Ước tính gánh nặng COVID-19

Estimating unreported cases, hospitalizations, and deaths helps to quantify the impact and severity of the COVID-19 pandemic on the healthcare system and society. Additionally, these estimates can inform how to direct and allocate healthcare resources; assist in planning for prevention and control measures, including vaccination; predict the future burden of COVID-19; and evaluate the potential impact of interventions.

The cumulative burden of COVID-19 is an estimate of the number of people who may have been infected, sick, hospitalized, or died as a result of a COVID-19 infection in the United States.  Confirmed COVID-19 cases and deaths are nationally reported, but these cases and deaths likely represent only a fraction of the true number that have occurred in the population. COVID-19 infections, symptomatic illnesses, hospitalizations, and deaths might be underdetected and go unreported for a variety of reasons.  Ví dụ:

• Some people infected with SARS-CoV-2 never show symptoms (asymptomatic infection), so their infection will likely go undetected.
• Báo cáo số ca bệnh được gửi đến CDC thường thiếu thông tin bệnh nhân, như tuổi tác hoặc tình trạng nhập viện, hoặc bị chậm trễ.
• Not everyone who is sick will seek medical care and/or be tested; and patients may not be tested for COVID-19 while hospitalized or if they die.
• Even if a sick outpatient or hospitalized patient is tested, an infection with COVID-19 may not be accurately captured if, for example:
  • the test was not completed correctly or within the appropriate timeframe for capturing infection;
  • kết quả xét nghiệm âm tính giả đối với tình trạng lây nhiễm COVID-19 do độ nhạy của xét nghiệm;
  • the test result was falsely negative for a COVID-19 infection due to the quality or quantity of the specimen collected; or
  • the confirmed COVID-19 case was never reported to the local and state public health agency and then to CDC.

• For patients with COVID-19, death can occur several days or weeks after being tested and reported, and the death might be incorrectly attributed to a cause other than COVID-19 because of the time between testing positive and death.
• COVID-19 may result in non-respiratory complications or it might increase the severity of chronic conditions, which can lead to death (e.g., sepsis, circulatory diseases, respiratory diseases, diabetes, or renal failure), and COVID-19 might be incorrectly omitted as a contributing cause of death on the death certificate.

Because current surveillance systems do not capture all cases or deaths of COVID-19 occurring in the United States, CDC provides these estimates to better reflect the larger burden of COVID-19. CDC sử dụng các loại ước tính này để thông báo các quyết định về chính sách và thông điệp công cộng.

To estimate COVID-19 infections, symptomatic illnesses, and hospitalizations, CDC uses a statistical model applied to confirmed cases of COVID-19, adjusted for missing age and hospitalization status. Several data sources and surveillance systems are used to identify and characterize potential sources of underdetection, which include:

• SARS-CoV-2 test sensitivity is lower. Những người được xét nghiệm SARS-CoV-2, vi-rút gây bệnh COVID-19, có thể không có xét nghiệm dương tính ngay cả khi đã bị nhiễm vi-rút do xét nghiệm được sử dụng có độ nhạy thấp hơn. SARS-CoV-2 test sensitivity has been reported in the literature; a range of 79%-98% sensitivity for RT-PCR tests is used to account for false negative test results.

• SARS-CoV-2 test is not ordered or not completed in a timely manner. Not all outpatients who seek care for acute respiratory illness (ARI) or inpatients hospitalized with ARI are tested for COVID-19, and not all ordered tests are correctly completed in a timely manner. CDC uses two data sources to approximate how many outpatients with ARI are not tested for COVID-19:
  • Cơ sở dữ liệu Hồ sơ Sức khỏe Điện tử IBM Watson Explorys, một kho lưu trữ dữ liệu hồ sơ sức khỏe điện tử từ hơn 39 đối tác y tế, 400 cơ sở chăm sóc cấp tính và 400.000 nhà cung cấp độc nhất; và
  • COVID Near Youexternal icon (CNY), a website application launched by Harvard University in March 2020 where participants can submit information on self-reported symptoms, efforts to obtain health care, and COVID-19 testing.
• Not all patients with symptoms seek care or testing services. Not all sick patients seek care or are tested for COVID-19, and therefore they are not included in national case reports.  To approximate the number of symptomatic people who never sought medical care, researchers use data from COVID Near Youexternal icon (CNY) and Flu Near Youexternal icon (FNY) sites on health care seeking behaviors. Trong khi COVID Near You mới ra mắt vào tháng 3 năm 2020, thì FNY đã thu thập dữ liệu liên quan đến bệnh cúm tự báo cáo kể từ năm 2011.
• Patients do not have symptoms. Some people infected with SARS-CoV-2 never show symptoms (they have asymptomatic infection). People with asymptomatic infection are very likely to go undetected. Tỷ lệ lây nhiễm không có triệu chứng được báo cáo trong tài liệu và thay đổi theo nhóm tuổi. In people 0-64 years old, a range of 5%-24% is used to estimate asymptomatic infections, and for people 65 years and older, a range of 5%-32% is used.

The statistical model used to adjust hospitalized and non-hospitalized case counts for the above sources of underdetection is based on methods that have been previously used to estimate the disease burden of influenza, detailed elsewhere. These methods are peer-reviewed and published in Clinical Infectious Diseases.

COVID-19 deaths are estimated using a statistical model to calculate the number of COVID-19 deaths that were unrecognized and those that were not recorded on death certificates and, as a result, were never reported as a death related to COVID-19.

To estimate these unrecognized COVID-19 deaths, all-cause deaths are obtained from the National Center of Health Statistics.  Before applying the statistical model, reported COVID-19 deaths are subtracted by age, state, and week from all-cause deaths, so that these reported COVID-19 deaths are not included in the calculation of the expected deaths for the statistical model.

Then, to understand how many deaths may have not been recognized as being related to COVID-19, CDC uses a statistical model to estimate the number of expected deaths from all causes assuming that there was no circulation of COVID-19 (that is, those deaths expected in the absence of any COVID-19 illnesses). Researchers then use the model to predict the number of all-cause deaths that would have occurred taking into account information on COVID-19 circulation,.  To obtain the number of unrecognized COVID-19 deaths, the number of expected all-cause deaths (without COVID-19 circulation) are subtracted from the number of predicted all-cause deaths (with COVID-19 circulation). The model is used to calculate estimates by state and age (for six age groups: 0-17, 18-49, 50-64, 65-74, 75-84, and ≥85 years).

Once investigators estimate unrecognized COVID-19 deaths, they add documented COVID-19 deaths to the unrecognized deaths to obtain an estimate of the total number of COVID-19-attributable deaths. These methods are peer-reviewed and published in The Lancet Regional Health – Americasexternal icon.

These estimates of COVID-19 infections, symptomatic illnesses, hospitalizations and deaths are subject to several limitations, either from the data inputs used or some statistical assumptions made in the methods. A detailed discussion of these limitations can be found in Clinical Infectious Diseasesexternal icon and The Lancet Regional Health – Americasexternal icon. CDC continues to explore data sources and statistical methodologies for estimating COVID-19 disease burden and will refine estimates over time.