The in-hospital death rate among adult COVID-19 patients in England early in the pandemic was 31% but declined significantly over time, with older age, male sex, low socioeconomic status, Asian or mixed ethnicity, and underlying conditions signaling poor outcomes, according to a retrospective, observational study published yesterday in The Lancet Respiratory Medicine.
A team led by a researcher from University College London used the National Health Service Hospital Episode Statistics administrative dataset to estimate in-hospital deaths and contributing factors among 91,541 COVID-19 patients at 500 hospitals from Mar 1 to May 31, 2020.
Of the 91,541 patients, 28,200 (30.8%) died in the hospital, with COVID-19 listed as the cause of death for 23,153 (88.2%) of 26,240 with a cause listed. The adjusted in-hospital death rate began to improve in early March to late May, falling from 52.2% to 16.8%.
Significant predictors of death were older age, male sex (odds ratio [OR], 1.5), low socioeconomic status (OR, 1.0), Asian ethnicity (OR, 1.2), mixed ethnicity (OR, 1.3), and all underlying illnesses other than mild liver disease and peptic ulcer, especially moderate to severe liver disease (OR, 5.4). Later date of hospital release was tied to a lower risk of death.
The most common underlying conditions were diabetes (27.7%), chronic lung disease (25.9%), kidney disease (18.3%), dementia (15.3%), and congestive heart failure (13.9%). The illnesses associated with the highest death rates by proportion of cases included moderate to severe liver disease, congestive heart failure, dementia, and kidney disease.
Median hospital length of stay rose steadily from Mar 1 to May 31, from 2 to 10 days in survivors and 4.5 to 11 days in those who died in the hospital. Over the study period, median length of hospitalization was 7 days.
Hospital strategy, clinical care improvements
The authors noted that the higher death rate in Asian patients and those of mixed ethnicity in their study, compared with those of previous studies, could partly have been caused by higher infection rates, the tendency toward severe illness, and differences in health-seeking behaviors.
"Additionally, public health messaging regarding prevention, early diagnosis, and treatment of COVID-19 might have been less effective in certain ethnic minority groups, leading to later presentation," they wrote.
The researchers said that their results don't prove that high death rates in the early phase of the pandemic were caused by a surge of COVID-19 patients that strained hospitals, because adjusted in-hospital death rates dropped even as coronavirus infections rose in the first half of April, especially among older, frail, white, and female patients. Death rates were halved over time in patients 80 years and older.
They added that the improved death rate over time could be attributed to advances in hospital strategy and clinical care as more became known about COVID-19; thus, the reasons for the lower death rates should be investigated to inform the response to future outbreaks.
"Learning from individual hospitals with the greatest levels of improvement, or consistently good performance throughout, could help to support others to improve," the researchers said. "Further examination of these temporal trends—with consideration of changes in public health strategy, central directives from arm's-length bodies, alterations in clinical processes, and availability of local health-care resources—will be vital to inform preparations and adjustments for ongoing and future case surges."
Implications for prioritizing interventions
In a commentary in the same journal, Purnema Madahar, MD, and Daniel Brodie, MD, of Columbia University, and Hannah Wunsch, MD; Prabhat Jha, MD, DPhil; and Arthur Slutsky, MD, of the University of Toronto, said that the initial high rate of COVID-19 death could have been caused in part by high community rates of infection coinciding with low levels of physical distancing and wearing of face coverings, as well as possible delays in older adults reaching the hospital.
"Even the perception of a strained health system can lead to unintentional excess deaths, from COVID-19 and other conditions, because individuals might avoid seeking care until later stages of disease or might die at home, leading to underestimates of the true mortality burden attributable to COVID-19," Madahar and colleagues wrote.
The commentary authors called for future studies linking data on community antibody prevalence, community and patient viral loads, the role of viral spread from healthcare workers to inpatients, and surveillance data to model the progression of the pandemic, determine the resulting case-fatality risk, and prioritize community interventions.
"As we continue to battle COVID-19, identifying high-risk patients in hospital and community settings will be crucial, as will insights from population-based studies, helping to focus our community-based and hospital-based public health initiatives," they said.