a “new way to construct clinical guidelines that seeks to incorporate the humility of uncertainty.” What do you mean by that? And how are wiki guidelines different from guidelines published by medical societies or the IDSA?

Brad Spellberg: Many guideline recommendations are strong recommendations based on no to weak evidence. Why are you doing that? Why are you creating standards of care based on your opinion, not based on high-quality data? If you want to treat a patient at the bedside, you need to make a decision. If you want to compel every doctor in the United States to do what you say, you better have some damn high-quality data behind it. It’s extremely arrogant of people to put out a guideline which creates a standard of care based on no data. 

If you want to compel every doctor in the United States to do what you say, you better have some damn high-quality data behind it. It’s extremely arrogant of people to put out a guideline which creates a standard of care based on no data. 

What you should do is be honest. Let’s be honest about what we actually know and what we think but don’t actually know. That’s where the humility of uncertainty comes into play. And with WikiGuidelines, we don’t use the grade-level evidence. The charter says, explicitly, that you can only make a recommendation if there are at least two prospective studies, one of which is a randomized controlled trial, showing that what you’re recommending is correct. If you don’t have that level of evidence, you don’t make a recommendation. What you do instead is describe options: Some doctors in this setting like to do this. Other people like to do this. There are some retrospective data that suggests this. You should be aware those data are flawed. 

Talk about it with your patient and be honest. And don’t pretend that there is a consensus that experts have based on no data. 

CIDRAP News: One topic you’ve written about extensively is the antibiotic pipeline. I’ve been covering antibiotic resistance, stewardship, and development for CIDRAP News since 2016, and, during that period, one of the big stories has been the dearth of new antibiotics and the related lack of investment in antibiotic development. Many people have argued for the need for government to get involved. And that’s where you see things like the PASTEUR Act. But you have argued that the antibiotic pipeline isn’t in as bad shape as many believe and that government funding models aren’t necessarily the answer. Let’s start with that antibiotic pipeline. Is it not as weak as has been portrayed?

Brad Spellberg: We just put out an editorial in Health Affairs Scholar, which is worth a review because it has the data in it. I don’t want to talk about people’s opinions. Let’s talk about the data, which is objective reality. 

In 2004, when we published the first article that documented the collapse of the pipeline of antibiotics, there were six antibiotics in development across all biotech and pharmaceutical companies. We searched all their pipelines, we found six. As of 2021, according to a report from the Pew Charitable Trusts, there were 42 antibiotics in development. Forty-two is a lot more than six. There are way more antibiotics in development today than there were 20 years ago. That’s an objective fact. That’s not subjective. You can say the earth is flat; it doesn’t make it so. This is an objective reality. 

Now what about approvals? We published the graph of FDA [Food and Drug Administration] approvals of antibiotics in the same editorial. Approvals petered out in the 2008 to 2012 period. There were three antibiotics approved during those 5 years. We had three antibiotics approved in 2024 alone. We have had way more FDA approvals despite the COVID pause, and there were no antibiotics developed for 3 or 4 years during COVID. Despite that pause, we’ve had way more antibiotics approved in the last decade than during the decade before that. 

The fact is, the pipeline is much better positioned today than it was 20 years ago. You can say, “Well, that’s not enough.” Okay, I’ll next ask you what is enough? What does enough mean? How many do you need? That’s a separate question. Objectively speaking, the pipeline is much better today than before, both in terms of drugs in development and in terms of drugs approved.

CIDRAP News: What’s the argument against government funding models like the ones being pursued through the PASTEUR Act or the United Kingdom? They’ve been called subscription-style models or Netflix-style models. What’s the argument against boosting antibiotic development in that way?

Brad Spellberg: We go through that in detail in the Health Affairs Scholar article, and the argument is quite simply this: It makes sense if you want to encourage people to use drugs. Take, for example, hepatitis C and HIV, where if people aren’t on therapy they will continue spreading those diseases to others, and you really want to get them on therapy as much as you can. Now, the drugs are expensive, and so there are states that have passed Medicaid subscription laws where the [pharmaceutical] companies allow them to use as much of those drugs as they want for a flat monthly or annual rate. That’s a win for the customer in the relationship. Think about it this way: You subscribe to Netflix, and you watch 1,000 hours a month of Netflix. That’s a good deal. You’re paying a flat rate and you can watch as much material as you want. 

Objectively speaking, the pipeline is much better today than before, both in terms of drugs in development and in terms of drugs approved.

Antibiotics are exactly the opposite. I don’t want you using all the antibiotics on everybody all the time. I want you to use as little antibiotics as possible. Now, let’s say you subscribe to Netflix and you never watched it. Would you keep paying the Netflix monthly subscription fee? No, you would not, because that would be a bad deal. The subscription model makes sense to the consumer if you want to encourage as much use of the drug as possible. It does not make financial sense to the consumer if you want to have as little use as possible. 

It does make good sense to pharma, who otherwise would sell almost no antibiotics, and now they have a guaranteed flat amount of money. This is a great deal for pharma. It is not a good deal for the taxpayer. That’s the point.

CIDRAP News: Another topic I know is important to you is the future of the infectious disease workforce. How would you characterize the current environment for ID practitioners?

Brad Spellberg: Again, I like to use data. By two objective markers, the ID specialty is not thriving. First, ID physicians continue in annual survey after survey to make less money than general internists. So you spend 2 more years in fellowship training to end up making less money. That is not a way to attract the top talent in the graduating classes. You’ve got to figure out how to fix that. The second marker is that ID fellowship spots simply aren’t filling. I think it was 40% of ID fellowships this last cycle that did not fill, and some went completely unmatched. Those are two signs of a sick specialty.

People have been talking about this for more than a decade now. In fact, I think there was an article in the New England Journal of Medicine in 1978 or ’79 which was predicting the end of ID as a specialty. So this is actually not a new conversation. It was actually the HIV epidemic that staved off the end for 15 or 20 years, lured people in for other reasons. 

But now we’re sort of back to that. And if you go back and read that editorial, it was remarkably prescient. The exact problems that that person was writing about at that time are the problems we face today. Reimbursement is low. You have to go to multiple hospitals to try to make a living. You are a loss leader in an academic environment. It’s the surgeons that bring in the money. Your only job is to support them. You’re second class. 

None of those structural elements of our health system have changed. So the pharmacoeconomics, or the business models, have not changed. And until somebody actually changes the business models, this is not going to get better.

CIDRAP News: Ten years from now, what do you think, or what do you hope, the conversation around antibiotic resistance, stewardship, and antibiotic development sounds like?

Brad Spellberg: What I’d like to get to is a point where we actually are talking about data. Let’s look at the actual numbers and talk about actual ways to adjust performance. Not superficially but actually having pay-for-performance measures in place that actually encourage people to do the right thing and discourage people from doing the wrong thing. That is the essence of quality-improvement work.

I’m not inventing this wheel. Every quality improvement book in the world says if you want to change an outcome, you redesign the process to sort of incentivize the thing you want done and disincentivize the thing you don’t want done. That’s what we need to do.