implied_ifr

Implied Infection Fatality Rate (IIFR)

New July 29: This is the accompanying dataset for our write-up, Estimating True Infections: A Simple Heuristic to Measure Implied Infection Fatality Rate. We aim to update this daily as new data becomes available.

Here we introduce the concept of implied infection fatality rate (IIFR), and present the raw data used to compute the IIFR. IIFR is a metric computed by taking a region’s reported deaths and dividing it by the true infections estimate (after accounting for the lag). Knowing the true number of people who are infected with COVID-19 in the US is an essential step towards understanding the disease.

Read more about how we compute the implied IFR in write-up, Estimating True Infections.

Note: Currently, we only support the computation of IIFR for the US on a state and national level. For all other regions, we only include the implied case fatality rate, which is computed by taking the 7-day moving average of daily confirmed deaths two weeks in the future and dividing by the 7-day moving average of daily confirmed cases.

Description of Columns

• Confirmed - Total reported cases
• Deaths - Total reported deaths
• daily_deaths, daily_deaths_7day_ma - Daily reported deaths and 7-day moving average
• daily_confirmed, daily_confirmed_7day_ma - Daily reported cases and 7-day moving average
• daily_tests, daily_tests_7day_ma - Daily reported tests and 7-day moving average
• net_hosp, net_hosp_7day_ma - Daily reported net hospitalizations and 7-day moving average (positive number means more admissions, negative number means more discharges)
• pos_rate_7day_ma - daily_confirmed_7day_ma / daily_tests_7day_ma
• rt_estimate - Rough estimate of the effective reproduction number, taken by taking the pos_rate_7day_ma and dividing it by the pos_rate_7day_ma from 5 days ago (the serial interval). Formula: rt_estimate(n) = pos_rate_7day_ma(n) / pos_rate_7day_ma(n-5).
• prevalence_ratio_7day_ma - Estimate of the ratio of true infections to confirmed cases. Formula explained here.
• true_inf_est_7day_ma - Estimate of the true number of newly infected individuals on that day, after accounting for a 14-day lag. Formula: true_inf_est_7day_ma(n) = daily_confirmed_7day_ma(n+14) * prevalence_ratio_7day_ma(n+14).
• implied_ifr_7day_ma - Implied infection fatality rate, after accounting for a 28-day lag. Formula: implied_ifr_7day_ma(n) = daily_deaths_7day_ma(n+28) / true_inf_est_7day_ma(n).

For the summary files (example), we have a few additional columns:

• cur_implied_cfr - Current impleid CFR, calculating by taking deaths and dividing by true cases from 14 days ago: cur_implied_ifr(n) = daily_deaths_7day_ma(n) / daily_confirmed_7day_ma(n-14)
• total_implied_cfr - Deaths / Confirmed (total deaths / total cases)
• cur_implied_ifr - Current implied IFR, calculated by taking deaths and dividing by true infections from 28 days ago: cur_implied_ifr(n) = daily_deaths_7day_ma(n) / true_inf_est_7day_ma(n-28)
• total_infections - sum(true_inf_est_7day_ma) (Sum of new daily infections)
• total_implied_ifr - Deaths / total_infections (total deaths / total infections)
• perc_infected - total_infections / population

Data Source

• Cases and deaths: Johns Hopkins CSSE
• Tests and hospitalizations: The COVID Tracking Project