Avoiding Disinformation Traps in COVID-19

Article in Review

Richardson S, Hirsch JS, Narasimhan M, et al. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA. 2020;323:2052-2059.

What Question Did This Investigation Aim to Answer?

What are the characteristics, clinical presentation, and outcomes of patients hospitalized with coronavirus disease 2019 (COVID-19) in the United States?

What Study Design Did the Authors Choose?

Design: Observational case series.

Setting: Twelve hospitals in New York City, Long Island, and Westchester County, NY.

Population: A total of 5,700 patients with confirmed severe acute respiratory syndrome coronavirus 2 by positive result on polymerase chain reaction testing of a nasopharyngeal sample.

Outcome Measures: Clinical outcomes, including invasive mechanical ventilation, kidney replacement therapy, and death.

Sponsors: National Institutes of Health and National Heart, Lung, and Blood Institute.

How Did the Authors Interpret the Results?

Of the 5,700 patients included in their analysis, 3,066 remained hospitalized without a final disposition recorded. Common presenting findings at initial hospitalization included fever (30.7%), an oxygen saturation less than 90% (20.4%), lymphocyte counts less than 1,000×10⁹/L (60%), and elevated liver function test results. At data collection, 2,081 patients had been discharged and 553 had died. Of the 373 patients admitted to the ICU, 291 (78.0%) had died, including 88.1% of those requiring mechanical ventilation. Mortality rates for patients with hypertension who were not receiving an angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker, were receiving an angiotensin-converting enzyme inhibitor, and were receiving an angiotensin II receptor blocker were 26.7%, 32.7%, and 30.6%, respectively. These data provide characteristics and early outcomes of hospitalized patients with COVID-19.

How Might This Study Affect Your Clinical Practice in the Emergency Department?

These data provide brief insight into the general characteristics and outcomes of patients hospitalized for COVID-19. Although there is little in this report to provide clinical guidance for assessing the severity of COVID-19 or informing decisions regarding who requires hospitalization, these data suggest that in its severe form, COVID-19 is a prolonged and resource-intense disease.

Discussion Points

• 1. The mortality rate in patients requiring mechanical ventilation was reported as 88.1%, higher than that of previous reports. What is censoring, and how might it affect the results of an analysis such as this?

Censoring occurs when data regarding an event of interest are incomplete. In this case, what has occurred is known as “right censoring,” or when the study period ends without complete outcome data on all participants. This typically occurs in 1 of 2 ways: either the event has not yet occurred for all study participants by the end of the study period or a number of patients are lost to follow-up and outcome data are unavailable.¹ In the case of Richardson et al, at data analysis, only 314 of the 1,152 patients requiring mechanical ventilation had either died or been discharged from the hospital. The remaining 838 patients (72.7%) continued to be intubated and mechanically ventilated. These patients could ultimately contribute to either the numerator or the denominator for true mortality reporting. Furthermore, the fate of the remaining 1,914 patients who remained hospitalized remains unclear. The potential exists that some of these patients not yet requiring mechanical ventilation will clinically decline and require mechanical ventilatory support. These patients would also contribute their effects to the final mortality and recovery statistics.

The net effect of this censoring leads to the appearance that mortality in patients receiving mechanical ventilation is higher than traditionally expected in patients with acute respiratory distress syndrome. Critically ill patients with COVID-19 who require mechanical ventilation may have a higher-than-expected mortality owing to disease-specific factors. However, because the majority of patients remain intubated and have yet to experience the outcome of interest (ie, death or hospital discharge), it is far too early to draw conclusions regarding overall mortality of the cohort.

• 2.How might these interim observations ideally be presented?

An alternative presentation of these data would be explicitly inclusive of the status of all patients receiving mechanical ventilation at data collection. Along these lines, on April 22, a clarification was published by the Journal of the American Medical Association stating, “As of April 4, 2020, for patients requiring mechanical ventilation (n = 1151, 20.2%), 38 (3.3%) were discharged alive, 282 (24.5%) died, and 831 (72.2%) remained in hospital.” The authors could have also further illustrated the effects of their necessary censoring by highlighting the resulting uncertainty. As is evident from the raw data, the final full cohort mortality could theoretically range from 24.5% to 96.7%, depending on the outcomes of the remaining patients still admitted to the hospital and requiring care.

Section editors: Tyler W. Barrett, MD, MSCI; Ryan P. Radecki, MD, MS; Rory J. Spiegel, MD