COVID-19 Resource Center

COVID-19 Resource Center


ATS. Guidance document on COVID-19. Available at: Accessed April 6, 2020.

Disclosures: Wilson reports no relevant financial disclosures.
April 09, 2020
8 min read

Q&A: Insight into ATS guidance for COVID-19


ATS. Guidance document on COVID-19. Available at: Accessed April 6, 2020.

Disclosures: Wilson reports no relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

As health care providers across the United States grapple with the unknowns of treating a novel disease, management of patients with COVID-19, especially those who are critically ill, has varied widely. Therefore, on April 6, the American Thoracic Society released interim guidance to help clinicians standardize care.

During a conversation with Healio Pulmonology, Kevin C. Wilson, MD, professor of medicine in the division of pulmonary, allergy, sleep and critical care medicine at Boston University School of Medicine and chief of guidelines and documents at the ATS, discussed the process for developing the guidance, the reasoning behind specific suggestions and potential questions for future updates. – by Melissa Foster

What prompted you and your colleagues to draft the interim guidance at this time?

Wilson: COVID-19 is a serious, new and emerging disease about which not much is known. At present, there are no randomized trials and very few published studies. Management of the disease varies widely from institution to institution because there is so much uncertainty about the appropriate course of action. We saw this as an opportunity to standardize management, which is known to improve quality of care and facilitate research.

Wilson Infographic

How did you approach drafting the guidance?

Wilson: To be clear, this is not a clinical practice guideline for which we normally would conduct a systematic review of the literature and then make recommendations. We did not do that in this case because we knew a systematic review would not yield results, as the disease is still too new. It is important to draw the distinction here because I do not want to overstate or embellish our findings or our work.

For this guidance, we used the previously developed Convergence of Opinion on Recommendations and Evidence (CORE) process, which is a specific type of modified Delphi process. Basically, it is a survey in which we ask a question in a very specific way and have participants select one of the following: a suggestion for, a suggestion against or no suggestion. We then collect the totals and redo the survey a second time during which participants are able to see the results of the first round. If more than 70% of participants select an option, then that becomes what we call a consensus suggestion.

Again, we should bear in mind that this is consensus guidance that is largely based on clinical experience, indirect evidence and limited direct evidence. I don’t want to imply that it is anything more.


Most of the suggestions focus on patients with COVID-19 pneumonia. Why is that?

Wilson: Everyone on the task force is very involved in clinical care; they are people on the front lines and have very limited time at this point. To secure their participation, I promised them that I would limit the size of the survey to 10 questions, and we saw the most pressing need was how to treat the sickest patients.

Additionally, by the time we started the project, it was well known that a fairly sizable proportion of patients develop only mild, self-limited disease. For those patients, there is more time to let evidence accumulate and let trials be completed.

The guidance does not make a suggestion for or against the use of hydroxychloroquine for outpatients with COVID-19 or hospitalized patients with COVID-19 but no evidence of pneumonia. However, it does suggest that it be used for hospitalized patients with COVID-19 pneumonia on a case-by-case basis under certain conditions. Can you explain how the task force settled on that suggestion?

Wilson: Given my interest in guideline development, this was actually the most interesting part of the whole project for me personally. For this question, we found that most participants fell into two groups. One group felt strongly that we should not be using hydroxychloroquine without randomized trial data because we might be giving an ineffective therapy to patients that could also be harmful and potentially deplete the supply from patients who need it for other conditions, such as lupus. The second group was less conservative in their approach. They viewed COVID-19 pneumonia as severe and potentially life-threatening and saw hydroxychloroquine as an intervention that has a low risk for harm and also has the potential to work. In that situation, the second group believed it is acceptable to use it as treatment while we await more definitive data. That second group’s approach is not unique; it is aligned with the GRADE working group guidance for making recommendations in the context of very low-quality evidence.

However, we saw that participants in the more conservative group changed their opinion as the degree of illness worsened. So, with outpatients with COVID-19 or with inpatients with COVID-19 but no evidence of pneumonia, we did not have enough participants from the first group cross over to the second group to make a consensus suggestion. However, when we addressed whether hospitalized patients with COVID-19 pneumonia should be treated with hydroxychloroquine, there was substantial crossover yielding more than 70% agreement and we were able to make a consensus suggestion.

Interestingly, I found out later from speaking with our participants after the survey that a number of participants who remained in that more conservative camp would have changed their opinion and favored treatment had we asked about patients with even more severe disease, such as those on mechanical ventilation. And when we think about it, this makes sense. Whenever you decide to treat a patient, you must consider the risk-benefit ratio. If you have a patient who has only a cough and fever and appears likely to have a self-limited illness, the treatment benefit would be very small, as you might only shorten their disease course by a day or so. However, if you have a patient who is on the verge of death and they survive the illness after treatment, then the benefit is huge compared with the risk. That’s what the task force was inherently applying here — the changing risk-benefit ratio as the severity of illness increased.


How does this suggestion differ from those from other organizations or associations?

Wilson: Several organizations, such as the Surviving Sepsis Campaign, have opted to say that they cannot make a suggestion or recommendation for or against the use of hydroxychloroquine in critically ill patients with COVID-19 due to insufficient evidence. In contrast, we suggest that clinicians consider giving hydroxychloroquine if the following conditions are met: shared decision-making in which the patient is informed about the possible benefits and side effects and then agrees to try the medication; data can be collected for interim comparisons of patients who received hydroxychloroquine vs. those who did not; the disease is severe enough to warrant investigational therapy; and the drug is not in short supply. I believe we are the first to suggest using it on a case-by-case basis in this way.

It appears we’re being more aggressive and taking a novel stance. What’s being missed here, though, is the fact that hydroxychloroquine is already being widely used in the United States. Moreover, some countries are using it routinely and have built it into their COVID-19 guidelines. What this means is that recommendations, such as those from the Surviving Sepsis Campaign, that make no suggestion for or against hydroxychloroquine are basically saying to maintain the status quo. This means that institutions will continue to use hydroxychloroquine as they are now and its use will remain highly variable. Our suggestion acknowledges that it’s already being used but suggests that it be used with specific conditions so that we can standardize treatment, collect data and hopefully eliminate outliers who are using the drug in a dangerous way. I want to draw the distinction that the difference between no suggestion and a suggestion in this case is really just an attempt to standardize how hydroxychloroquine is being used since it’s already in use and to emphasize the possibility of doing real-world research to evaluate effectiveness.


In terms of other treatments, the task force also makes no suggestions for or against the use of remdesivir (Gilead Science), lopinavir/ritonavir, tocilizumab (Actemra, Genentech) or systemic corticosteroids in patients with COVID-19 pneumonia. Why is that? And would any of these treatments theoretically be appropriate for outpatients or patients without evidence of pneumonia?

Wilson: I can’t comment on whether these therapies would be appropriate for outpatients or those without pneumonia because we didn’t ask it of the task force. For these specific treatments, we stuck to patients with COVID-19 pneumonia. I will tell you, however, that trials are being done for hydroxychloroquine in different contexts, such as treatment of outpatients, post-exposure prophylaxis or general prophylaxis, so I assume the same is probably true for these other agents as well.

I will point out, though, that even though there wasn’t enough agreement to warrant a consensus suggestion for these medications, the task force considered each one a little bit differently. For example, participants’ thought process for remdesivir was very similar to that for hydroxychloroquine in that we had two groups — one that wanted to wait for clinical trials and one that was willing to try it. We only fell a couple of votes short of making a suggestion for remdesivir, and I think that was mainly because hydroxychloroquine is a very familiar drug that’s been around for more than 60 years and people are comfortable with it, which is not the case with remdesivir.

The discussion around lopinavir/ritonavir was different in that a randomized trial, which has been published in The New England Journal of Medicine, did not conclusively show any benefit. Therefore, the group was not willing to suggest using it. However, they also weren’t willing to suggest against using it because the trial was uncertain enough and they thought additional investigation was warranted before rolling it out.

For steroids, the task force nearly made a suggestion against using them because there is some evidence that using steroids allows virus replication to increase and viral shedding to be prolonged, which we don’t want. For tocilizumab, there just wasn’t enough data.

The guidance suggests the use of prone ventilation, followed by extracorporeal membrane oxygenation if prone ventilation fails, in hospitalized patients with COVID-19 pneumonia. What were these suggestions based on?

Wilson: The suggestion for use of prone ventilation was largely based on the fact that there was an assumption that acute respiratory distress syndrome due to COVID-19 pneumonia is pretty similar to ARDS due to other causes, such as sepsis, and there is an ATS guideline recommending prone ventilation for ARDS. There is also a fair amount of experience within the task force showing that prone ventilation was helpful in patients with COVID-19. What’s interesting about this, though, is that since we finished the survey, just in the last couple of days, research letters have been published suggesting that ARDS due to COVID-19 pneumonia is a bit different from traditional ARDS in that the lungs are less likely to get stiff. That said, even those who are reporting that ARDS is different in patients with COVID-19 are still saying that they have seen some benefit, albeit smaller, with prone ventilation. Prone ventilation also is not harmful or expensive, so it is basically worth a try.


How often will the guidance be updated?

Wilson: The guidance will be updated on an as-needed basis. The task force is very enthusiastic, so our plan is to move on and address additional questions while also keeping an eye on the literature and address any new research that we think is significant.

The beauty of the CORE process is that it’s very easy. With a clinical practice guideline, you have to go back and redo the systematic review when new evidence emerges. With the CORE process, it’s very quick because it’s consensus-based. We can inform the task force about additional evidence and ask the question again.

Is there anything in particular that you’re looking to address in the future?

Wilson: I’m not sure yet. I’m actually going to poll the task force soon and ask them to suggest questions. Then, we’ll get our steering committee together to look at those questions and choose which ones to address.

I will say that I’ve received a ton of emails with interesting questions that I never even thought of. For example, do patients on high-flow oxygen aerosolize, meaning do they need an isolation room? Other questions include whether or not vitamin D may prevent COVID-19, whether or not there’s a role for anticoagulation or thrombolysis in patients with COVID-19 or whether or not statins may help prevent or treat COVID-19. I’m not saying these are questions that will be addressed anytime soon, but we certainly have a lot to consider.

Do you have a message for clinicians at this time?

Wilson: I would say that this is a rapidly evolving situation. Optimal management is rapidly evolving, so we need to continue to reassess what we’re doing and modify as new evidence becomes available.


ATS. Guidance document on COVID-19. Available at: Accessed April 6, 2020.

For more information:

Kevin C. Wilson, MD, professor of medicine in the division of pulmonary, allergy, sleep and critical care medicine at Boston University School of Medicine and chief of guidelines and documents at the American Thoracic Society, can be reached at or

Disclosure: Wilson reports no relevant financial disclosures.