Final project for Pf. Hyman's epidemic modeling course Fall 2020
When the COVID-19 pandemic began one of the major concerns of public health officials was how to avoid overwhelming the healthcare resources of a region given the newly inflated healthcare needs of the population. An overwhelmed healthcare system can have devastating secondary consequences for a population where everyday, treatable conditions are no longer treatable due to a lack of hospital space, healthcare workers, or healthcare supply resources which leads to unnecessary loss and suffering.
One particularly important concern is that the overlap of the COVID-19 pandemic with an unusually bad influenza season would result in hospitals becoming full. For this reason, public health officials have instructed the American public that it is imperative for as many people as possible to get their flu vaccine while the COVID-19 pandemic is ongoing. For seasonal influenza during normal circumstances, vaccination is the most cost effective mitigation strategy. However, COVID-19 mitigation strategies such as social distancing, mask wearing, and stay-at-home orders do not only affect the spread of COVID-19, but also affect the spread of any disease that is transmitted by person-to-person contact, such as seasonal influenza. In this project we analyze the effect of COVID-19 mitigation strategies on seasonal influenza in order to understand the risks associated with a lack of compliance in the flu vaccine directives and the cost-benefit trade off of a flu vaccination campaign during the pandemic.
In the southern hemisphere, which usually experiences it’s peak influenza season July through October, the 2020 influenza season was not as severe as in past years; even in countries with weak COVID-19 mitigation policies, the 2020 influenza season was lighter than normal. In Australia, where there are high rates of social distancing and mask compliance, the 2020 influenza season was exceptionally light, with only 315 reported cases of the flu, down 99.8% from average levels. This suggests that compliance with COVID-19 specific mitigation strategies is sufficient to prevent an influenza epidemic.
Public health officials use epidemiological models to understand how diseases spread through a population. Compartmental models are deterministic models of disease transmission that segment the population into different groups based on infectivity and susceptibility parameters and allow researchers to identify the effects of human behaviors and mitigation strategies on the spread of a disease. These models can also be used to analyze the short- and long-term outcomes of an epidemic.
In order to model the spread of the influenza viruses we develop a compartmental based SVEIRS (Susceptible, Vaccinated, Exposed, Infected, Recovered, Susceptible) model with four stage progressions for the infected population. In our model, we introduce a vaccinated population which will be immune from influenza infection for a period of time, eventually returning to the susceptible population. This waning of vaccine immunity can be interpreted as the change in circulating influenza strains each year which are assumed not to be covered by the previous year’s vaccine. We also use a latent state incubation period and stage progression for the infected states which allows us to vary the contacts and infectivity over the course of illness. This flexibility allows us to incorporate behavior change and understand more carefully how mask wearing, particularly in the later stages of the disease, when most people will have returned to work or school, will affect the spread of influenza causing viruses. Finally, we allow recovered individuals to become susceptible again which is also a result of the change in viral strains from year to year.
In our model, we assume that once an individual is exposed to the virus they will go through an incubation period before becoming infectious. We also assume vaccination offers 100% immunity from the influenza virus, but only for a period of time after which a vaccinated individual will return to the susceptible population. When an individual progresses to the first infectious state they are not symptomatic but are now slightly infectious. The second stage of infection will be the peak of the illness when malaise has forced a behavior change to reduce the number of contacts and the person is at their peak infectivity. In the third stage of illness, malaise has begun to recede and the individual will start to increase their contacts as their infectivity decreases. In the last stage of infection, the individual may have little to no symptoms, will have returned to normal contact levels, and will still be slightly infectious. We also have a parameter in our model that accounts for a reduction in infectivity if a mask is worn by an infectious individual.
We incorporate mask wearing into our model under two different circumstances. The first circumstance is referred to as a pandemic situation. In a pandemic situation it is assumed that mask wearing is universal due to the mask mandates for the pandemic of a different disease. In this case, individuals wear masks around their contacts throughout the course of their illness, regardless of their symptoms. This includes mask wearing around individuals in their household when staying home for the second infectious stage when malaise is at a peak, since in a pandemic individuals will assume that their illness is the disease of the pandemic and they will take extra precautions to prevent the spread of that disease. The second circumstance is referred to as the typical situation or pandemic free situation. In this case we want to understand the effect of a mask mandate for those infected with the flu. Since an individual may not know they have the flu until symptoms begin, they will not be wearing masks for the first infectious stage. And we assume the mask mandate can only be enforced outside of the household, so individuals will not wear masks while in the second infectious stage due to them staying home for Malaise. In the recovered stage individuals will have immunity from infection, however this immunity will eventually end and they will reenter the susceptible population.
Our model showed that in a pandemic situation, much like the COVID-19 pandemic, where there is universal mask wearing to prevent the spread of the pandemic disease, universal mask wearing is sufficient to prevent an influenza epidemic in the absence of vaccination. Furthermore, combining vaccination intervention with universal mask wearing does not significantly change the outcome of the influenza model meaning that in such a pandemic situation, vaccination may not be necessary or cost efficient for preventing the spread of influenza.