Researchers detected an unexpected 50% increase in patients presenting with respiratory illnesses at UCLA Health facilities in the months before the pandemic.
UCLA researchers and colleagues who analyzed electronic health records found that there was a significant increase in patients with coughs and acute respiratory failure at UCLA Health hospitals and clinics beginning in late December 2019, suggesting that COVID-19 may have been circulating in the area months before the first definitive cases in the U.S. were identified.
This sudden spike in patients with these symptoms, which continued through February 2020, represents an unexpected 50% increase in such cases when compared with the same time period in each of the previous five years.
The findings, the study authors say, demonstrate the importance of analyzing electronic health records to monitor and quickly identify irregular changes in patient populations. The researchers’ novel approach, in which they focused not only on hospitalization data but also on data from outpatient settings, may help epidemiologists and health systems detect future epidemics sooner.
The study appears in the peer-reviewed Journal of Medical Internet Research.
“For many diseases, data from the outpatient setting can provide an early warning to emergency departments and hospital intensive care units of what is to come,” said Dr. Joann Elmore, the study’s lead author and a professor of medicine in the division of general internal medicine and health services research at the David Geffen School of Medicine at UCLA. “The majority of COVID-19 studies evaluate hospitalization data, but we also looked at the larger outpatient clinic setting, where most patients turn first for medical care when illness and symptoms arise.”
As scientists and doctors continue to learn more about SARS-CoV-2, the virus that causes COVID-19, health systems and public health agencies are also attempting to predict and monitor cases. Analyzing electronic patient records, the researchers say, could help health authorities more effectively identify and control outbreaks like the current pandemic, which has killed hundreds of thousands worldwide and disrupted billions of lives.
“The pandemic has really highlighted our need for agile health care analytics that enable real-time symptom and disease surveillance using electronic health records data,” said Dr. Michael Pfeffer, a study co-author and chief information officer for UCLA Health. “Technology, including artificial intelligence powered by machine learning, has further potential to identify and track irregular changes in health data, including significant excesses of patients with specific disease-type presentations in the weeks or months prior to an outbreak.”
The researchers evaluated more than 10 million health system and patient visit records for UCLA Health outpatient, emergency department, and hospital facilities, comparing data from the period between December 1, 2019, and February 29, 2020 — the months prior to increased public awareness of COVID-19 in the U.S. — with data from the same period over the previous five years.
They found that outpatient clinic visits by UCLA patients seeking care for coughs increased by over 50% and exceeded the average number of visits for the same complaint over the prior five years by more than 1,000. Similarly, they discovered a significant excess in the number of patients seen in emergency departments for reports of coughs and of patients hospitalized with acute respiratory failure during this time period. These excesses remained even after accounting for changes in patient populations and seasonal variation.
The researchers noted that other factors could be responsible for some of this unexpected increase. For instance, their search of outpatient visit records included only the word “cough” as the reason for clinic visits, which may not have been sufficiently specific, and respiratory illnesses could have been due to vaping, though the use of e-cigarettes had been declining since September 2019. In addition, they could not rule out that the excess cases were due to flu.
“We may never truly know if these excess patients represented early and undetected COVID-19 cases in our area,” Elmore said. “But the lessons learned from this pandemic, paired with health care analytics that enable real-time surveillance of disease and symptoms, can potentially help us identify and track emerging outbreaks and future epidemics.”
Reference: “Excess Patient Visits for Cough and Pulmonary Disease at a Large US Health System in the Months Prior to the COVID-19 Pandemic: Time-Series Analysis” by Joann G Elmore, MD, MPH; Pin-Chieh Wang, PhD; Kathleen F Kerr, PhD; David L Schriger, MD, MPH; Douglas E Morrison, MS; Ron Brookmeyer, PhD; Michael A Pfeffer, MD; Thomas H Payne, MD and Judith S Currier, MD, 10 September 2020, Journal of Medical Internet Research.
Additional study authors are Dr. Judith Currier, Dr. David Schriger, Pin-Chieh Wang, Douglas Morrison and Ron Brookmeyer, all of UCLA, and Kathleen Kerr and Dr. Thomas Payne of the University of Washington.
This study was supported by the UCLA Department of Medicine.
It got an early start in somewhat remote, New Orleans too. Mardi Gras celebrations went ahead, starting in mid February, since nobody warned the City officials that it could be spread through the air, even thought inside the White House, Trump knew it could be, by late January, if not sooner. We know that now, since they held an emergency COVID-19 threat meeting inside the White House on January 28. Peter Navarro had, even earlier, written a memo warning that between a half million, and two million Americans could die from COVID-19. Yet nobody warned local officials about what was might be coming.
With no warning of airborne spread potential, Mardi Gras celebrations went ahead during the second half of February. Days after that celebration ended on February 25, little New Orleans suddenly became a global outbreak hot spot, probably due to visitors from Europe, and the New York area, bringing the virus to New Orleans during Mardi Gras. After cases started appearing, Louisiana became the first state in the South to lock down. But by then it was too late. COVID-19 had spread everywhere. Thousands of people who weren’t sick were unknowingly spreading the virus, after attending the Mardi Gras parades and balls. They had no idea that they might be inhaling a deadly virus in a screaming parade crowd. Everyone thought that the Mardi Gras celebration would be over before COVID-19 could get to New Orleans from a few cases in a distant Washington State nursing home. Xi Jinping never told them it was airborne, like he had told Trump in January.
A January speech by Trump on national prime time TV, warning what Xi Jinping had told him in January, about it spreading through the air, could have saved tens of thousands of lives. “Panicing” people as soon as possible into wearing masks, staying away from crowds, and out of bars, would have been a great idea. Instead, we were told for MONTHS that COVID-19 was a “hoax” and would, “soon just disappear”. Trump LIED, and thousands more Americans DIED unnecessarily as a result. That is a FACT, as admitted by Donald Trump himself, on tape, to Bob Woodward. Trump put his reelection, and keeping the Dow Jones Industrial average up, ahead of thousands of American lives, which could have been saved with an early national lock down, like has been largely successful in other countries. The only “panic” was by Trump, concerning his reelection chances. He did virtually nothing, hoping that not enough people would die to make a difference at the polls in November. As the body count began to mount, he tried blaming everything on the state governors. Like they had the resources to fight a new deadly disease, never seen on Earth before 2019. How many microbiology research labs does Mississippi have?
Now, we will be lucky to avoid another significant spike in COVID-19 cases, starting around Thanksgiving. With so many cases now spread throughout the entire country, even with a vaccine, it will be a significant health threat for at least a couple more years. With early national action, that risk could have been significantly reduced.
A certain death rate increase would be predicted from a Covid-19 outbreak beginning in December, especially in the hospitalized group of “Excess” cases related to “cough”. The same medical records surveyed in this paper should also contain patient case fatality rates to be estimated for this target group in December through February.
The current stats on Covid-19 show a very rapid rise in cases and fatalities from late February onward…it seems intuitive that if the outbreak had started in December rather than February, there should be a similar shaped curve of cases and deaths within the studied group of cases in this paper. The synopsis of this paper doesn’t address whether the paper considered these points.
LAX is a major international airport. If they didn’t have Covid-19 in December, then that’d be news. But we are in regime where we don’t want to test lest we find Covid-19. And even deaths are under reported.
I saw in some study that deaths are under reported – marked to other causes even when covid precipitated the cause.
All of this is pure speculation. Why don’t the researchers use the same patient information and do antibody follow up. For those with antibodies, couldn’t they see whether the genomic info of those with antibodies correlates with early versions of Covid 19 or later versions?