Updated: 16 Nov 2020

Author: Marc Bevand

This project studies the age-stratified infection fatality ratio (IFR) of COVID-19:

• compare COVID-19 to seasonal influenza (flu)
• calculate the expected overall IFR based on countries' population pyramids
• calculate the age-stratified IFR of COVID-19 from the Spanish ENE-COVID serosurvey

Comparing COVID-19 to seasonal influenza

The above chart compares the IFR of COVID-19 to the IFR of seasonal influenza. We find that COVID-19 is definitely significantly more fatal than influenza at all ages above 30 years. The source code producing this chart is covid_vs_flu.py.

The vertical indicators represent the difference in fatality between COVID-19 and influenza at various ages, from 30 to 80 years at 10-year intervals. The top/bottom of the indicators are anchored at the geometric means of the COVID-19/influenza IFR estimates.

The COVID-19 IFR curves represent these estimates:

1. ENE-COVID Spanish serosurvey (calculated by calc_ifr.py, see this section)
2. US CDC COVID-19 Pandemic Planning Scenarios (table 1); which is based on Estimation of SARS-CoV-2 mortality during the early stages of an epidemic: A modeling study in Hubei, China, and six regions in Europe
3. Verity et al.: Estimates of the severity of coronavirus disease 2019: a model-based analysis (table 1)
4. Levin et al.: Assessing the age specificity of infection fatality rates for COVID-19: systematic review, meta-analysis, and public policy implications (table 3)
5. Perez-Saez et al.: Serology-informed estimates of SARS-CoV-2 infection fatality risk in Geneva, Switzerland
6. Poletti et al.: Age-specific SARS-CoV-2 infection fatality ratio and associated risk factors, Italy, February to April 2020 (table 1, column "Any time")
7. Picon et al.: Coronavirus Disease 2019 Population-based Prevalence, Risk Factors, Hospitalization, and Fatality Rates in Southern Brazil (table 2)
8. Gudbjartsson et al.: Humoral Immune Response to SARS-CoV-2 in Iceland, specifically Supplementary Appendix 1 (table S7)
9. PHAS - Public Health Agency of Sweden: The infection fatality rate of COVID-19 in Stockholm – Technical report (table B.1)
10. O’Driscoll et al.: Age-specific mortality and immunity patterns of SARS-CoV-2 infection in 45 countries (table S4)
11. Ward et al.: Antibody prevalence for SARS-CoV-2 in England following first peak of the pandemic: REACT2 study in 100,000 adults, specifically Supplementary Appendix (table S2a, column "Based on confirmed COVID-19 deaths")
12. Yang et al.: Estimating the infection fatality risk of COVID-19 in New York City during the spring 2020 pandemic wave (table 1)
13. Molenberghs et al.: Belgian Covid-19 Mortality, Excess Deaths, Number of Deaths per Million, and Infection Fatality Rates (table 6)

The seasonal influenza IFR curves represent data from the US CDC on multiple seasons of flu:

1. 2019-2020 influenza burden
2. 2018-2019 influenza burden
3. 2017-2018 influenza burden
4. 2016-2017 influenza burden
5. 2015-2016 influenza burden
6. 2014-2015 influenza burden

However, these CDC statistics (eg. table 1 in "2018-2019 influenza burden",) only give the estimated number of symptomatic illnesses. We must account for asymptomatic ones as well to calculate the IFR.

In Key Facts About Influenza (Flu) the CDC implies 55-60% of illnesses are symptomatic:

«on average, about 8% of the U.S. population gets sick from flu each season, with a range of between 3% and 11%, depending on the season. [...] The commonly cited 5% to 20% estimate was based on a study that examined both symptomatic and asymptomatic influenza illness, which means it also looked at people who may have had the flu but never knew it because they didn’t have any symptoms. The 3% to 11% range is an estimate of the proportion of people who have symptomatic flu illness.»

Thus, the CDC acknowledges that 55-60% of illnesses are symptomatic (3/5 = 60%, and 11/20 = 55%.) We use the mid-point, 57.5%, to infer the number of asymptomatic illnesses:

total_illnesses = symptomatic_illnesses / .575

Age-stratified IFR applied to countries' population pyramids

The script apply_ifr.py uses a handful of age-stratified IFR estimates for COVID-19 and the seasonal flu and applies them to countries' population pyramids, to find their expected average IFR. The calculation assumes equal prevalence of the disease among all age groups.

IFR estimates are a subset of the same sources as in covid_vs_flu.py. The flu IFR is from the US CDC (last flu season, 2019-2020.)

The real-world overall IFR will, of course, dependent on many factors: varying prevalence among age groups, underlying health conditions, access to healthcare, socioeconomic status, ethnicity, etc.

Data for the population pyramids comes from the United Nations, specifically the first sheet of Population by Age Groups - Both Sexes. This excel file was converted to CSV format: WPP2019_POP_F07_1_POPULATION_BY_AGE_BOTH_SEXES.csv

Results

The overall expected IFR percentages are summarized in this table (sorted on the ENE-COVID column):

ENE-COVID	COVID: US CDC	COVID: Verity	COVID: Levin	Flu: US CDC	Region
1.274	1.311	1.605	2.660	0.125	Japan
1.065	1.092	1.382	2.177	0.106	Italy
1.041	1.043	1.339	2.135	0.102	Greece
0.993	1.012	1.305	1.997	0.100	Germany
0.984	1.045	1.320	2.027	0.104	Portugal
0.931	0.985	1.270	1.938	0.100	Martinique
0.919	0.933	1.221	1.953	0.095	Lithuania
0.916	0.942	1.207	1.950	0.093	Spain
0.914	0.941	1.201	1.899	0.095	France
0.899	1.008	1.248	1.819	0.102	Finland
0.881	0.919	1.192	1.910	0.095	Latvia
0.875	0.957	1.205	1.789	0.096	Puerto Rico
0.868	0.892	1.171	1.773	0.093	Estonia
0.865	0.932	1.209	1.733	0.098	Croatia
0.855	0.965	1.205	1.716	0.099	Malta
0.846	0.881	1.137	1.740	0.089	Belgium
0.843	0.905	1.181	1.740	0.095	Slovenia
0.840	0.943	1.154	1.722	0.093	Sweden
0.836	0.910	1.149	1.709	0.090	Austria
0.825	0.897	1.132	1.697	0.089	Switzerland
0.810	0.861	1.117	1.633	0.089	Europe
0.802	0.909	1.141	1.625	0.093	Netherlands
0.801	0.949	1.181	1.592	0.098	Bulgaria
0.798	0.864	1.113	1.637	0.090	Guadeloupe
0.797	0.935	1.136	1.593	0.093	Denmark
0.794	0.868	1.089	1.622	0.087	United Kingdom
0.791	0.811	1.102	1.607	0.087	China, Hong Kong SAR
0.759	0.830	1.090	1.494	0.089	Romania
0.753	0.854	1.101	1.534	0.093	Hungary
0.745	0.800	1.064	1.502	0.087	Poland
0.739	0.828	1.059	1.538	0.084	Channel Islands
0.738	0.878	1.092	1.510	0.092	Czechia
0.734	0.810	1.047	1.505	0.085	Canada
0.719	0.756	0.998	1.480	0.079	Barbados
0.713	0.808	1.038	1.403	0.084	Curaçao
0.709	0.802	1.007	1.447	0.082	Norway
0.680	0.800	1.029	1.337	0.088	Serbia
0.676	0.694	0.904	1.407	0.071	Uruguay
0.675	0.745	0.967	1.379	0.079	Northern America
0.671	0.736	0.944	1.374	0.076	Australia
0.671	0.733	0.983	1.271	0.080	Ukraine
0.669	0.738	0.958	1.365	0.078	United States of America
0.659	0.757	1.011	1.243	0.085	Bosnia and Herzegovina
0.654	0.742	0.945	1.323	0.077	New Zealand
0.644	0.735	0.941	1.314	0.077	Cuba
0.641	0.877	1.059	1.308	0.094	United States Virgin Islands
0.633	0.725	0.963	1.259	0.077	Republic of Korea
0.628	0.686	0.938	1.280	0.077	China, Taiwan Province of China
0.624	0.663	0.916	1.191	0.074	Russian Federation
0.624	0.657	0.921	1.252	0.075	Belarus
0.615	0.668	0.881	1.261	0.070	Luxembourg
0.607	0.670	0.896	1.155	0.073	Georgia
0.605	0.683	0.888	1.236	0.073	Iceland
0.604	0.709	0.935	1.222	0.079	Slovakia
0.582	0.667	0.891	1.138	0.075	Montenegro
0.563	0.653	0.838	1.126	0.070	Ireland
0.561	0.649	0.839	1.106	0.069	Cyprus
0.548	0.659	0.857	1.079	0.071	Albania
0.531	0.652	0.874	1.049	0.072	Aruba
0.518	0.580	0.754	1.052	0.062	Oceania
0.510	0.588	0.795	1.017	0.065	Thailand
0.500	0.569	0.763	1.020	0.063	Réunion
0.499	0.545	0.710	1.021	0.060	Israel
0.497	0.522	0.748	0.915	0.059	Armenia
0.497	0.616	0.816	0.968	0.070	North Macedonia
0.493	0.550	0.736	1.001	0.061	Chile
0.479	0.550	0.796	0.952	0.067	Singapore
0.454	0.515	0.671	0.917	0.056	Argentina
0.449	0.514	0.738	0.887	0.062	China, Macao SAR
0.447	0.544	0.732	0.885	0.063	Mauritius
0.435	0.501	0.721	0.876	0.062	Republic of Moldova
0.415	0.511	0.681	0.820	0.058	Trinidad and Tobago
0.413	0.475	0.638	0.823	0.053	Costa Rica
0.413	0.468	0.626	0.818	0.051	Saint Lucia
0.409	0.507	0.695	0.806	0.061	China
0.394	0.497	0.648	0.822	0.052	Antigua and Barbuda
0.387	0.495	0.651	0.762	0.056	Sri Lanka
0.377	0.487	0.635	0.706	0.050	Dem. People's Republic of Korea
0.376	0.437	0.594	0.743	0.050	Brazil
0.365	0.464	0.611	0.703	0.052	Guam
0.364	0.425	0.569	0.725	0.047	Jamaica
0.361	0.454	0.605	0.764	0.051	Saint Vincent and the Grenadines
0.359	0.399	0.540	0.724	0.045	Panama
0.358	0.418	0.566	0.709	0.048	WORLD
0.356	0.410	0.555	0.710	0.047	Latin America and the Caribbean
0.355	0.412	0.560	0.707	0.047	Colombia
0.354	0.438	0.595	0.740	0.050	Grenada
0.347	0.399	0.528	0.687	0.045	El Salvador
0.343	0.363	0.528	0.701	0.043	Viet Nam
0.342	0.408	0.552	0.675	0.047	Turkey
0.339	0.397	0.537	0.671	0.045	Peru
0.336	0.394	0.553	0.664	0.047	Tunisia
0.327	0.432	0.573	0.616	0.049	New Caledonia
0.325	0.394	0.541	0.640	0.047	Asia
0.310	0.350	0.463	0.622	0.039	Bolivia (Plurinational State of)
0.309	0.348	0.480	0.616	0.040	Dominican Republic
0.307	0.348	0.503	0.574	0.042	Kazakhstan
0.306	0.352	0.485	0.615	0.041	Mexico
0.305	0.365	0.502	0.604	0.042	Venezuela (Bolivarian Republic of)
0.303	0.347	0.474	0.607	0.040	Ecuador
0.295	0.346	0.482	0.576	0.041	Lebanon
0.293	0.400	0.542	0.527	0.045	Seychelles
0.287	0.384	0.527	0.579	0.046	French Polynesia
0.282	0.326	0.456	0.521	0.038	Guyana
0.280	0.335	0.463	0.536	0.039	Suriname
0.273	0.313	0.474	0.515	0.039	Azerbaijan
0.272	0.337	0.474	0.522	0.041	Morocco
0.269	0.358	0.495	0.517	0.042	Bahamas
0.263	0.325	0.453	0.504	0.039	Malaysia
0.260	0.305	0.426	0.512	0.037	Algeria
0.257	0.307	0.418	0.500	0.037	Paraguay
0.251	0.297	0.403	0.486	0.035	Bhutan
0.246	0.301	0.431	0.476	0.037	Iran (Islamic Republic of)
0.234	0.293	0.414	0.452	0.037	India
0.229	0.293	0.416	0.437	0.036	Indonesia
0.228	0.259	0.373	0.452	0.032	Nicaragua
0.221	0.269	0.368	0.428	0.031	Bangladesh
0.221	0.232	0.349	0.432	0.029	Cabo Verde
0.217	0.284	0.375	0.342	0.032	Tonga
0.215	0.277	0.401	0.412	0.036	Myanmar
0.208	0.235	0.334	0.406	0.029	Belize
0.206	0.235	0.331	0.407	0.029	Honduras
0.204	0.233	0.323	0.403	0.029	Guatemala
0.204	0.255	0.360	0.388	0.032	Philippines
0.202	0.267	0.360	0.388	0.033	Nepal
0.197	0.249	0.381	0.373	0.032	Brunei Darussalam
0.194	0.237	0.331	0.376	0.030	Haiti
0.193	0.251	0.355	0.366	0.032	South Africa
0.193	0.222	0.339	0.360	0.029	Turkmenistan
0.192	0.247	0.343	0.363	0.031	Egypt
0.192	0.262	0.376	0.354	0.033	Fiji
0.188	0.220	0.335	0.353	0.029	Kyrgyzstan
0.188	0.245	0.353	0.392	0.031	French Guiana
0.187	0.219	0.340	0.359	0.029	Uzbekistan
0.185	0.224	0.323	0.361	0.029	Syrian Arab Republic
0.182	0.226	0.324	0.349	0.028	Libya
0.181	0.229	0.319	0.350	0.029	Lesotho
0.173	0.216	0.320	0.333	0.027	Mongolia
0.170	0.225	0.317	0.324	0.028	Djibouti
0.170	0.238	0.325	0.299	0.029	Samoa
0.168	0.221	0.314	0.325	0.029	Cambodia
0.164	0.210	0.292	0.310	0.026	Pakistan
0.158	0.204	0.273	0.303	0.026	Eritrea
0.157	0.183	0.277	0.285	0.024	Maldives
0.157	0.194	0.268	0.324	0.024	Mayotte
0.156	0.199	0.279	0.294	0.025	Jordan
0.155	0.203	0.289	0.296	0.027	Botswana
0.154	0.199	0.287	0.289	0.026	Lao People's Democratic Republic
0.152	0.209	0.279	0.285	0.025	Timor-Leste
0.146	0.180	0.282	0.281	0.024	Saudi Arabia
0.145	0.187	0.254	0.271	0.024	Eswatini
0.139	0.173	0.248	0.265	0.023	Namibia
0.135	0.176	0.302	0.246	0.024	Kuwait
0.134	0.173	0.244	0.254	0.023	Sudan
0.133	0.173	0.245	0.248	0.023	Gabon
0.130	0.167	0.231	0.244	0.022	Ethiopia
0.130	0.174	0.242	0.252	0.023	Solomon Islands
0.128	0.158	0.245	0.233	0.022	Tajikistan
0.127	0.165	0.236	0.239	0.022	Africa
0.125	0.157	0.229	0.240	0.022	Iraq
0.125	0.150	0.253	0.224	0.021	Bahrain
0.124	0.163	0.225	0.230	0.021	South Sudan
0.123	0.168	0.244	0.216	0.023	Vanuatu
0.122	0.165	0.247	0.227	0.023	Papua New Guinea
0.121	0.157	0.226	0.228	0.021	Liberia
0.121	0.157	0.222	0.229	0.021	Benin
0.120	0.155	0.224	0.224	0.021	Mauritania
0.120	0.157	0.222	0.220	0.021	State of Palestine
0.119	0.150	0.219	0.191	0.020	Sao Tome and Principe
0.118	0.178	0.278	0.250	0.027	Micronesia (Fed. States of)
0.118	0.138	0.222	0.222	0.020	Oman
0.118	0.160	0.260	0.214	0.024	Western Sahara
0.116	0.157	0.230	0.212	0.021	Ghana
0.116	0.146	0.216	0.218	0.020	Madagascar
0.113	0.149	0.220	0.217	0.021	Comoros
0.112	0.142	0.205	0.211	0.020	Zimbabwe
0.112	0.146	0.216	0.218	0.021	Rwanda
0.111	0.146	0.208	0.206	0.020	Senegal
0.110	0.144	0.203	0.205	0.019	Democratic Republic of the Congo
0.107	0.139	0.202	0.196	0.020	Yemen
0.106	0.141	0.203	0.195	0.019	Sierra Leone
0.103	0.165	0.239	0.217	0.023	Kiribati
0.103	0.136	0.192	0.189	0.019	Mozambique
0.102	0.135	0.192	0.189	0.019	Somalia
0.102	0.139	0.202	0.187	0.019	Togo
0.101	0.137	0.201	0.185	0.019	Congo
0.101	0.133	0.191	0.186	0.019	Central African Republic
0.100	0.134	0.194	0.185	0.019	Guinea
0.100	0.135	0.197	0.184	0.019	Côte d'Ivoire
0.098	0.130	0.189	0.178	0.018	Cameroon
0.097	0.130	0.192	0.178	0.019	Guinea-Bissau
0.097	0.129	0.183	0.176	0.018	Malawi
0.096	0.129	0.186	0.175	0.018	United Republic of Tanzania
0.096	0.128	0.184	0.175	0.018	Afghanistan
0.094	0.122	0.186	0.172	0.018	Kenya
0.094	0.130	0.190	0.171	0.019	Nigeria
0.092	0.118	0.179	0.167	0.018	Equatorial Guinea
0.091	0.126	0.177	0.162	0.017	Gambia
0.091	0.116	0.170	0.168	0.017	Chad
0.090	0.112	0.203	0.158	0.018	Qatar
0.088	0.121	0.172	0.162	0.017	Niger
0.088	0.118	0.173	0.160	0.017	Burkina Faso
0.088	0.109	0.169	0.161	0.017	Burundi
0.087	0.117	0.169	0.162	0.017	Mali
0.085	0.111	0.164	0.155	0.016	Angola
0.083	0.104	0.190	0.146	0.016	United Arab Emirates
0.083	0.107	0.158	0.150	0.016	Zambia
0.074	0.099	0.147	0.136	0.015	Uganda

Note that in addition to countries, there are rows for each continent and for the world.

Findings

The overall IFR estimates of COVID-19, with the exception of Levin et al., are relatively consistent with each other, usually within 30-40%. Levin et al. is often up to 2-fold higher than the others, depending on the country.

The country with the oldest population is expected to have the highest overall IFR: Japan at 1.274-1.605% (excluding Levin et al.)

The country with the youngest population is expected to have the lowest overall IFR: Uganda at 0.074-0.147%.

The overall IFR varies dramatically by more than 10-fold between countries with a young population and those with an old population.

In fact, the young age of the population in Africa is a major factor explaining the relatively small number of deaths on this continent. We find (ENE-COVID) IFR=0.127% for Africa, and IFR=0.810% in Europe, a 6-fold difference.

The overall IFR of COVID-19 is, for each world region and the world:

• 0.810% Europe
• 0.675% Northern America
• 0.518% Oceania
• 0.356% Latin America and the Caribbean
• 0.325% Asia
• 0.127% Africa
• 0.358% World

Our code, with the ENE-COVID Spanish serosurvey data from June 2020, accurately predicted an overall IFR of 0.669% in the United States, which is very close to overall US CDC estimate of 0.65% published in July 2020.

The IFR of COVID-19 is one order of magnitude (10×) higher than the seasonal flu for all regions. For example, in the US the average flu IFR is 0.078%, compared to 0.669-1.169% for COVID-19.

Calculating the age-stratified IFR of COVID-19 from the Spanish ENE-COVID study

One of the largest serological prevalence surveys of COVID-19 was conducted by Spain during the second round of the ENE-COVID study that analyzed 63 564 samples between 18 May 2020 and 01 June 2020. We used its provisional results published on 03 June to calculate the overall and age-stratified IFR of COVID-19 with the Python script calc_ifr.py:

$ ./calc_ifr.py
Ages  0 to   9:  115013 infected,     4 deaths,  0.003% IFR
Ages 10 to  19:  177929 infected,     7 deaths,  0.004% IFR
Ages 20 to  29:  212099 infected,    32 deaths,  0.015% IFR
Ages 30 to  39:  281290 infected,    86 deaths,  0.030% IFR
Ages 40 to  49:  447942 infected,   287 deaths,  0.064% IFR
Ages 50 to  59:  410213 infected,   874 deaths,  0.213% IFR
Ages 60 to  69:  334709 infected,  2404 deaths,  0.718% IFR
Ages 70 to  79:  270572 infected,  6451 deaths,  2.384% IFR
Ages 80 to  89:  131703 infected, 11150 deaths,  8.466% IFR
Ages 90 to 199:   46631 infected,  5827 deaths, 12.497% IFR
Ages  0 to 199: 2428102 infected, 27121 deaths,  1.117% IFR

The average IFR for Spain is 1.117%. However the true IFR may be higher due to right-censoring, under-reporting of deaths, or low specificity of the serological test; or the true IFR may be lower due to low sensitivity of the serological test.

The age-stratified IFR was calculated from three sources:

1. Detailed prevalence data for age brackets, from the serosurvey (table 1)
2. Total deaths and deaths per age bracket from the Ministry of Health's daily report for 29 May (table 2 and table 3)
3. Population pyramid for Spain, from worldpopulationreview.com

In order to minimize right-censoring (deaths lagging infections,) the parameters total deaths and deaths per age bracket should be obtained from a point in time as close as possible to when the serosurvey was conducted (18 May to 01 June, preferably closer to the mid-point 25 May.) This is because the seroconversion time is roughly the same as the time between infection and death. We found only two Ministry of Health reports in this time period that document deaths per age bracket: 18 May, 29 May. However the Ministry of Health has made significant corrections to deaths statistics on 25 May by subtracting approximately 2 000 deaths. Therefore we trusted the statistics from 29 May over those of 18 May. Furthermore, 29 May is closer to the mid-point.

Important detail to note: there were 27 121 total deaths, however age information was only available for 20 585 deaths, and was missing for 6 536 deaths. We assume that these 6 536 deaths were distributed proportionally—not equally—among age brackets, which seems to be a reasonable assumption.

Regarding the specificity of the commercial test used (COVID-19 IgG Rapid Test Cassette by Zhejiang Orient Gene Biotech Co Ltd) we found various claims, all 100% or close, so no significant false positives are expected:

• 100% claimed by the manufacturer (serosurvey, page 3)
• 100% measured by the Ministry of Health (serosurvey, page 3)
• 99.2% measured by a third-party

However the sensitivity is more uncertain:

• 97% claimed by the manufacturer (serosurvey, page 3)
• 79% measured by the Ministry of Health (serosurvey, page 3)
• 93.1% measured by a third-party

So a false negative rate anywhere from 3% to 21% could be possible, and we think it is premature to adjust IFR calculations given the exact sensitivity is not known.