Key PointsQuestion
What trends for in-hospital mortality by age among hospitalized patients with SARS-CoV-2–positive tests are evident between March 1 and November 21, 2020?
Findings
This cohort study of 503 409 patients from 209 US acute care hospitals found a decrease in in-hospital mortality among patients with SARS-CoV-2–positive tests. Mortality rates were 10.6% in March, increased to 19.7% in April, and decreased to 9.3% in November.
Meaning
This large study supported findings of smaller, regional studies that reported in-hospital mortality decreased for all age groups during the pandemic period and suggests that this decrease cannot be solely attributed to hospital admission of increased proportions of younger patients.
Importance
Mortality is an important measure of the severity of a pandemic. This study aimed to understand how mortality by age of hospitalized patients who were tested for SARS-CoV-2 has changed over time.
Objective
To evaluate trends in in-hospital mortality among patients who tested positive for SARS-CoV-2.
Design, Setting, and Participants
This retrospective cohort study included patients who were hospitalized for at least 1 day at 1 of 209 US acute care hospitals of variable size, in urban and rural areas, between March 1 and November 21, 2020. Eligible patients had a SARS-CoV-2 polymerase chain reaction (PCR) or antigen test within 7 days of admission or during hospitalization, and a record of discharge or in-hospital death.
Exposure
SARS-CoV-2 positivity.
Main Outcomes and Measures
SARS-CoV-2 infection was defined as a positive SARS-CoV-2 PCR or antigen test within 7 days before admission or during hospitalization. Mortality was extracted from electronically available data.
Results
Among 503 409 admitted patients, 42 604 (8.5%) had SARS-CoV-2–positive tests. Of those with SARS-CoV-2–positive tests, 21 592 (50.7%) were male patients. Hospital admissions among patients with SARS-CoV-2–positive tests were highest in the group aged 65 years or older (19 929 [46.8%]), followed by those aged 50 to 64 years (11 602 [27.2%]) and 18 to 49 years (10 619 [24.9%]). Hospital admissions among patients 18 to 49 years of age increased from 1099 of 5319 (20.7%) in April to 1266 of 4184 (30.3%) in June and 2156 of 7280 (29.6%) in July, briefly exceeding those in the group 50 to 64 years of age (June: 1194 of 4184 [28.5%]; 2039 of 7280 [28.0%]). Patients with SARS-CoV-2–positive tests had higher in-hospital mortality than patients with SARS-CoV-2–negative tests (4705 [11.0%] vs 11 707 of 460 805 [2.5%]; P < .001). In-hospital mortality rates increased with increasing age for both patients with SARS-CoV-2–negative tests and SARS-CoV-2–positive tests. In patients with SARS-CoV-2–negative tests, mortality increased from 45 of 11 255 (0.4%) in those younger than 18 years to 4812 of 107 394 (4.5%) in those older than 75 years. In patients with SARS-CoV-2–positive tests, mortality increased from 1 of 454 (0.2%) of those younger than 18 years to 2149 of 10 287 (20.9%) in those older than 75 years. In-hospital mortality rates among patients with SARS-CoV-2–negative tests were similar for male and female patients (6273 of 209 086 [3.0%] vs 5538 of 251 719 [2.2%]) but higher mortality was observed among male patients with SARS-CoV-2–positive tests (2700 of 21 592 [12.5%]) compared with female patients with SARS-CoV-2–positive tests (2016 of 21 012 [9.60%]). Overall, in-hospital mortality increased from March to April (63 of 597 [10.6%] to 1047 of 5319 [19.7%]), then decreased significantly to November (499 of 5350 [9.3%]; P = .04), with significant decreases in the oldest age groups (50-64 years: 197 of 1542 [12.8%] to 73 of 1341 [5.4%]; P = .02; 65-75 years: 269 of 1182 [22.8%] to 137 of 1332 [10.3%]; P = .006; >75 years: 535 of 1479 [36.2%] to 262 of 1505 [17.4%]; P = .03).
Conclusions and Relevance
This nationally representative study supported the findings of smaller, regional studies and found that in-hospital mortality declined across all age groups during the period evaluated. Reductions were unlikely because of a higher proportion of younger patients with lower in-hospital mortality in the later period.
In-hospital mortality among patients with COVID-19 was high early in the pandemic and was reported to range from 12% to 28% in early case series.1-3 These studies also described a linear association between age and mortality.1,3,4 Since then, in-hospital mortality rates have decreased, likely reflecting improvements in survival.2,5,6 However, it has been proposed that at least some of the reduction in mortality rates in recent months can be attributed to the higher proportion of younger patients, who have lower mortality rates, among those hospitalized with COVID-19.7 In this analysis, in-hospital mortality rates for patients with SARS-CoV-2 infection were evaluated over time, and factors associated with observed trends were examined.
We conducted a multicenter, retrospective cohort analysis of data from 209 acute care facilities in the BD Insights Research Database (Becton, Dickinson and Company). The study data set was approved as a limited, deidentified data set for retrospective analysis and was exempted from patient consent by the New England institutional review board because it is a retrospective analysis of deidentified data. This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.8
The hospitals included in the analysis have a range of bed sizes and are located in urban and rural areas throughout the United States. The database includes electronically captured laboratory results, US Census data, pharmacy orders, and admission, discharge, and transfer (ADT) data feeds.9 Hospitals in this analysis are a subset of facilities that are similar in bed size, census division, urban/rural designation, and teaching status to US hospitals included in the American Hospital Association survey.10
Eligible patients were hospitalized for at least 1 day, had results reported from a SARS-CoV-2 polymerase chain reaction (PCR) or antigen test 7 days prior to admission or during hospitalization, and a record of discharge or death in the hospital between March 1 and November 21, 2020. In this database, in-hospital mortality was defined as a designation of death, mortality, or presence in morgue in the ADT data feeds. We performed a validity check by randomly selecting 50 mortality cases and evaluating them for encounter-level clinical signs of mortality (eg, uncorrected severe metabolic acidosis determined by pH from arterial blood gases and uncorrected electrolyte changes incompatible with life). SARS-CoV-2 infection was defined as a positive SARS-CoV-2 PCR or antigen test 7 days prior to admission or during hospitalization.
Frequencies were calculated for categorical variables. Means with SDs and medians with ranges were calculated for continuous variables. We used the χ2 and 2-sample t test to compare categorical and continuous variables, respectively. Linear regression was used to evaluate trends for mortality over time. All statistical analyses were performed using SAS software version 9.4 (SAS Institute). Statistical significance was set at P < .05, and tests were 2-tailed.
We identified 503 409 unique patients hospitalized between March 1 and November 21, 2020, comprising 42 604 patients (8.5%) with laboratory-confirmed SARS-CoV-2–positive tests and 460 805 patients (91.5%) with SARS-CoV-2–negative tests. A minority of hospitalized patients were younger than 18 years (11 709 [2.3%]) (Table).
Overall, the youngest age group (ie, <18 years) had the lowest proportion of patients with SARS-CoV-2–positive tests (454 of 11 709 [3.9%]), and patients 50 to 64 years of age had the highest proportion (11 602 of 119 673 [9.7%]) of patients with SARS-CoV-2–positive tests. Patients with SARS-CoV-2–positive tests, compared with those with SARS-CoV-2–negative tests, were more likely to be male (21 592 [50.7%] vs 209 086 [45.4%]; P < .001) and to have at least 1 underlying severe illness or condition (33 248 [78.0%] vs 277 396 [60.2%]; P < .001). Among patients with SARS-CoV-2–positive tests vs with SARS-CoV-2–negative tests, mean (SD) hospital length of stay was longer (8.5 [13.6] days vs 5.2 [8.0] days; P < .001), mean intensive care unit (ICU) length of stay was longer (8.3 [9.8] days vs 3.8 [5.5] days; P < .001), and ICU admission was more common (9586 [22.5%] vs 65 717 [14.3%]; P < .001) (Table).
Hospital Admission Trends by Age
Hospital admissions among patients with SARS-CoV-2–positive tests were highest in the group aged 65 years or older (19 929 [46.8%]), followed by those aged 50 to 64 years (11 602 [27.2%]) and 18 to 49 years (10 619 [24.9%]) (Figure 1). Hospital admissions in patients aged 18 to 49 years increased from 1099 of 5319 (20.7%) in April to 1266 of 4184 (30.3%) in June and 2156 of 7280 (29.6%) in July and exceeded rates in patients aged 50 to 64 years (June: 1194 of 4184 [28.5%]; 2039 of 7280 [28.0%]). The proportion of patients admitted with SARS-CoV-2–positive tests who were younger than 18 years, while extremely low overall, also increased in June (Figure 1). These data indicate that hospital admissions among younger persons increased briefly in June and July but declined again thereafter.
Patients with SARS-CoV-2–positive tests were significantly more likely to experience in-hospital mortality compared with those with SARS-CoV-2–negative tests (4705 [11.0%] vs 11 707 [2.5%]; P < .001) (Table; Figure 2). In-hospital mortality rates increased with increasing age in both patients with SARS-CoV-2–negative tests and those with SARS-CoV-2–positive tests. In patients with SARS-CoV-2–negative tests, mortality increased from 45 of 11 255 (0.4%) in those younger than 18 years to 4812 of 107 394 (4.5%) in those older than 75 years. In patients with SARS-CoV-2–positive tests, mortality increased from 1 of 454 (0.2%) of those younger than 18 years to 2149 of 10 287 (20.9%) in those older than 75 years (Table; Figure 2). While observed in-hospital mortality rates were similar in male and female patients with SARS-CoV-2–negative tests (6273 of 209 086 [3.0%] vs 5538 of 251 719 [2.2%]), higher in-hospital mortality was observed among male patients with SARS-CoV-2–positive tests compared with female patients with SARS-CoV-2–positive tests (2700 of 21 592 [12.5%] vs 2016 of 21 012 [9.6%]) (Figure 2). In the oldest age groups, in-hospital mortality rates were significantly higher among patients with SARS-CoV-2–positive tests compared with patients with SARS-CoV-2–negative tests (>75 years: 2149 of 10 287 [20.9%] vs 4812 of 107 394 [4.5%]; P < .001) (Figure 2).
Figure 3 shows in-hospital mortality by month of admission for patients with SARS-CoV-2–negative tests and in-hospital mortality by month of admission overall and by age group for patients with SARS-CoV-2–positive tests. In-hospital mortality among patients with SARS-CoV-2–negative tests decreased slightly and then remained stable throughout the study period. In-hospital mortality among patients with SARS-CoV-2–positive tests increased from March to April (63 of 597 [10.6%] to 1047 of 5319 [19.7%]) and then decreased significantly during the study period (1047 of 5319 [19.7%] to 499 of 5350 [9.3%]; P = .04). Significant decreases in in-hospital mortality were also noted in the oldest age groups, including patients aged 50 to 64 years (197 of 1542 [12.8%] to 73 of 1341 [5.4%]; P = .02), 65 to 75 years (269 of 1182 [22.8%] to 137 of 1332 [10.3%]; P = .006), and older than 75 years (535 of 1479 [36.2%] to 262 of 1505 [17.4%]; P = .03) (Figure 3).
This large multicenter, geographically representative, retrospective cohort study found a significant decrease in in-hospital mortality among patients with SARS-CoV-2–positive tests from March to November. Mortality among patients with SARS-CoV-2–negative tests remained stable throughout the study period after the initial peak that occurred from March to May. In-hospital mortality decreases were not associated with a shift toward a higher proportion of younger (ie, 18 to 49 years of age) hospitalized patients with SARS-CoV-2–positive tests, who tend to have lower in-hospital mortality. While there was a slight shift in the proportion of hospitalized patients with SARS-CoV-2–positive tests to younger age groups in June and July, the proportion in this age group returned to levels seen earlier in the pandemic shortly afterward. In-hospital mortality decreased in all age groups, including the oldest age groups, who accounted for the highest proportion of all hospitalized patients with SARS-CoV-2–positive tests and had the highest in-hospital mortality rates. In-hospital mortality decreased overall, from a monthly peak of 19.7% in April to 9.3% in November, with every age group experiencing a decrease of approximately 50%. This large, national study is consistent with recently published smaller studies demonstrating decreases in in-hospital mortality and COVID-19 risk-adjusted mortality rates.2,11 Reasons for decreases in mortality since the start of the pandemic may include increased clinical experience in caring for and ventilating patients and use of prone positioning, systemic corticosteroids, and remdesivir.12-14
Strengths of this study include the extraordinarily large sample size, which included a diverse population across multiple geographic regions, including a large number of pediatric patients. Previously published studies have been conducted in single health systems in specific geographic areas with fewer patients.6,15 Longitudinal data for pediatric patients in our study showed that this age group had an increased proportion of all admissions with SARS-CoV-2–positive tests over time compared with other age groups—a finding that should be confirmed in other studies. In-hospital mortality among pediatric patients with SARS-CoV-2–positive tests was extremely rare (1 of 454 [0.2%]). Reasons for lower mortality among pediatric patients are likely multifactorial, ranging from lack of high-risk comorbid chronic conditions to differences in T cell–mediated humoral immune response in pediatric vs older populations.16-18 Most patients in our cohort were adults, and mortality in patients with SARS-CoV-2–positive tests was associated with male sex and having at least 1 underlying medical condition, consistent with other studies.1,15
Limitations of this study include testing type and access. For the time interval assessed, the predominant diagnostic test for SARS-CoV-2 detection was PCR. Antigen testing was only readily available later in the pandemic and as a point-of-care test. These results may have been inconsistently captured in acute care emergency medical records. In addition, this study did not assess any discordance in test results among the 2 testing assays; for example, if an antigen test was positive and then a subsequent PCR test was negative for the same patient, the patient was considered as having a positive test. Furthermore, each hospital admission was counted as a single observation in the denominator. We did not link index hospital admissions and readmissions. We evaluated mortality trends by month and assumed that readmission rates were equal over time; therefore, the trends should not be affected. Additionally, the availability of diagnostic tests and public policy lockdowns increased during the pandemic and may have affected earlier mortality rates.19 We believe there was some ascertainment bias in the identification of mortality among patients with SARS-CoV-2–positive tests and those with SARS-CoV-2–negative tests in the month of March when the pandemic was beginning. Mortality among patients with SARS-CoV-2–negative tests declined from March to May and mortality among those with SARS-CoV-2–positive tests increased from March to April (Figure 3). It is likely that this was because of testing practices and a shortage of available SARS-CoV-2 diagnostics led to an initial inability to test all patients with suspected COVID-19 infections. Moreover, mortality was delineated in this database through ADT demarcations of expiration or death, and the completeness of such data was dependent on the governing institutional policies.
This nationally representative cohort study supported the findings of smaller, more geographically focused studies and found decreases in in-hospital mortality across all age groups throughout the pandemic period. Reductions in mortality rates did not appear to be associated with the age distribution of hospitalized patients with SARS-CoV-2–positive tests and were likely because of new therapies and improvements in the clinical management of patients with SARS-CoV-2 infection.
Accepted for Publication: February 27, 2021.
Published: April 8, 2021. doi:10.1001/jamanetworkopen.2021.6556
Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License. © 2021 Finelli L et al. JAMA Network Open.
Corresponding Author: Lyn Finelli, DrPH, MS, Outcomes Research, CORE, Merck Research Labs, 351 North Sumneytown Pike, North Wales, PA 19454 (lynn.finelli@merck.com).
Author Contributions: Drs Gupta and Puzniak had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Finelli, Gupta, Sutter, Yu, Puzniak.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Finelli, Gupta, Yu, Puzniak.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Finelli.
Obtained funding: Gupta.
Administrative, technical, or material support: Finelli, Gupta, Yu, Sutter, Puzniak.
Supervision: Gupta, Yu, Puzniak.
Conflict of Interest Disclosures: Drs Finelli, Petigara, Bauer, and Puzniak reported being employees of Merck Sharpe and Dohme, a subsidiary of Merck and Co, Inc, and they may own stock and/or hold stock options in Merck and Co, Inc. Drs Gupta and Yu reported being employees of Becton, Dickinson and Company (BD), and they may own stock and/or hold stock options in Becton, Dickson and Company.
Funding/Support: Funding for this research was provided by Merck Sharp and Dohme, a subsidiary of Merck and Co, Inc.
Role of the Funder/Sponsor: Drs Finelli, Petigara, Bauer, and Puzniak are employees of the funding source and played a role in the design and conduct of the study; interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Medical writing and editorial assistance was provided by Alanna Kennedy, PhD, CMPP, of The Lockwood Group. This assistance was funded by Merck Sharp and Dohme Corp, a subsidiary of Merck and Co, Inc.
Additional Information: The data sharing policy of Merck Sharp and Dohme Corp, a subsidiary of Merck and Co, Inc, including restrictions, is available at http://engagezone.msd.com/ds_documentation.php. Requests for access to the study data can be submitted through the EngageZone site or via email to dataaccess@merck.com.
1.Richardson
S, Hirsch
JS, Narasimhan
M,
et al; the Northwell COVID-19 Research Consortium. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area.
JAMA. 2020;323(20):2052-2059. doi:
10.1001/jama.2020.6775PubMedGoogle ScholarCrossref 7.Rossen
LM, Branum
AM, Ahmad
FB, Sutton
P, Anderson
RN. Excess deaths associated with COVID-19, by age and race and ethnicity—United States, January 26-October 3, 2020.
MMWR Morb Mortal Wkly Rep. 2020;69(42):1522-1527. doi:
10.15585/mmwr.mm6942e2PubMedGoogle ScholarCrossref 8.von Elm
E, Altman
DG, Egger
M, Pocock
SJ, Gøtzsche
PC, Vandenbroucke
JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.
Lancet. 2007;370(9596):1453-1457. doi:
10.1016/S0140-6736(07)61602-XPubMedGoogle ScholarCrossref 9.McCann
E, Srinivasan
A, DeRyke
CA,
et al. Carbapenem-nonsusceptible gram-negative pathogens in ICU and non-ICU settings in US hospitals in 2017: a multicenter study.
Open Forum Infect Dis. 2018;5(10):ofy241. doi:
10.1093/ofid/ofy241PubMedGoogle Scholar 13.Sterne
JAC, Murthy
S, Diaz
JV,
et al; WHO Rapid Evidence Appraisal for COVID-19 Therapies (REACT) Working Group. Association between administration of systemic corticosteroids and mortality among critically ill patients with COVID-19: a meta-analysis.
JAMA. 2020;324(13):1330-1341. doi:
10.1001/jama.2020.17023PubMedGoogle ScholarCrossref 15.Petrilli
CM, Jones
SA, Yang
J,
et al. Factors associated with hospital admission and critical illness among 5279 people with coronavirus disease 2019 in New York City: prospective cohort study.
BMJ. 2020;369:m1966. doi:
10.1136/bmj.m1966PubMedGoogle ScholarCrossref 19.Chaudhry
R, Dranitsaris
G, Mubashir
T, Bartoszko
J, Riazi
S. A country level analysis measuring the impact of government actions, country preparedness and socioeconomic factors on COVID-19 mortality and related health outcomes.
EClinicalMedicine. 2020;25:100464. doi:
10.1016/j.eclinm.2020.100464PubMedGoogle Scholar