Appendix

Maximum Likelihood Estimation

In this section, we introduce the parameters, transmission rates among age groups, and estimation

process using the maximum likelihood estimation (MLE). To incorporate the history of infection and

vaccination of individuals, we categorized the population into the following groups:

• Λ 𝐼 : infected hosts,

• Λ𝑆 : uninfected and unvaccinated individuals,

• Λ 𝑉−𝑝𝑟𝑒 : uninfected and vaccinated before August 1, 2021, and

• Λ 𝑉−𝑝𝑜𝑠𝑡 : uninfected and vaccinated after August 1, 2021.

In our likelihood estimation, we did not consider the breakthrough infection group because the

available data did not contain the information whether a confirmed individual is vaccinated or not.

Since we considered eight age groups (0 to 17, 18 to 29, 30 to 39, 40 to 49, 50 to 59, 60 to 69, 70 to

79, and over 80 years old), there are 32 subgroups for MLE. Individual case data, which was provided

by the Korea Disease Control and Prevention Agency, contains the date of symptom onset and

confirmation. We assumed that the individuals are infected 4 days before the date of symptom onset.1

If the data on the date of symptom onset was not available, we assumed that the individual was

infected 8 days (2 days of latent period + 6 days of transmission period) before.2,3 Between two

discrete time (a day), from time 𝑡 − 1 to 𝑡, there are two possibilities for a host in age group 𝑖:

probability of being infected (𝑃𝐼,𝑖 (𝑡)), or not infected (𝑃𝑆,𝑖 (𝑡)). Assuming that there is a homogeneous

mixing of infectors and infectees in the community and that the distribution of the infection event is

exponential, then the probabilities are given by

𝐼(𝑡)
∑𝑗 𝛽𝑖𝑗
𝑃𝐼,𝑖 (𝑡) = 1 − 𝑒 − 𝑁 ,

𝐼(𝑡)
∑𝑗 𝛽𝑖𝑗
𝑃𝑆,𝑖 (𝑡) = 𝑒 − 𝑁 .

Note that we estimated the number of infectors (𝐼(𝑡)) by assuming that individuals can transmit the

disease 2 days after they are infected until they have confirmation. To consider vaccination, we set

that a vaccinated individual has partial vaccine effectiveness against infection (0.8 × 𝑒𝑖 ) for rounded
1
value of 1/ 𝜔𝑖 days, which is the average time to have vaccine-induced immunity considering

secondary dose, and after 1/ 𝜔𝑖 days, a vaccinated individual has full effectiveness against infection

(𝑒𝑖 ). Vaccine effectiveness against infection reduces the probability of being infected ((1 −

0.8 × 𝑒𝑖 ) × 𝑃𝐼,𝑖 (𝑡) or (1 − 𝑒𝑖 ) × 𝑃𝐼,𝑖 (𝑡)). The parameters 𝜔𝑖 and 𝑒𝑖 will be described in the next

subsection. Likelihood consists of the probabilities of not being infected and infected. If a host is in

Λ 𝐼 , the host has 𝑃𝑆,𝑖 2 days before the host was infected and has 𝑃𝐼,𝑖 a day before the host was

infected and became infected. A host in the other groups only has 𝑃𝑆,𝑖 until the final time of MLE. A

more detailed explanation of MLE process can be found in a previous work.4

The initial time of MLE is August 1, 2021, and the final time is set to December 31, 2021. During

this period, the Delta variant was dominant. Therefore, the estimated matrix 𝑀1 represents the

transmission dynamics of the Delta variant. However, 𝑀1 contains the effect of NPIs. Hence, the

matrix is adjusted to exclude this effect so that we can incorporate NPIs as a separate parameter 𝜇 in

𝑅0𝛿
the mathematical model. The adjusted matrix 𝑀2 is calculated as 𝑀2 = 𝑀1 , where 𝑅𝑒𝑠𝑡 is the
𝑅𝑒𝑠𝑡

reproductive number calculated using 𝑀1 and 𝑅0𝛿 = 6.22 is the basic reproductive number of the

Delta variant.5,6

2
Mathematical modeling of COVID-19 considering Delta and Omicron variants

The mathematical formulation of the model is as follows:

𝑋 ∈ {𝑢, 𝑤, 𝑣1, 𝑣2, 𝑤𝑣, 𝑏, 𝑤𝑏}, 𝑖 = 1, … ,8,

𝑑𝑋𝑖 𝛿 𝑜
= −𝜆𝑋,𝑖 𝑋𝑖 − 𝜆𝑋,𝑖 𝑋𝑖 − 𝑂𝑈𝑇𝑋,𝑖 + 𝐼𝑁𝑋,𝑖 ,
𝑑𝑡
𝛿 𝑜
𝑑𝐸𝑋,𝑖 𝛿 𝛿
𝑑𝐸𝑋,𝑖 𝑜 𝑜
= 𝜆𝑋,𝑖 𝑋𝑖 − 𝜅𝐸𝑋,𝑖 , = Γ𝑖 + 𝜆𝑋,𝑖 𝑋𝑖 − 𝜅𝐸𝑋,𝑖 ,
𝑑𝑡 𝑑𝑡
𝛿 𝑜
𝑑𝐼𝑋,𝑖 𝛿 𝛿
𝑑𝐼𝑋,𝑖 𝑜 𝑜
= 𝜅𝐸𝑋,𝑖 − 𝛼𝐼𝑋,𝑖 , = 𝜅𝐸𝑋,𝑖 − 𝛼𝐼𝑋,𝑖 ,
𝑑𝑡 𝑑𝑡
𝛿 𝑜
𝑑𝑄𝑋,𝑖 𝛿 𝛿
𝑑𝑄𝑋,𝑖 𝑜 𝑜
= 𝛼𝐼𝑋,𝑖 − 𝜎𝑄𝑋,𝑖 , = 𝛼𝐼𝑋,𝑖 − 𝜎𝑄𝑋,𝑖 ,
𝑑𝑡 𝑑𝑡
𝛿 𝑜
𝑑𝑀𝑋,𝑖 𝛿 𝛿 𝛿 𝛿
𝑑𝑀𝑋,𝑖 𝑜 𝑜 𝑜 𝑜
= (1 − 𝑝𝑋,𝑖 )𝜎𝑄𝑋,𝑖 − 𝛾𝑚 𝑀𝑋,𝑖 , = (1 − 𝑝𝑋,𝑖 )𝜎𝑄𝑋,𝑖 − 𝛾𝑚 𝑀𝑋,𝑖 ,
𝑑𝑡 𝑑𝑡
𝛿 𝑜
𝑑𝐶𝑋,𝑖 𝛿 𝛿 𝛿
𝑑𝐶𝑋,𝑖 𝑜 𝑜 𝑜
= 𝑝𝑋,𝑖 𝜎𝑄𝑋,𝑖 − 𝛾𝑐𝛿 𝐶𝑋,𝑖 , = 𝑝𝑋,𝑖 𝜎𝑄𝑋,𝑖 − 𝛾𝑐𝑜 𝐶𝑋,𝑖 ,
𝑑𝑡 𝑑𝑡

𝑑𝑅𝑋,𝑖 𝛿 𝛿 𝑜 𝑜 𝛿 𝛿 𝑜 𝑜
= 𝛾𝑚 𝑀𝑋,𝑖 + 𝛾𝑚 𝑀𝑋,𝑖 + (1 − 𝑓𝑋,𝑖 )𝛾𝑐𝛿 𝐶𝑋,𝑖 + (1 − 𝑓𝑋,𝑖 )𝛾𝑐𝑜 𝐶𝑋,𝑖 − 𝜁𝑋 𝑅𝑋,𝑖 ,
𝑑𝑡

𝛿
𝜆𝑋,𝑖 𝛿
= (1 − 𝜇)(1 − 𝑒𝑋,𝑖 ) ∑ ∑ 𝛽𝑖𝑗 (1 − 𝑒̃
𝛿 𝛿
𝑋,𝑗 ) 𝐼𝑋,𝑗 ,
𝑋 𝑗

𝑜
𝜆𝑋,𝑖 𝑜
= (1 − 𝜇)(1 − 𝑒𝑋,𝑖 ) ∑ ∑ 𝛽𝑖𝑗 (1 − 𝑒̃
𝑜 𝛿
𝑋,𝑗 )𝐼𝑋,𝑗 ,
𝑋 𝑗

𝑂𝑈𝑇𝑢,𝑖 = min(𝑢𝑖 , 𝜈𝑖 (𝑡)) , 𝐼𝑁𝑢,𝑖 = 0,

𝑂𝑈𝑇𝑤,𝑖 = 𝜈𝑖 (𝑡) − min(𝑢𝑖 , 𝜈𝑖 (𝑡)) , 𝐼𝑁𝑤,𝑖 = 𝜁(𝑅𝑢,𝑖 + 𝑅𝑤,𝑖 ),

𝑂𝑈𝑇𝑣1,𝑖 = 𝜔𝑖 𝑣1𝑖 , 𝐼𝑁𝑣1,𝑖 = 𝜈𝑖 (𝑡),

𝑂𝑈𝑇𝑣2,𝑖 = 𝜏𝑣,𝑖 𝑣2𝑖 + 𝑏𝑖 (𝑡) − min (𝑤𝑣𝑖 , 𝜈𝑖𝑏 (𝑡)) , 𝐼𝑁𝑣2,𝑖 = 𝜔𝑖 𝑣1𝑖 ,

𝑂𝑈𝑇𝑤𝑣,𝑖 = min (𝑤𝑣𝑖 , 𝜈𝑖𝑏 (𝑡)) , 𝐼𝑁𝑤𝑣,𝑖 = 𝜏𝑣 𝑣2𝑖 + 𝜁𝑣 (𝑅𝑣1,𝑖 + 𝑅𝑣2,𝑖 + 𝑅𝑤𝑣,𝑖 ),

𝑂𝑈𝑇𝑏,𝑖 = 𝜏𝑏 𝑏𝑖 , 𝐼𝑁𝑏,𝑖 = 𝑏𝑖 (𝑡),

𝑂𝑈𝑇𝑤𝑏,𝑖 = 0, 𝐼𝑁𝑤𝑏,𝑖 = 𝜏𝑏 𝑏𝑖 + 𝜁𝑣 (𝑅𝑏,𝑖 + 𝑅𝑤𝑏,𝑖 ),

where 𝜇 indicates the reduction rate caused by NPIs. For example, ignoring other factors, if the basic

3
reproductive number is 2 and 𝜇 is 0.7, then the effective reproductive number becomes (1 − 0.7) ×

2 = 0.6 . We estimated the value of 𝜇 every week using least square curve fitting method,

𝛿 𝑜
minimizing the difference of cumulative incidence comes from model (∫ ∑𝑋 ∑𝑖 𝛼(𝐼𝑋,𝑖 + 𝐼𝑋,𝑖 )) 𝑑𝑡) and

data. Model simulation time was from 1 August 2021 to 2 February 2022, because the testing policy

has been changed since 3 February 2022.7 The parameter Γ𝑖 represents the number of overseas

entrants cases from age group i who are not screened but entered the local community. Then using the

population data of the age groups, Γ𝑖 is computed using ratio and proportion and using the formula

40/38 = ∑𝑋 ∑𝑖 Γ𝑖 , where 40/38 is the average daily number of overseas entrants cases (40 in total)

across all ages from 24 November to 31 December 2021 (38 days). The model parameters are listed in

Tablea and Tableb. Note that there are three effectiveness of vaccine, against infection, transmission,

and severity. Daily number of vaccine administration, 𝜈𝑖𝑏 (𝑡) and 𝜈𝑖 (𝑡), are obtained from data.8 In

Korea, booster shots were administered to anyone who finished primary doses more than 90 days ago.

That means, targets for the booster shot administration are waned after primary dose (𝑤𝑣𝑖 ) and 2

weeks after primary dosed (𝑣2𝑖 ). Therefore, to keep nonnegativity condition and prioritize waned

hosts to be boostered, number of flows from 𝑤𝑣𝑖 and 𝑣2𝑖 to being boostered (𝑏𝑖 ) are set as

min (𝑤𝑣𝑖 , 𝜈𝑖𝑏 (𝑡)) and 𝜈𝑖𝑏 (𝑡) − min (𝑤𝑣𝑖 , 𝜈𝑖𝑏 (𝑡)), respectively.

4
Tablea Model parameters, non-age dependent
Symbol Description Value Reference
1,3
1/𝜅 Latent period 2
2,3
1/𝛼 Infectious period 6
𝜎 Progression rate 2 Assumed
9
𝛾𝑚𝛿 Recovery rate of mild Delta case 1/10
9
𝛾𝑐𝛿 Recovery rate of severe Delta case 1/14
𝑜 9
𝛾𝑚 Recovery rate of mild Omicron case 1/7
𝑜 9
𝛾𝑐 Recovery rate of severe Omicron 1/7
case
10,11,12
𝜁𝑢 , 𝜁𝑤 Waning rate of infection-induced 1/120
immunity for unvaccinated case
13
𝜁𝑣1 , 𝜁𝑣2 , 𝜁𝑤𝑣 , 𝜁𝑏 , 𝜁𝑤𝑏 Waning rate of infection-induced 1/480
immunity for vaccinated case
𝜏𝑏 Waning rate of booster shot 1/300 Assumed

5
Tableb Model parameters, age-and-strain dependent. Note that subscript and superscript indicate strain
and age group, respectively
Symbol Description 1 2 3 4 5 6 7 8 Reference
1/𝜔𝑖 Average duration from 35 38 39 43 43 86 66 41 14

first dose to 2 weeks

after second dose
𝜏𝑣,𝑖 Waning rate after 0.002 0.002 0.002 0.002 0.002 0.003 0.003 0.002 8,15,16

finishing primary 0 1 5 3 3 6 0 2
vaccine
𝛿 Severe rate of 0.000 0.003 0.008 0.020 0.432 0.074 0.132 0.258 17
𝑝𝑢,𝑖
unvaccinated 4 0 7 0 3 9 9 7
𝑜 0.000 0.000 0.002 0.005 0.010 0.018 0.033 0.064
𝑝𝑢,𝑖
1 8 2 0 6 7 2 7
0 0 0.003 0.005 0.008 0.033 0.024 0.054 17
𝛿
𝑝𝑣1,𝑖 𝛿
, 𝑝𝑣2,𝑖 , Severe rate, except
8 1 0 9 4 1
𝛿
𝑝𝑤𝑣,𝑖 , 𝑝𝑏,𝑖𝛿
, unvaccinated
𝛿 𝛿
𝑝𝑤𝑏,𝑖 , 𝑝𝑤,𝑖
𝑜 𝑜 0 0 0.000 0.001 0.002 0.008 0.006 0.013
𝑝𝑣1,𝑖 , 𝑝𝑣2,𝑖 ,
𝑜
𝑝𝑤𝑣,𝑖 𝑜
, 𝑝𝑏,𝑖 , 9 3 0 5 1 5
𝑜 𝑜
𝑝𝑤𝑏,𝑖 , 𝑝𝑤,𝑖
𝛿 𝑜 Fatal rate 0 0.000 0.000 0.000 0.002 0.006 0.025 0.127 17
𝑓𝑋,𝑖 , 𝑓𝑋,𝑖
1 3 6 0 4 4 5
𝛿 Natural-infection- 0.396 0.414 0.381 0.393 0.401 0.273 0.322 0.381 8,15,16
𝑒𝑤,𝑖
induced partial 0 7 8 8 5 7 6 0
𝑜
𝑒𝑤,𝑖 0.071 0.076 0.059 0.065 0.069 0.006 0.032 0.063 8,15,16
immunity against
infection 0 5 6 6 1 4 3 0
𝛿 Vaccine effectiveness 0.621 0.631 0.610 0.618 0.622 0.545 0.575 0.612 8,15,16
𝑒𝑣1,𝑖
against infection; 2 6 3 9 3 8 0 7 2
𝑜
𝑒𝑣1,𝑖 0.232 0.243 0.188 0.207 0.218 0.020 0.105 0.205 8,15,16
weeks before finishing
primary doses 0 7 8 9 7 4 3 8
𝛿 Vaccine effectiveness 0.777 0.789 0.763 0.772 0.778 0.681 0.719 0.765 8,15,16
𝑒𝑣2,𝑖
against infection; 2 0 1 6 8 5 2 6 2
𝑜 0.290 0.304 0.236 0.259 0.273 0.025 0.131 0.257 8,15,16
𝑒𝑣2,𝑖 weeks after finishing
primary doses 0 6 0 9 4 5 7 2
𝛿 𝛿 Waned vaccine 0.396 0.414 0.381 0.393 0.401 0.273 0.322 0.381 8,15,16
𝑒𝑤𝑣,𝑖 , 𝑒𝑤𝑏,𝑖
effectiveness against 0 7 8 8 5 7 6 0
𝑜
𝑒𝑤𝑣,𝑖 𝑜
, 𝑒𝑤𝑏,𝑖 0.071 0.076 0.059 0.065 0.069 0.006 0.032 0.063 8,15,16
infection
0 5 6 6 1 4 3 0
𝛿 Vaccine effectiveness 0.899 0.896 0.899 0.898 0.897 0.909 0.905 0.900 8,15,16
𝑒𝑏,𝑖
against infection; 0 2 1 0 2 4 3 3
𝑜
𝑒𝑏,𝑖 0.510 0.501 0.496 0.497 0.497 0.486 0.496 0.507 8,15,16
boostered
0 2 2 5 5 9 3 0
Natural-infection- 0.050 0.054 0.050 0.052 0.053 0.034 0.040 0.048 8,15,16
𝑒̃
𝛿
𝑤,𝑖
induced 0 7 3 0 2 8 8 2
0.006 0.006 0.005 0.005 0.005 0.000 0.002 0.005 8,15,16
𝑒̃
𝑜
𝑤,𝑖 transmissibility
reduction 0 5 1 6 9 5 7 3
̃
𝛿 Vaccine-induced 0.216 0.233 0.225 0.229 0.232 0.190 0.200 0.213 8,15,16
𝑒𝑣1,𝑖
transmissibility 0 8 6 2 5 8 7 0
0.032 0.035 0.027 0.030 0.031 0.002 0.014 0.028 8,15,16
𝑒̃
𝑜
𝑣1,𝑖 reduction; 2 weeks
before finishing 0 0 3 0 7 9 6 4
primary doses
0.216 0.233 0.225 0.229 0.232 0.190 0.200 0.213 8,15,16
𝑒̃
𝛿
𝑣2,𝑖
Vaccine-induced
transmissibility 0 8 6 2 5 8 7 0
0.032 0.035 0.027 0.030 0.031 0.002 0.014 0.028 8,15,16
𝑒̃
𝑜
𝑣2,𝑖 reduction; 2 weeks
after finishing primary 0 0 3 0 7 9 6 4
doses
̃ 0.050 0.054 0.050 0.052 0.053 0.034 0.040 0.048 8,15,16
𝑒𝑤𝑏,𝑖 , 𝑒𝑤𝑣,𝑖 Waned
𝛿 ̃𝛿 vaccine-
0 7 3 0 2 8 8 2

6
 induced 0.006 0.006 0.005 0.005 0.005 0.000 0.002 0.005 8,15,16
𝑒̃
𝑜 ̃𝑜
𝑤𝑏,𝑖 , 𝑒𝑤𝑣,𝑖
transmissibility 0 5 1 6 9 5 7 3
reduction
̃
𝛿 Vaccine-induced 0.410 0.402 0.404 0.403 0.402 0.414 0.412 0.410 8,15,16
𝑒𝑏,𝑖
transmissibility 0 6 4 5 4 2 6 4
0.077 0.074 0.074 0.074 0.074 0.073 0.074 0.076 8,15,16
𝑒̃
𝑜
𝑏,𝑖 reduction; boostered
0 7 0 1 0 4 9 5

7
REFERENCES
1. Korea Disease Control and Prevention Agency; Korean situation report of COVID-19; 2020

Feb 16

2. Ki M; Task Force for 2019-nCoV. Epidemiologic characteristics of early cases with 2019

novel coronavirus (2019-nCoV) disease in Korea. Epidemiol Health 2020;42:e2020007.

3. World Health Organization; Transmission of SARS-CoV-2: implications for infection

prevention precautions: scientific brief, 09 July 2020. World Health Organization; 2020

4. Ko Y, Lee J, Seo Y, Jung E. Risk of COVID-19 transmission in heterogeneous age groups and

effective vaccination strategy in Korea: a mathematical modeling study. Epidemiology and

health. 2021;43.

5. Billah MA, Miah MM, Khan MN. Reproductive number of coronavirus: A systematic review

and meta-analysis based on global level evidence. PloS one. 2020;15(11):e0242128.

6. Campbell F, Archer B, Laurenson-Schafer H, Jinnai Y, Konings F, Batra N, Pavlin B,

Vandemaele K, Van Kerkhove MD, Jombart T, Morgan O. Increased transmissibility and

global spread of SARS-CoV-2 variants of concern as at June 2021. Eurosurveillance.

2021;26(24):2100509.

7. Korea Disease Control and Prevention Agency. Regular briefing. February 3, 2022.

8. Korea Disease Control and Prevention Agency. Vaccination status. Available at

https://ncv.kdca.go.kr/

9. Sould Metropolitan Government. The daily news review of COVID-19. March 4, 2022.

10. Brouqui P, Colson P, Melenotte C, Houhamdi L, Bedotto M, Devaux C, Gautret P, Million M,

Parola P, Stoupan D, La Scola B. COVID‐19 re‐infection. European journal of clinical

investigation. 2021;51(5).

11. Silva MS, Demoliner M, Hansen AW, Gularte JS, Silveira F, Heldt FH, Filippi M, Pereira

VM, Silva FP, Mallmann L, Fink P. Early detection of SARS-CoV-2 P. 1 variant in Southern

Brazil and reinfection of the same patient by P. 2. Revista do Instituto de Medicina Tropical

de São Paulo. 2021;63.

8
12. Siqueira JD, Goes LR, Alves BM, da Silva AC, de Carvalho PS, Cicala C, Arthos J, Viola JP,

Soares MA. Distinguishing SARS-CoV-2 bonafide re-infection from pre-existing minor

variant reactivation. Infection, Genetics and Evolution. 2021 1;90:104772.

13. Yang Y, Yang M, Peng Y, Liang Y, Wei J, Xing L, Guo L, Li X, Li J, Wang J, Li M.

Longitudinal analysis of antibody dynamics in COVID-19 convalescents reveals neutralizing

responses up to 16 months after infection. Nature Microbiology. 2022:1-1.

14. Korea Disease Control and Prevention Agency, What vaccines will be administered?.

Available at https://ncv.kdca.go.kr/menu.es?mid=a10119000000

15. Gardner BJ, Kilpatrick AM. Estimates of reduced vaccine effectiveness against

hospitalization, infection, transmission and symptomatic disease of a new SARS-CoV-2

variant, Omicron (B. 1.1. 529), using neutralizing antibody titers. MedRxiv. 2021.

16. Andrews N, Stowe J, Kirsebom F, Toffa S, Rickeard T, Gallagher E, Gower C, Kall M,

Groves N, O’Connell AM, Simons D. Covid-19 vaccine effectiveness against the omicron (B.

1.1. 529) variant. New England Journal of Medicine. 2022.

17. Korea Disease Control and Prevention Agency, Public health weekly report, 2021;14(37).

Available at

https://www.kdca.go.kr/board/board.es?mid=a30501000000&bid=0031&list_no=716914&ac

t=view#.