Why Are COVID Patients Returning to the ER After Discharge?

Robert D. Glatter, MD; Austin S. Kilaru, MD, MSHP; Ali S. Raja, MD, MBA, MPH


October 22, 2020

Find the latest COVID-19 news and guidance in Medscape's Coronavirus Resource Center.

This transcript has been edited for clarity.

Robert D. Glatter, MD: Hi. I'm Dr Robert Glatter, medical advisor for Medscape Emergency Medicine.

Today we'll be discussing an interesting study that was published by Dr Austin Kilaru, an adjunct assistant professor at Penn Medicine, and his colleagues. They looked at return visits and readmissions for ED patients diagnosed with COVID-19 within 1 week of their initial diagnosis and visit. The study found, interestingly, that slightly under 5% of the patients returned to the hospital and were admitted within 3 days of their initial ED visit, and just under 4% were hospitalized within a week. Overall, that's about 9% of patients being readmitted to the hospital after the first ED visit due to COVID-19.

Dr Ali Raja, associate professor of emergency medicine and executive vice chair of emergency medicine at Massachusetts General Hospital, will also be joining us to discuss this important research.

Welcome, gentlemen.

Austin S. Kilaru, MD, MSHP: It's good to be with you.

Ali S. Raja, MD, MBA, MPH: Thank you for having us, Rob.

Glatter: We really appreciate it. Dr Kilaru, I'd like to begin with you. You can maybe review your study design, your objectives, your implications for the study, and your motivation to do the study.

Kilaru: Absolutely. The motivation for this study came out of a conversation I had in March with one of my colleagues who was working next to me in the emergency department. We had just started to see more and more patients with COVID-19 in Philadelphia. We both had patients that we were managing and trying to decide what would be the best plan for those patients.

My colleague had a patient who looked very well, was older, had findings of pneumonia on a chest x-ray, and asked me, "What do I do with this patient? Do I send them home?" At that point, our hospital was not having any issues with capacity, so we would have been able to find a bed for that patient.

With that in mind, as we continued to progress with the pandemic through March, April, May, and June, we noticed that some patients were coming back to the hospital after being discharged from the emergency department.

That, in combination with the fact that this particular virus seemed to be acting differently than other viral illnesses that we see, in that patients have somewhat delayed worsening in symptoms — the traditional message that we tell patients is that they'll be better in 3 to 5 days — that didn't necessarily seem to apply to all patients with COVID-19.

Glatter: What were the main reasons for return visits when you looked across the board? What were the main criteria? Could you talk about the age distribution and what you were finding in that context?

Kilaru: This specific paper doesn't mention the exact reasons for the return admission. That being said, data we did not report in the paper indicate that most of the patients who returned had respiratory issues, worsening shortness of breath, and in some cases, respiratory failure. We also had some patients who returned reporting worsening fatigue, dehydration, and other sequelae of prolonged viral illness.

As you noted, we found that age was a significant risk factor for returning to the hospital after initially appearing well and going home. When we adjusted for other characteristics of patients, we found that patients over the age of 60 had almost three times the rate of returning to the hospital. Those issues around shortness of breath, fatigue, and a decreasing ability to remain at home and to be supported at home were the reasons that older patients were coming back to the hospital.

Glatter: When we talk about return visits, we are always thinking, Did we miss something? when a patient bounces back. With COVID-19, though, certainly there's a spectrum or evolution of the illness itself that we're learning. Certainly, when patients want to go home and they're stable, they can go home.

In your summation, do you feel, in looking at the patient population, that anything was being missed? Or was this just the spectrum of the illness in and of itself that we are learning about?

Kilaru: That's a great question. Just like you mentioned, we're trapped on our bounce-backs. Bounce-backs are never desirable. Again, we didn't specifically look at whether some of these return visits could be called a mistake. Particularly for the patients we were looking at in March, April, and May, we were all learning about this illness and how to manage it.

I will say that, for the most part, the patients who were sent home were stable and wanted to go home, so the physician and the patient probably together decided that was the appropriate plan of care. I don't see these return admissions being bounce-backs or failures. I think it is mostly a sign of the illness that continues to progress and may worsen for some patients somewhat unpredictably, but also with some predictable risk factors that you can expect.

One initial evaluation in an emergency department or by a clinician is just a single snapshot in time. As things evolve, patients may very well need an additional evaluation or a higher level of care.

Glatter: Ali, I want to bring you into the discussion. In your experience at Massachusetts General Hospital, what type of situation are you seeing in terms of bounce-backs, revisits, and readmissions? Can you talk to us about your experience?

Raja: Thanks, Rob. I think that we are seeing very similar findings to those noted by Dr Kilaru in this study. I have to say that this is a great study and I love it for a number of different reasons. The first is that it began when you were on shift, talking to another physician, and she or he brought this question up. That's how the best studies are begun. It's from a real clinical need, a real clinical question.

The fact that [1 in 10] patients are bouncing back within a week of discharge is not surprising to me. Remember that when all of this began a few months ago, like you talked about when the study was first thought of, we had plenty of room in the hospital. Unfortunately, as the pandemic spread and we found our hospitals getting more and more full — in our hospital in the months of April and May, we often had 140 patients who were in the ICU; typically, we have between 70 and 90 ICU beds — this was a situation where we had to limit resources, so we were discharging patients we might otherwise have kept in the hospital.

I think this study is also great because of the fact that your main findings — age, fever, oxygen saturation — those are things that patients can measure at home, especially the fever and the oxygen saturation. Like you said, Dr Kilaru, we have a snapshot of what they look like, but when they go home, your study has actually given us the fact that if their oxygen saturation was initially bad when they first came in or if they had a fever, they were more likely to bounce back.

This supports the fact that, like many of us do, we tell patients, "Keep checking your temperature and your oxygen saturation. If it starts dropping, call your doctor and consider coming back."

Glatter: Austin, if you could expand a little bit about your monitoring system at home. You mentioned in the article that it's a text messaging–based system. Is there any video capability?

Kilaru: That's a great question. We actually have a few different monitoring systems that we've launched at the University of Pennsylvania. Some have video capability, some don't.

The COVID Watch program — and I'll refer the audience to an article published in NEJM Catalyst on this program — is an automated, twice-daily text message–based system where patients who are enrolled in this program, who can be from the emergency department, an outpatient setting, or even patients discharged from the hospital, receive an automated text asking about their symptoms.

If their symptoms are worsening or there are other problems that these automatic tests are detecting, then the case is escalated to a pool of clinicians who can further triage or answer questions for the patient.

We're currently evaluating the effectiveness of that program in terms of being able to keep healthy patients at home while bringing the patients who need to come back, back to the hospital. In general, this program is loved by both patients and clinicians because it really reduces some of the burden on outpatient offices, for instance, that might otherwise be inundated with calls about symptoms.

We've also partnered with our home health colleagues in our Penn Medicine Home Health agency to enroll particularly older and sicker patients who want to recover at home and are then provided with home health services, including visiting nurses. There is some telemedicine built into the monitoring that the home health providers are offering.

Glatter: Ali, would you echo that experience in terms of home monitoring? Do you have any kind of system set up with Mass General to monitor your discharged patients for decompensation?

Raja: It's a great question. We don't have as fancy a system as the Penn Medicine folks do. COVID Watch sounds fantastic. I just started looking it up in NEJM Catalyst and it seems like there have been a few thousand patients who have already gone through that.

In general, what we had was a tiered response. For most patients, we would actually just tell them to stop by the drugstore and pick up a $20 or $30 oxygen saturation monitor and a thermometer, and to check their oxygen saturation, especially with ambulation — not just while sitting or lying around.

We also set up a system where local paramedics partnered with us and did home health visits. We also have a home hospital system, much like the one at UPenn, that involves either nurse practitioners or physicians visiting patients in their houses.

I think this was a great opportunity for all of us to expand upon our somewhat nascent telemedicine systems. This allowed us to set up systems where patients could telephone in or video call in with their primary care physicians so that if they started feeling bad, they didn't have to try to come right back to the emergency department. They could get their doctor on the phone and she could guide them based on what symptoms they were feeling and what kind of temperatures and oxygen saturation they were experiencing.

Glatter: Moving on to a different topic, let's discuss racial disparities and how that may have contributed to the readmission rate in patients who were discharged. Austin, do you have any data on that from this study to really link that to the different demographics of your patients — the rates of infection and so forth?

Kilaru: Thank you for asking that. I think the relationship between race and COVID-19 is so complicated. In our study, we did include race as a potential risk factor. In the patient population that we studied, over 50% of the patients were African American.

We did not find that race was a statistically significant predictor in whether patients needed to return to the hospital for admission. That being said, these are patients who were already seen and evaluated in the emergency department.

We know that in our community in West Philadelphia, where my hospital is located, Black patients have a much higher rate of contracting the virus in the first place. The mortality rates are higher for our population. Although we didn't find that specific finding in our study to be a risk factor, we're still very concerned about the differential outcomes for patients with COVID-19 who are Black.

Glatter: Ali, if you could comment on that as well — your experience at your institution.

Raja: Dr Kilaru hit the nail on the head. Although that may have not borne out in this particular study with the subset of patients who were discharged and bounced back, the fact is that the hardest-hit parts of Boston have been the places that have the patients with the lowest socioeconomic status. Unfortunately, these also happen to be places where there are predominantly people of color and non-English-speaking patients.

We found that not only were they unfortunately hit harder than White English-speaking patients in the city, but on top of that, even when they came in and seemed to be equally ill, care was just a lot more difficult.

For example, having end-of-life or goals-of-care discussions with patients while using a phone interpreter, when everybody's wearing a mask and there's a fan going off in a negative-pressure room, is so much harder than being able to just have a face-to-face conversation with an English-speaking patient in the same scenario.

We found that we had to set up systems of care to help give those patients care. We, for example, had a Spanish-language care group where we had everybody from general surgeons to psychiatrists who spoke Spanish who specifically came in to the emergency department to have these critical conversations with our critically ill patients so that we didn't have to go through an interpreter or try to communicate in ways that weren't optimal.

We all had to set up systems in place — and I'm sure the same is true in New York and Philadelphia — to make sure that we gave patients equitable care despite the language barriers and the fact that they may have come from communities of color where primary care might not have been as available as it was in other settings.

Glatter: The takeaway I'm getting is that people older than 60 and people who are hypoxic in presentation are the people we need to watch and maybe reconsider about sending home. That said, if there is appropriate monitoring in place, it's something to consider upon discharge.

Kilaru: I'll just add that our goal for the study is not necessarily to direct emergency clinicians as to say whether you should be sending home these patients or admitting these patients, because ultimately that decision is part of our practice. We're trained to make those decisions.

I'm hoping that the impact of this study could be that we are able to have more informed conversations with patients and to really take the time, which can honestly be challenging in the emergency department.

We want to limit our contact in some cases with patients who are suspected to have COVID-19, but we do really want to take the time to have these conversations with patients, to advise them that their symptoms could potentially worsen, and to make sure they have a plan in place for, if they are getting sicker, knowing what to do, whether it's through the programs that we've talked about or just through calling their doctor.

I think that is the goal. We were finding that there are predictable risk factors for reasons that patients come back to the hospital. With any patient that you're sending home with COVID-19 that you're concerned about, we want to provide them with the most up-to-date information on what they can expect during the course of their illness.

Glatter: Dr Raja, do you want to add anything to those comments?

Raja: This is such a scary diagnosis, so being able to provide patients with any statistics that allow them to be somewhat reassured about being discharged home helps.

A study like this is great because we know that the right thing for these patients is to be treated at home, from a hospital capacity standpoint, but also so that they don't get other hospital-acquired infections.

Sometimes patients come in and think, Oh, I've tested positive for COVID-19. I should automatically stay in the hospital. Being able to use data like these to say, "No, you can actually be safely cared for at home" allow us to fall back on this kind of great study. Thank you, Dr Kilaru, for your team's work.

Glatter: I want to thank you both. This has really been an informative discussion. It gave us some guidelines to look at, but certainly, shared decision-making and the patient's comfort are the most important things. Thank you again.

Robert D. Glatter, MD, is an attending physician at Lenox Hill Hospital in New York City and assistant professor of emergency medicine at Zucker School of Medicine at Hofstra/Northwell in Hempstead, New York. He is an editorial advisor and hosts the Hot Topics in EM series on Medscape. He is also a medical contributor for Forbes.

Austin S. Kilaru, MD, MSHP, is an adjunct professor at the Perelman School of Medicine at the University of Pennsylvania and an attending physician at Pen Presbyterian Medical Center in Philadelphia, Pennsylvania. His research has been featured in The New England Journal of Medicine, JAMA Network Open, Health Affairs, and Annals of Emergency Medicine.

Ali S. Raja, MD, MBA, MPH , is associate professor of emergency medicine and executive vice chair at Massachusetts General Hospital in Boston, Massachusetts. A practicing emergency physician and author of over 200 publications, his federally funded research focuses on improving the appropriateness of resource utilization in emergency medicine.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.