The Susceptible-Infectious-Recovered (SIR) model, introduced by Kermack and McKendrick in 1927, provides a fundamental framework for studying infectious disease dynamics [10] . It categorizes the population into three compartments:

The model is described by the following set of differential equations:

$\begin{array}{l}\frac{\text{d}S}{\text{d}t}=-\beta IS/N\\ \frac{\text{d}I}{\text{d}t}=\frac{\beta IS}{N}-\lambda I\\ \frac{\text{d}R}{\text{d}t}=\lambda I\end{array}$ (1)

where $\beta$ represents the transmission rate and $\lambda$ is the recovery rate (see Figure 1 ). The SIR model assumes a closed population with homogeneous mixing, lifetime immunity, and no demographic changes. However, real-world epidemics often involve mortality, prompting modifications such as the SIRD model. Equation (1) is subject to the initial conditions, $S\left(t_0\right)>0$, $I\left(t_0\right)\ge 0$, and $R\left(t_0\right)\ge 0$ at the initial time $t_0$.

The model is based on a large, closed population, a short-lived outbreak, immediate infectiousness, lifetime immunity, and mass-action mixing. It also assumes that all people who recover from the disease are either fully immune or no longer able to help it spread. However, real-world epidemics often involve significant mortality, which the SIR model does not explicitly address [11] . The model is expanded into the susceptible-infected-recovered-deceased (SIRD) model.

The model adds a fourth compartment to SIR model called “deceased” (D) to account for people who died from the disease. This makes the model more useful and accurate. The deceased (D) component represents individuals who have died as a direct result of the infection. This compartment differentiates between those who recover and those who do not survive the disease.

A key assumption in the SIRD model is that the total population, represented by the sum of $S\left(t\right),I\left(t\right),R\left(t\right)$, and $D\left(t\right)$, remains constant over time. Additionally, the model assumes immediate infection upon exposure, with no latent period between exposure and the onset of infection. Factors such as quarantine measures or confinement are not incorporated into this model [12] .

The SIRD model is a widely utilized framework for characterizing epidemiology and classifying diseases within a community [13] . The model extends the SIR model by incorporating a variable for the deceased population as a function of time. This model has been widely used in epidemiological studies to analyze the spread and control of infectious diseases, including COVID-19, AIDS, and Ebola. Researchers have applied it to understand transmission dynamics, estimate key epidemiological parameters, and evaluate the effectiveness of intervention strategies such as vaccination, social distancing, and quarantine measures [13] - [17] .

The model is governed by a set of differential equations that describe the rate of change in each compartment over time:

$I+S\stackrel{\beta }{\to }I$

$\frac{\text{d}S\left(t\right)}{\text{d}t}=-\beta S\left(t\right)I\left(t\right)/N$

$I\stackrel{\lambda }{\to }R$

$\frac{\text{d}I\left(t\right)}{\text{d}t}=\beta S\left(t\right)I\left(t\right)/N-\lambda I\left(t\right)-k_{d}I\left(t\right)$ (2)

$I\stackrel{k_d}{\to }D$

$\frac{\text{d}R\left(t\right)}{\text{d}t}=\lambda I\left(t\right)$

$\frac{\text{d}D\left(t\right)}{\text{d}t}=k_{d}I\left(t\right)$

where: $\beta$ is a transmission rate, representing the likelihood of disease spread through contact between Susceptible and Infected individuals. $\lambda$ is a recovery rate, representing the fraction of infected individuals who recover per unit time, and $k_d$ is a mortality rate, representing the fraction of infected individuals who die per unit time.

While the SIRD model is an improvement over the SIR model, it still has limitations. These limitations are: 1) No Latent Period: Diseases with significant incubation periods, such as COVID-19, may require the inclusion of an exposed compartment, 2) Simplistic Assumptions: The SIRD model assumes homogeneity in the population, ignoring variations in age, behavior, or geography. The transition from the SIR to the SIRD model marks an important step in improving the realism of epidemic modeling by incorporating disease-related mortality. This modification provides a clearer picture of the epidemic’s toll, helping researchers and policymakers design better interventions to manage and mitigate outbreaks. However, further extensions and refinements are needed to adapt the model to specific diseases and real-world complexities.

Sen and Sen (2021), introduced a significant enhancement to the standard SIRD model to better analyze and predict the spread of COVID-19 by accounting for real-world complexities that the original model does not address [14] . Their modified model incorporates the effects of exposure, quarantine, confinement, and asymptomatic populations [15] [18] [19] , reflecting the unique dynamics of COVID-19 transmission and control measures. These additions make the model more aligned with the realities of managing a global pandemic. The governing equations of the modified model are as follows (See Table 1 for description of parameter):

$\frac{\text{d}S\left(t\right)}{\text{d}t}=-\alpha S\left(t\right)-\frac{\beta S\left(t\right)I\left(t\right)}{N}-\frac{\sigma S\left(t\right)A\left(t\right)}{N}-\eta S\left(t\right)$ (3)

$\frac{\text{d}A\left(t\right)}{\text{d}t}=-\tau A\left(t\right)+\xi E\left(t\right)$ (4)

$\frac{\text{d}C\left(t\right)}{\text{d}t}=\alpha S\left(t\right)-\mu C\left(t\right)$ (5)

$\frac{\text{d}E\left(t\right)}{\text{d}t}=-\gamma E\left(t\right)+\beta S\left(t\right)I\left(t\right)N+\mu C\left(t\right)+\eta S\left(t\right)+\sigma S\left(t\right)A\left(t\right)N-\xi E\left(t\right)$ (6)

$\frac{\text{d}I\left(t\right)}{\text{d}t}=\tau A\left(t\right)+\gamma E\left(t\right)-\delta I\left(t\right)$ (7)

$\frac{\text{d}R\left(t\right)}{\text{d}t}=\lambda \left(t\right)\varrho \left(t\right)$ (8)

$\frac{\text{d}\varrho \left(t\right)}{\text{d}t}=\delta I\left(t\right)-\lambda \left(t\right)\varrho \left(t\right)-k_d\left(t\right)\varrho \left(t\right)$ (9)

$\frac{\text{d}D\left(t\right)}{\text{d}t}=k_d\left(t\right)\varrho \left(t\right)$ (10)

$S\left(t\right)+C\left(t\right)+E\left(t\right)+A\left(t\right)+I\left(t\right)+Q\left(t\right)+R\left(t\right)+D\left(t\right)=N$ (11)

In this model, N represents the total population, while different compartments describe the progression of individuals through various stages of infection and containment.

The interaction between A(t) (asymptomatic carriers) and Q(t) (quarantined individuals) is critical in shaping the disease’s trajectory. Since asymptomatic individuals unknowingly spread the virus, early detection, contact tracing, and isolation strategies are essential in breaking transmission chains [20] . Efficient public health interventions—such as rapid testing, timely quarantine, and community awareness campaigns—play a crucial role in reducing the overall infection rate and preventing healthcare system overload. By swiftly identifying and isolating infected individuals, health authorities can protect high-risk populations, minimize socio-economic disruptions, and effectively manage disease outbreaks.

Several methods exist for estimating the basic reproduction number ( $R_0$), each suited to different types of data and levels of disease model complexity. The choice of method depends on factors such as available epidemiological data, assumptions about disease transmission, and the structure of the underlying mathematical model. Below are some of the most used approaches:

The next-generation matrix method is widely used in deterministic compartmental models to estimate $R_0$. It calculates the expected number of secondary infections by examining the dominant eigenvalue of the next-generation matrix, which describes disease transmission within a structured population [23] . Mathematically, $R_0$ is derived from the spectral radius of the next-generation matrix $G$, given by:

$R_0=\rho \left(G\right)$ (12)

Here, $\rho \left(G\right)$ represents the dominant eigenvalue (i.e., the largest absolute eigenvalue) of the next-generation matrix $G$, which is constructed from the new infection terms and transition terms of the model. The next-generation matrix method decomposes the infection process into two key components: (1) the New Infections Matrix $F$, which represents the rate at which new infections arise in each compartment, and (2) the Transition Matrix $V$, which describes the movement of infected individuals between compartments, including recovery and death. Mathematically:

$G=F{V}^{-1}$ (13)

where: $F$ is a non-negative matrix describing new infections and $V$ is an invertible matrix describing the movement of infected individuals. Consider an SEIR (Susceptible-Exposed-Infected-Recovered) model with:

$\frac{\text{d}t}{\text{d}E}=\beta SI-\sigma E$ (14)

$\frac{\text{d}I}{\text{d}t}=\sigma E-\gamma I$ (15)

where: $\beta$ represents the transmission rate, $\sigma$ represents the rate of progression from exposed to infected, and $\gamma$ represents the recovery rate. From eq. 14 – 15, the matrices F and V are:

$F=\left[\begin{array}{cc}0& \beta \\ 0& 0\end{array}\right]$ (16)

$V=\left[\begin{array}{cc}\sigma & 0\\ -\sigma & \gamma \end{array}\right]$ (17)

Computing $G=F{V}^{-1}$:

$V^{-1}=\left[\begin{array}{cc}\frac{1}{\sigma }& 0\\ \sigma /\left(\gamma \sigma \right)& \frac{1}{\gamma }\end{array}\right]$ (18)

$G=F{V}^{-1}=\left[\begin{array}{cc}0& \frac{\beta }{\gamma }\\ 0& 0\end{array}\right]$ (19)

The eigenvalue $G$ are $\lambda =0$ and $\lambda =\frac{\beta }{\gamma }$, that is: $R_0=\frac{\beta }{\gamma }$.

For diseases with an explicit generation time distribution, the Euler-Lotka equation provides an alternative way to estimate $R_0$ using epidemiological data. To derive Euler-Lotka equation, assume the epidemic is growing exponentially with rate $r$. Then the incidence (number of new cases) at time $t$ is given by:

$n\left(t\right)=n\left(0\right){\text{e}}^{rt}$ (20)

where $n\left(0\right)$ represents the incidence at time $t=0$.

Following a renewal equation for incidence, the number of new infections at time $t$ is the sum of contributions from all previous infections. Each infected individual generates new cases according to the generation time distribution $b\left(\tau \right)$, where $\tau$ is the time since infection. Thus, we can write [24] :

$n\left(t\right)={\displaystyle {\int }_0^{\infty }b}\left(\tau \right)n\left(t-\tau \right)\text{d}\tau$ (21)

Replace $n\left(t-\tau \right)$ by the exponential form: $n\left(t-\tau \right)=n\left(0\right){\text{e}}^{r\left(t-\tau \right)}$, eq. (21) can be rewritten as:

$n\left(t\right)=n\left(0\right){\text{e}}^{rt}{\displaystyle {\int }_0^{\infty }b}\left(\tau \right){\text{e}}^{r\tau }\text{d}\tau$ (22)

Consider eq. (20), we can write:

$n\left(0\right){\text{e}}^{rt}=n\left(0\right){\text{e}}^{rt}{\displaystyle {\int }_0^{\infty }b}\left(\tau \right){\text{e}}^{r\tau }\text{d}\tau$ (23)

Hence, the Euler–Lotka equation follows.

$1=n\left(0\right){\text{e}}^{rt}{\displaystyle {\int }_0^{\infty }b}\left(\tau \right){\text{e}}^{r\tau }\text{d}\tau$ (24)

where: $b\left(\tau \right)$ represents the generation time distribution (or infectivity profile), describing the probability that an infected individual generates a new case at time $\tau$ infection, and ${\text{e}}^{r\tau }$: Acts as a discount factor, accounting for the fact that secondary cases occurring later contribute less to the current growth of the epidemic. This derivation shows how the Euler–Lotka equation links the epidemic growth rate $r$ with the generation time distribution $b\left(t\right)$ and provides a framework for estimating the basic reproduction number R0 when the total reproduction can be related as:

$R_0={\displaystyle {\int }_0^{\infty }b}\left(\tau \right)\text{d}\tau$ (25)

Thus, by knowing $r$ and $b\left(t\right)$, one can assess the transmission dynamics of an infectious disease.

MLE is a statistical approach for estimating $R_0$ using case incidence data. It fits a likelihood function to observed epidemic data and derives $R_0$ by optimizing parameters [25] . The likelihood function can be expressed as:

$L\left(R_0\right)={\displaystyle {\prod }_{t=1}^{T}P}\left(I_t|R_0,I_{t-1},\cdots \right)$ (26)

where $I_t$ represents the number of new infections at time $t$. This method is particularly useful for real-time epidemic assessments.

Stochastic models incorporate random variation in disease transmission, making them suitable for small population settings or emerging outbreaks where deterministic models may not fully capture transmission dynamics [3] . The branching process approximation is a common approach, defining $R_0$ as:

$R_0={\displaystyle {\sum }_{k=0}^{\infty }kP\left(k\right)}$ (27)

where $P\left(k\right)$ represents the probability of an infected individual generating $k$ secondary cases. Stochastic models can also use Markov chains or agent-based simulations to model individual-level transmission.

The probability $P\left(k\right)$ represents the likelihood that an infected individual transmits the infection to exactly $k$ secondary cases. This probability distribution is typically derived from empirical outbreak data or assumed based on known epidemiological characteristics of the disease, such as transmission modes, contact structure, and variability in infectiousness. Common choices for $P\left(k\right)$ include the Poisson, negative binomial, or geometric distributions, depending on whether transmission exhibits homogeneity or overdispersion. Let us consider the choice of Poisson distribution, if each infected individual has independent and identical probabilities of transmitting the infection to others, and contacts occur randomly in a large, well-mixed population, the number of secondary infections follows a Poisson distribution:

$P\left(k\right)=\frac{{\text{e}}^{-\lambda }{\lambda }^k}{k!}$ (28)

where: $\lambda =R_0$ represent the average number of secondary infections per case. This assumption is valid when transmission is relatively uniform across individuals.

In real-world outbreaks, $P\left(k\right)$ plays a crucial role in determining the likelihood of disease extinction or persistence. If most infected individuals generate few or no secondary infections (high probability for small k), the outbreak is likely to die out early. Conversely, if $P\left(k\right)$ has a heavy tail (e.g., due to superspreading events), even a small fraction of highly infectious individuals can sustain an epidemic. This branching process framework is particularly useful in modeling early outbreak dynamics, where stochastic effects dominate and deterministic models may overestimate transmission potential [26] .

Time-series models, such as the Wallinga-Teunis method, estimate R₀ from epidemic curves by analyzing serial intervals and transmission patterns over time [27] . These approaches are particularly useful when estimating $R_0$ dynamically during outbreaks. Each method for estimating $R_0$ has distinct strengths and limitations, influenced by the quality and type of available data. Deterministic approaches, such as the next-generation matrix method, are well-suited for structured population models, while the Euler-Lotka equation is particularly effective in estimating $R_0$ during early outbreak stages. Stochastic and likelihood-based methods, including MLE, provide robust estimates using incidence data, whereas time-series methods enable dynamic, real-time monitoring of disease transmission. By integrating multiple estimation techniques, researchers can enhance the accuracy and reliability of $R_0$ estimates, ultimately improving public health responses and intervention strategies [28] .

The interpretation of $R_0$ follows a clear epidemiological framework:

By understanding $R_0$, public health authorities can implement timely interventions such as vaccination programs, quarantine measures, and hygiene campaigns to control disease spread effectively.

The basic reproduction number ( $R_0$) varies significantly across infectious diseases, reflecting differences in their transmissibility and outbreak potential. For COVID-19, early pandemic variants the $R_0$ range between 2 - 3 [29], and Omicron variant with $R_0$ of around 8.2 [30]. This variation demonstrates how viral mutations can dramatically change transmission potential.

Seasonal influenza, for instance, has an estimated $R_0$ ranging from 0.9 to 2.1, indicating moderate transmission rates. The 1918 flu pandemic had a slightly higher $R_0$ of 1.4 to 2.8, contributing to its widespread impact. In contrast, measles, one of the most highly contagious diseases, has an $R_0$ between 12 and 18, meaning a single infected person can spread the virus to a large number of susceptible individuals [31] . Monkeypox (MPXV) presents a more variable transmission pattern. Earlier studies estimated an $R_0$ of 0.83, suggesting self-limiting outbreaks [32] [33] . However, recent modeling in non-endemic regions found $R_0$ values between 1.10 and 2.40, raising concerns about its potential for sustained transmission [34] . The 2014 Ebola virus outbreak in West Africa is the largest outbreak of the genus Ebolavirus to date. The maximum likelihood estimates of the basic reproduction number are 1.51 for Guinea, 2.53 for Sierra Leone and 1.59 for Liberia [35] .

These variations highlight the necessity for disease-specific control measures, emphasizing the importance of tailored public health strategies to mitigate transmission risks effectively. Estimating $R_0$ comes with inherent uncertainties, particularly during the early stages of an outbreak. Most modeling simulations produce a range of $R_0$ values, reflecting difficulties in accurately determining key parameters. One major challenge is identifying the true number of cases, as many mild or asymptomatic infections may go undetected, contributing to unobserved transmission [36] . These uncertainties underscore the importance of continuous epidemiological surveillance and refinement of disease models to improve public health response strategies.

Several factors influence $R_0$, extending beyond the intrinsic properties of a pathogen to include environmental and social determinants. Pathogen characteristics such as virulence, which affects disease severity and infectious duration, play a crucial role—higher virulence often leads to increased $R_0$. The mode of transmission also impacts $R_0$, with airborne diseases like measles exhibiting higher values compared to contact-based infections. Additionally, the incubation period influences transmission dynamics, as shorter incubation times accelerate infection cycles [3] . Antigenic variation, observed in viruses like influenza, allows pathogens to evade host immunity, prolonging outbreaks and sustaining transmission [37] . Beyond pathogen biology, environmental and social factors significantly affect $R_0$. Population density facilitates frequent interactions, increasing disease spread, whereas behavioral practices such as improved hygiene, sanitation, and social distancing can effectively reduce contact rates and lower $R_0$ [38] . The efficiency of the healthcare system further modulates $R_0$, as rapid diagnosis, effective treatment, and robust containment strategies help suppress transmission and mitigate outbreak severity [23] . Together, these factors shape disease dynamics, highlighting the complexity of controlling infectious disease spread and the necessity of integrated public health strategies.

$R_0$ serves as a critical metric in public health decision-making, helping to predict whether an infectious disease will spread or decline. It provides essential insights into the severity of an outbreak, the level of intervention required (such as vaccination coverage or quarantine duration), and the effectiveness of control measures. Reducing $R_0$ below 1 is fundamental to epidemic prevention, achievable through widespread immunization, improved healthcare access, and public awareness campaigns [39] . However, while $R_0$ is a valuable tool for assessing disease spread, its interpretation must account for pathogen properties, population behavior, and healthcare system efficiency, as these factors significantly influence transmission dynamics. Integrating epidemiological models with real-world data allows public health officials to refine outbreak response strategies, ensuring timely and effective interventions [40] .

The basic reproduction number, $R_0$, is a fundamental metric in epidemiology that determines the potential spread of an infectious disease in a fully susceptible population. It is a critical parameter for informing public health interventions such as vaccination strategies. The concept is widely used in the literature concerning infectious disease models, exhibiting varying degrees of affection [41] . When applied to control measures, it establishes a threshold for the proportion of the population that must be immunized to halt the spread of the disease, a concept known as herd immunity.

Herd immunity occurs when a sufficient proportion of the population becomes immune—either through vaccination or natural infection—thereby reducing disease transmission. This protects both immunized individuals and those who cannot be vaccinated, such as people with contraindications or compromised immune systems. Public health measures, such as vaccination, social distancing, and improved hygiene, are essential for reducing $R_0$. Vaccination directly lowers $R_0$ by decreasing the proportion of susceptible individuals. The critical vaccination threshold required to achieve herd immunity is calculated using:

Herd Immunity Threshold $\left(HIT\right)=1-1/R_0$ (29)

For example, diseases with higher $R_0$ values require a larger proportion of the population to be immunized. Measles, with an $R_0$ of approximately 12 - 18, necessitates immunizing 92% - 95% of the population, whereas seasonal influenza, with an $R_0$ of about 1.3, requires only around 23% immunity to prevent outbreaks. This relationship follows from the idea that if ( $R_0-1$) out of the $R_0$ individuals an infected person might have transmitted the disease to are vaccinated, each infected individual will generate fewer than one secondary infection. Thus, in general, HIT can be expressed as:

$HIT=\left(R_0-1\right)/R_0$, as stated in eq. (29)

Vaccination strategies aim to lower $R_0$ by reducing susceptibility through immunization or by minimizing contact rates via social distancing and other public health interventions. For highly contagious diseases, targeted vaccination of high-transmission groups (e.g., healthcare workers and essential personnel) can be especially effective [42] .

When designing vaccination programs, the effectiveness of the vaccine is a pivotal factor in determining the coverage needed to achieve herd immunity and effectively control the spread of infectious diseases. Vaccine efficacy refers to the percentage reduction in disease incidence among vaccinated individuals compared to those who are unvaccinated, typically measured under controlled conditions such as clinical trials. The success of a vaccination program hinges not only on achieving the herd immunity threshold (HIT) but also on accounting for the effectiveness of the vaccine. This adjustment ensures that enough individuals in the population are protected to reduce transmission. The relationship between vaccine efficacy ( $Ev$) and the actual proportion of the population that must be vaccinated ( $Vc$) can be expressed as:

$Vc=HIT/Ev$ (30)

If HIT is 80% (0.8) and the vaccine efficacy is 90% (0.9), then:

$Vc=0.8/0.9\approx 89\%$

This means nearly 89% of the population must be vaccinated to achieve herd immunity. If vaccine efficacy is lower—say 70%—the required coverage rises to approximately 114%, indicating that additional interventions or booster doses would be necessary.

Vaccine efficacy plays a crucial role in shaping vaccination programs and determining the required coverage to achieve herd immunity. When vaccines have high efficacy, a smaller proportion of the population needs to be immunized, making it easier to control disease spread. Conversely, for vaccines with moderate efficacy, a larger portion of the population must be vaccinated, which can present logistical and resource challenges. Additionally, the efficacy observed in clinical trials may not always translate directly to real-world effectiveness due to factors such as improper storage, delays in dosing, or individual variations in immune response. To account for these discrepancies, vaccination programs must incorporate a buffer in coverage calculations to maintain adequate protection. Furthermore, population heterogeneity—including differences in susceptibility, contact patterns, and vaccine access—necessitates tailored strategies to ensure that adjusted coverage goals are met across diverse demographics and regions.

Herd immunity is often perceived as a definitive threshold beyond which disease transmission ceases, but this oversimplified view does not align with real-world epidemiology [3] . The commonly used formula for the herd immunity threshold, $1-1/R_0$, assumes homogeneous mixing, stable population dynamics, and lifelong immunity. For instance, if $R_0$ is 3, the threshold is estimated at 66%. However, real-world populations exhibit heterogeneous mixing, where individuals with high contact rates acquiring immunity may reduce transmission, but susceptible subgroups can sustain localized outbreaks. Additionally, population turnover through births and migration continuously replenishes the susceptible pool, explaining why diseases like measles persist despite strong lifelong immunity and high vaccination coverage [43] . Moreover, waning immunity and viral evolution further complicate herd immunity dynamics. Many respiratory viruses, including influenza, RSV, and coronaviruses, allow reinfections due to immune waning or antigenic drift. The COVID-19 pandemic demonstrated that vaccines, while effective in reducing severe illness, provide only temporary protection and do not entirely block transmission [44] . The emergence of immune-evasive variants, such as Omicron, has raised the effective reproduction number ( $R_t$), increased herd immunity thresholds, and contributed to breakthrough infections [45] . Furthermore, vaccine hesitancy remains a critical challenge, as misinformation, distrust in healthcare systems, and logistical barriers hinder widespread uptake, preventing communities from reaching theoretical immunity levels [46] . These complexities highlight that herd immunity is not a fixed endpoint but a dynamic process influenced by epidemiological, immunological, and behavioral factors. As a result, infectious diseases rarely disappear entirely but often establish endemicity, necessitating ongoing public health interventions to manage their impact.

While $R_0$ is a powerful metric in epidemiology, it has several limitations that must be considered when interpreting disease dynamics and public health strategies. The following details illustrate how these limitations impact $R_0$’s estimate:

Traditionally, $R_0$ is calculated under the assumption of an infinitely large, homogeneously mixed population. However, real-world populations exhibit heterogeneity in various forms—demographic structure, spatial distribution, and individual contact patterns—all of which can significantly impact $R_0$ estimates and disease dynamics [24] .

Demographic factors such as age, immunity levels, and susceptibility create variations in $R_0$. Different age groups have varying levels of susceptibility and contact rates. For instance, children often have higher contact rates, increasing transmission potential, while the elderly may have weaker immune responses, prolonging infectious periods [47] . Age-structured contact matrices reveal that certain groups interact more frequently, influencing transmission dynamics and leading to heterogeneous $R_0$ values [48] . Additionally, birth and death rates affect $R_0$; high birth rates continuously introduce new susceptible individuals, maintaining disease transmission and increasing $R_0$ over time [3] .

Geographical factors also shape disease spread and $R_0$ estimates. Densely populated urban areas facilitate higher contact rates, increasing $R_0$, whereas rural areas experience slower transmission due to lower population density [49] . Movement between regions, such as commuting and migration, can introduce infections to new populations, modifying local $R_0$values and influencing regional outbreaks [50] . Additionally, natural barriers (e.g., mountains, rivers) and artificial interventions (e.g., lockdowns, travel restrictions) alter contact networks and may lower $R_0$ by limiting transmission pathways [51] .

Variability in individual interactions plays a crucial role in shaping $R_0$. Some individuals, known as super-spreaders, may have disproportionately high contact rates, leading to localized outbreaks with much higher $R_0$ values [52] . Social and professional networks create household and workplace transmission clusters, causing local variations in $R_0$ and influencing epidemic wave patterns [53] . Furthermore, public health measures such as mask-wearing, social distancing, and vaccination uptake impact transmission rates and the effective $R_0$ [54] .

The effect of heterogeneity on $R_0$ differs between infinite and finite populations. In infinite populations, heterogeneity generally increases $R_0$ because high-contact individuals spread the disease more effectively, making eradication more difficult [55] . In finite populations, the effects become more complex. When $R_0$ is small relative to the population size, heterogeneity increases $R_0$ by enhancing transmission efficiency. However, when $R_0$ is large relative to the population size, heterogeneity can reduce $R_0$, making disease control easier than predicted by classical models [56] .

Heterogeneity in population structure and behavior influences epidemic dynamics and intervention strategies. Simplistic models that assume homogeneous mixing may misestimate $R_0$, leading to inaccurate predictions of disease spread [57] . Population heterogeneity can also cause staggered outbreaks as different subpopulations experience transmission at different times [58] . Additionally, vaccination strategies prioritizing high-contact individuals can lower $R_0$ more effectively than uniform approaches [59] .

Population heterogeneity plays a crucial role in determining $R_0$ and shaping disease dynamics. Epidemiological models must account for variations in demographic structure, spatial distribution, and individual contact patterns to improve predictive accuracy and optimize intervention strategies. Network-based, spatially explicit, and age-structured models provide more realistic estimates of $R_0$, aiding in more effective disease control and prevention measures.

Many $R_0$ calculations assume that individuals in a population mix homogeneously, meaning that everyone has an equal chance of coming into contact with an infected individual. Populations have structured interactions influenced by factors such as age, geography, and social behavior. For example, diseases spread differently in urban areas versus rural communities due to variations in population density and movement patterns [60] .

$R_0$ is a theoretical measure based on an entirely susceptible population, meaning it does not account for dynamic changes over time. Factors such as acquired immunity, vaccination campaigns, behavioral adaptations (e.g., social distancing), and government interventions all influence disease transmission. The effective reproduction number ( $R_t$) is often used instead to measure real-time disease spread, as it evolves based on current population immunity and control measures [43] .

The estimated value of $R_0$ depends on the underlying epidemiological model used, including assumptions about transmission pathways, incubation periods, and intervention effectiveness. This variability makes it difficult to compare $R_0$ values across studies and diseases. For instance, different studies on COVID-19 reported varying $R_0$ estimates due to differences in data sources, population settings, and modeling techniques [60] .

Examples: The following examples illustrate how $R_0$ and the corresponding herd immunity thresholds vary across diseases, and they highlight the importance of tailored public health strategies that consider real-world complexities such as vaccine efficacy, population behavior, and viral evolution.

1) Measles: Measles is one of the most contagious diseases, with an $R_0$ estimated between 12 and 18. This high $R_0$ means that, in a completely susceptible population, one individual could infect 12 to 18 others. To achieve herd immunity against measles, approximately 92% - 95% of the population must be immunized. This example underscores why even small declines in vaccination rates can lead to significant outbreaks [61] . 2) Influenza: Seasonal influenza typically has a lower $R_0$, around 1.3. The herd immunity threshold in theory would be about 23% of the population needing immunity. However, due to antigenic drift (changes in the virus) and waning immunity, annual vaccination is necessary to manage and control influenza outbreaks effectively. 3) COVID-19: At the start of the COVID-19 pandemic, early estimates of SARS-CoV-2’s $R_0$ were around 2 to 3, indicating moderate transmissibility. With the emergence of more contagious variants, such as Omicron—where $R_0$ was estimated to be as high as 8—the challenge of controlling the spread increased dramatically. Despite high vaccination rates, factors like waning immunity and variant evolution necessitated booster programs and reinstated non-pharmaceutical interventions (e.g., mask mandates, social distancing) to keep the effective reproduction number ( $R_t$) below 1 [1] .

These limitations do not diminish the utility of $R_0$; rather, they define how it should be applied in public health interventions. Instead of viewing $R_0$ as a fixed threshold for disease control, it should be used alongside other metrics, such as the effective reproduction number ( $R_t$), to guide dynamic response strategies. For instance, while $R_0$ helps estimate the herd immunity threshold for vaccination planning, real-world factors like vaccine efficacy, population heterogeneity, and waning immunity must also be considered. Similarly, outbreak containment relies on continuous monitoring of $R_t$ rather than a static $R_0$ value. By integrating multiple data sources and adapting strategies based on real-time surveillance, public health officials can make informed decisions that effectively mitigate disease spread.

While $R_0$ provides insight into a pathogen’s theoretical transmission potential in a fully susceptible population, real-world disease spread is far more complex. This is where another crucial metric, the effective reproduction number ( $R_t$), becomes essential. Unlike $R_0$, which remains constant under idealized conditions, $R_t$ fluctuates over time as immunity builds, public health measures are implemented, and behavioral adaptations occur, making it a vital metric for real-time epidemic monitoring.

The growth and decline of an epidemic depend on changes in disease transmission over time. At the beginning of an outbreak, when the entire population is susceptible, transmission rates are typically high. As the epidemic progresses, transmission decreases due to behavioral changes, acquired immunity through infection, or immunization efforts. The epidemic reaches its peak when transmission drops below a critical threshold, meaning each infected individual no longer spreads the disease to more than one person on average. The effective reproduction number ( $R_t$) quantifies this dynamic by measuring the average number of new infections generated by each case at a given time within a population, considering the impact of immunity, interventions, and behavioral changes [61] . It measures the actual transmission of a disease in a population where some people may be immune and control measures are in place. $R_t$ is derived from $R_0$ but adjusts for factors such as vaccination, acquired immunity, social distancing, and public health interventions. It provides a real-time assessment of disease transmission. When $R_t>1$, an outbreak continues to grow; when $R_t<1$, the outbreak declines and eventually dies out.

Several factors influence the effective reproduction number ( $R_t$), determining how a disease spreads within a population over time. Immunity and vaccination play a significant role, as increasing immunity—either through natural infection or vaccination—reduces the proportion of susceptible individuals, thereby lowering $R_t$ [61] . However, if immunity wanes over time or new variants emerge that evade prior immunity, $R_t$ may rise again, leading to renewed outbreaks. Public health measures, such as mask mandates, lockdowns, and social distancing, directly impact $R_t$ by limiting transmission opportunities and reducing contact between infected and susceptible individuals. Additionally, behavioral changes at the individual level, such as avoiding crowded places, practicing good hygiene, and adhering to health recommendations, can further decrease $R_t$, slowing disease spread even in the absence of formal interventions [29] . Understanding these factors is crucial for implementing effective, adaptive public health strategies.

At the onset of the COVID-19 pandemic, $R_0$ for SARS-CoV-2 was estimated to be between 2 and 3, indicating that each infected individual could transmit the virus to two or three others in a completely susceptible population. However, as governments implemented various public health measures—including lockdowns, mask mandates, and vaccination programs—the effective reproduction number ( $R_t$) fluctuated. During periods of strict restrictions, $R_t$ dropped below 1, leading to a decline in case numbers. Conversely, when these measures were relaxed, $R_t$ increased again, triggering new waves of infections.

Compared to $R_0$, $R_t$ provides a more practical and real-time assessment of disease transmission under prevailing conditions. While $R_0$ is useful for estimating the initial risk and informing control strategies, tracking $R_t$ allows for ongoing evaluation of an outbreak’s trajectory. By continuously monitoring $R_t$, public health officials can determine whether infections are increasing, decreasing, or stabilizing, enabling them to adapt interventions and optimize response strategies accordingly.