The number of new COVID-19 cases in the United States last week rose 13% to more than 393,000, approaching levels last seen during the summer peak, according to a Reuters analysis.
Deaths fell 2% to about 4,900 people for the week ended Oct. 18, according to the analysis of state and county reports. Since the outbreak started, nearly 220,000 people in the country have died and over 8.1 million have become infected with the novel coronavirus.
The United States recorded 69,478 new cases on Friday — the highest single-day total since July 24 and the fifth-highest single-day total since the start of the pandemic.
Thirty-four out of 50 states have seen cases increase for at least two weeks in a row, up from 29 the prior week. Cases are rising in Pennsylvania, Ohio, Michigan and North Carolina — all battleground states for the Nov. 3 U.S. presidential election.
In Nevada and South Dakota, more than 35% of tests came back positive for COVID-19 last week, the highest positive test rate in the country, according to data from The COVID Tracking Project, a volunteer-run effort to track the outbreak. A total of 14 states have a positive test rate of over 10%.
Nationally, the percentage of tests that came back positive rose to 5.4% from 5.0% the prior week.
The World Health Organization considers rates above 5% concerning because it suggests there are more cases in the community that have not yet been uncovered.
As of July 27, reports of COVID-19 cases and deaths are sourced to Reuters reporting in order to streamline data reconciliation efforts when discrepancies arise. Testing data continues to come from The COVID Tracking Project.
About the data: On July 27, this page began using case and mortality data collected by Reuters after previously relying on The COVID Tracking Project. By tracking data in-house, Reuters is able to account for and follow up on reporting discrepancies on a state-by-state basis. This page will continue to rely on The COVID Tracking Project’s testing data.
Reuters collates and checks this data by hand and the figures largely come from state, county and territory government/public health department websites. Reuters also occasionally sources information from press conferences, press releases and verified tweets and social media posts by state officials. While some states and counties report fresh numbers daily, others only update on weekdays or less frequently.
Reuters’ total cases and deaths include both confirmed and probable cases and deaths where data is available. If probable cases and deaths are not reported, only confirmed cases and deaths are shown.
As testing numbers are not reported in a standardized format nationally, states vary in the way they record testing figures. Some include all tests performed while others only count the number of individuals tested. The COVID Tracking Project only includes polymerase chain reaction (PCR) tests, which detect current coronavirus infections. Antibody tests, which may indicate a past exposure to the virus, are excluded from the overall count if they are reported separately.
Florida did not report data on Saturday, Oct. 10 because the state required extra time to remove duplicates due to a private lab resubmitting 400,000 results that have previously already been submitted.
Missouri reported that a database error caused an “incorrect inflation” of case figures for Saturday, Oct. 10. The state has not yet provided a revision as of Monday, Oct. 12.
Mississippi began reporting test data weekly instead of daily on Sept. 25, so complete figures for the week ended Oct. 4 are not yet available.
A mismatch in the Iowa Health Department and Covid Tracking Project’s test data for the week ending Oct. 18 has led to a discrepancy in the state’s positive test rate. It has been omitted from the testing table.
Puerto Rico did not report negative tests for the week ended Oct. 4 and has been omitted from this week’s testing table.
On Sept. 2, Massachusetts adopted a more restrictive definition for probable cases, which resulted in a significant fall in the number of cases. Reuters has reconciled all historical data to reflect this change in methodology
Cumulative deaths in Michigan fell by one on Aug. 9 as the state continued its data review to remove fatalities recorded in error.
Virginia reported a large backlog of positive infections on Aug. 7, resulting in an abnormal spike in cases.
On July 30, Minnesota began reporting the total number of people tested instead of total specimens tested. This changed the state total by more than 173,000 tests for July 31.
A spike observed in deaths in New Jersey and nationally in late June is due to New Jersey including over 1,800 probable deaths from earlier this year.
Wyoming’s spike in late April is due to the inclusion of probable cases.
Reporting by Emily Isaacman, Arundhati Sarkar, Yajush Gupta, Sabahatjahan Contractor, Chinmay Rautmare, Roshan Abraham, Lisa Shumaker, Christine Chan, Wen Foo, Aditya Munjuluru, Anurag Maan, Nikhil Subba, K. Sathya Narayanan, Ahmed Farhatha, Aniruddha Chakrabarty, Mrinalika Roy, Abhishek Manikandan, Arpit Nayak, Chaithra J, Shaina Ahluwalia, Shreyasee Raj, Shashank Nayar, Nallur Sethuraman, Harshith Aranya and Gautami Khandke. Additional reporting by David Gregorio and Aurora Ellis.
Sources: Deaths and cases data from state and local governments and health authorities; Reuters reporting; U.S. Navy; Testing data from The COVID Tracking Project
Editing by Christine Chan and Tiffany Wu