Replacement Personnel May Cause Disease to Spread Faster During an Outbreak

Replacement Personnel May Cause Disease to Spread Faster During an Outbreak

A mathematical biologist at the University of Vermont and his colleagues have discovered that the pace and tempo of an epidemic can be impacted by how a hospital manages its personnel based on outbreak response and its institutional sick leave policies. Samuel Scarpino, PhD, who led this research, asks people to imagine a scenario in which a healthcare worker contracts influenza; he or she goes home to recover and a healthy replacement worker comes to the hospital. This sort of substitution, called a "relational exchange," can actually accelerate the spread of some epidemics. Scarpino explains that their model is a departure from the traditional "mass-action" disease models that are used by the Centers for Disease Control and Prevention (CDC) and other public health entities that do not account for individuals' social contacts that can influence pathogen transmission.

Samuel Scarpino, PhD

By Kelly M. Pyrek

A mathematical biologist at the University of Vermont and his colleagues have discovered that the pace and tempo of an epidemic can be impacted by how a hospital manages its personnel based on outbreak response and its institutional sick leave policies. Samuel Scarpino, PhD, who led this research, asks people to imagine a scenario in which a healthcare worker contracts influenza; he or she goes home to recover and a healthy replacement worker comes to the hospital. This sort of substitution, called a "relational exchange," can actually accelerate the spread of some epidemics. Scarpino explains that their model is a departure from the traditional "mass-action" disease models that are used by the Centers for Disease Control and Prevention (CDC) and other public health entities that do not account for individuals' social contacts that can influence pathogen transmission.

Scarpino explains that their relational exchange model is different in two key ways. "What makes other models mass-action models is that contacts between individuals happen randomly," says Scarpino, an epidemiology expert who leads university's Emergent Epidemics Lab. "For example, say I am going to decide with whom I am going to have dinner tonight. I put the names of everyone who lives in my town of Burlington, Vt. into a hat and then I draw one name out of that hat. Obviously that’s not how our social contacts are structured. The change that we implemented in our paper was to impose a slightly more realistic assumption about social contacts – mainly that we have structured contacts and some people have more connections than others and some people have fewer, and this gives us what we call a network model, where we are imposing this more realistic social network structure into our analysis. When healthcare workers are replaced in a mass-action or random mixing case, we don’t see the same type of effect because the contacts are already random, and it is already mixed maximally. However, in the realistic social contact network case, the relational exchange model, it matters when we replace these individuals because we are keeping track of the social contacts."

Scarpino adds, "One take-home of our study is that it may be very difficult to predict the size of a disease outbreak and that mass-action models can't really account for the kind of sudden speed-up and slow-down in transmission that many real-world epidemics show."

Logic tells us that replacing sick workers with healthy ones will stop an outbreak, but that's not always the case, according to the findings from Scarpino's team.
"We’re not suggesting that sick people keep working; in fact, the conclusion from the model would be that sick people need to stay home even earlier than they are staying home," Scarpino emphasizes. "You replace people quick enough, then you will slow down the spread of disease – people’s intuition is right in that scenario, it just turns out that in our analysis of influenza data in the U.S., we see that people are not being replaced quickly enough. Say a school teacher isn’t feeling well; they come into work on the first day and while they are sick they are going to be very likely infecting some of their social contacts at work. Then on the second day that teacher feels too ill to come into work and so they get replaced with a healthy worker. We are taking a healthy individual who is very likely not in a situation where there is as much disease and we are moving them into a place where we know there are a lot of sick people because someone is so sick that they have to stay home from work. So we are placing a healthy person into a high-risk situation with respect to the disease transmission. That healthy person is also going to have a whole bunch of additional social contacts that are essentially brought closer to the source of infection because that person may become infected and take their illness back to their house."

The new research suggest that as epidemics approach their peak, replacement workers are in a more dangerous situation than conventional disease models would suggest. The implications for outbreak management as well as personnel and resource budgeting are significant.

"If you're making strategic decisions about how many healthcare workers you need, how many people you might expect to show up in the hospital, or how many courses of antivirals or antibiotics you might need, then the pace and tempo of cases matters deeply," Scarpino says. Being able to quickly replace sick workers not just at the beginning of an epidemic but at its peak is one key to limiting its spread. But if you can't predict that it's about to exponentially speed up "with a huge number of new cases coming in," Scarpino says, "that could easily overwhelm the healthcare system and hospitals."

Scarpino and his team are actively following up on a few aspects of the Nature Physics paper, including writing a new paper that is focused on issues faced by stakeholders involved in clinical and public health decision-making. "With this first paper we were trying to convince the broader scientific community of the validity of our data findings," Scarpino confirms. "The second step is to translate that data into the types of recommendations that can be acted upon, and we think there are two main concepts. The first is that the very early predictions for the rate at which people would become infected and seek care at hospitals during outbreaks such as SARS, H1N1 and Ebola were all off substantially. Our explanation for that error is because in all of those cases, the early predictions were based on these mass-action or random mixing models, and we know the world doesn’t work like that."

Scarpino continues, "For SARS, scientists applied one of these random-mixing models to crowded apartment buildings in Southeast Asia and then they applied that model directly to Vancouver and Toronto -- yet we know that a whole city is not going to act like a crowded apartment building. So even if they had applied that to a city in Southeast Asia, it’s already not going to work. But then to apply it to a city in another part of the world that may have a very different distribution of population, etc., it is going to lead to the types of mass prediction errors that we saw. I think the first take-home for us is that we really need to be incorporating some of these more realistic assumptions into our predictions for infectious disease preparedness and response. The second, and this relates more to our model, is if we want to prevent this kind of speed-up from happening, then we need to ensure that we have enough staff on hand so that we can replace people quickly and we need to ensure that those individuals who are at high risk for being infected, are vaccinated and have adequate PPE. That type of preparedness can help slow down both the rate of transmission and prevent any kind of speed-up as a result of this relational exchange.

The Nature Physics study had its genesis during the recent Ebola outbreak in West Africa. During that time, Scarpino observed how many healthcare and funeral workers would get sick and a healthy person would replace them--only to get sick themselves. He wondered how this kind of substitution affects the spread and speed of the virus and various other kinds of epidemics.

"Once the world community realized it was Ebola and not another hemorrhagic fever in western Africa, it started taking serious control measures that we know work," Scarpino says. "As soon as anyone showed symptoms they could isolate them, and we saw that the rate of disease transmission slow and eventually become much more controllable. I suspect that the longer we had waited to start on those quick replacements of sick personnel, the more disease spread we would have seen in an accelerating."

Disease modeling is becoming an increasingly important part of healthcare and epidemic planning. As we have seen, mass-action models assume that infected people interact with other people at random, like so many molecules bouncing in the air. This approach has been defended on these grounds: if it's not a perfect reflection of the real world then at least these models give forecasters a sense of the worst-case scenarios. But Scarpino and his colleagues disagree. Their approach represents this essential worker behavior in a dynamic network, where infected individuals, referred to in scientific jargon as "infected nodes," get "rewired" when a replacement worker comes in. An emergent property of their model shows that there can be critical transitions where the epidemic accelerates faster than the supposed worst-case scenario of a mass-action model -- and just as suddenly burns out. Their model points to another worrisome dynamic. Imagine during a disease epidemic that the worker replacement rate is "high enough to keep an outbreak under control, but that after some time the rate is slightly reduced. This might occur, for example, after the initial fear wears off," the scientists write. Their model indicates that a slight reduction in the rate at which, say, sick teachers or doctors are replaced "can push the system over a discontinuous transition," such that a "microscopic change" in the rate of replacement can lead to a very large change in disease prevalence. Suddenly, the epidemic is now in a higher gear, and if public health officials then wish to "bring the system back to its previous state," they write. "The replacement rate must be increased well beyond its previous value for the system to return to the initial state."

The authors of the Nature Physics paper were able to back up their modeling results with data from actual epidemics. Scarpino and his two co-authors, Laurent Hebert-Dufrene at the Santa Fe Institute, and Antoine Allard at University of Barcelona in Spain, analyzed existing national data from 17 flu outbreaks in the U.S., 25 years of state-level flu data, and 19 years of dengue fever data from Puerto Rico. They found the predicted pattern -- accelerating exponential transmission near the outbreak peak -- in most of the outbreaks of influenza, but almost not at all in dengue.

"In this paper we use dengue fever as a kind of null model," explains Hebert-Dufrene, who says he didn't expect to see the same accelerations in dengue from replacement behavior, "because in dengue, behavior is not as important," he says. "You can stay home, but the mosquitoes will find you."
Scarpino has been keeping a watchful eye on the issue of pandemics and antibiotic resistance and notes, "It’s challenging for scientists because it takes time to fully understand what’s going on for each new pathogen, so there’s a tension between wanting to be involved in the response as quickly as possible and then also wanting to make sure that we’re spending enough time diagnosing the things that have already happened historically and learning from that. I think we need to make sure that we’re taking both approaches and that we are employing our tools and knowledge for current and ongoing outbreaks and also spending time reflecting and learning from past outbreaks. What I find to be most sobering is that for many of these pathogens that have 'surprised' us, I see it as a failure of awareness or surveillance. For example, regarding Ebola, there’s evidence going back to the 1880s of individuals with Ebola and Ebola antibodies in those West African countries. Whether they are getting exposed abroad and coming back or getting exposed in-country, we have known for decades that Ebola was present there -- and so we should not have been caught off guard about emergence in those countries. Similarities could be suggested for Chickungunya, Zika and influenza, and we should have been more prepared for these diseases to arise." 
Scarpino continues, "Another challenge is trying to get out in front of antibiotic resistance. I think the solution in the near term has to be the development of new antibiotics and I am not sure what it will take to convince governments and pharmaceutical companies that we have to invest more seriously in the development of antibiotics. I think there is plenty of convincing evidence across all spectrums of science and medicine that we need to be doing this, and we’re not.  So I am at a loss as to what it will take to convince people this has to happen today – it probably needed to happen five or 10 years ago."
For now, Scarpino and his colleagues are turning their attention to next steps involving their research. "We would like to test some of the predictions of our model in different contexts than the one we did in the published paper. One of the predictions would be is, if you have two different companies – one of which has a more liberal sick leave policy that people take advantage of -- we would see less of an effect of this relational exchange there.  Similar things could be said about countries with different codified requirements surrounding paid sick leave. So we’d like to evaluate what kind of effect these different possible interventions may have had in the past. I think we could demonstrate that company A gives out twice as many sick days but maybe has a lower rate of absenteeism because people stay home more often and they break these chains of transmission and you don’t get these large workplace outbreaks. That’s another way to convince people it matters to quantitatively demonstrate with concrete examples from different settings. Second, to get a sense for how this process relates to different disease beyond dengue and influenza, we'd like to categorize different pathogens as being more or less likely to be affected by this type of a placement process. As we allude to in the paper, to start bringing more of these realistic and prevalent behaviors into our mathematical models. As we say at the beginning of the paper, we all know people get sick and they stay home and especially in the U.S., we all know that people come into work more often than they should because they don’t want to burn sick days, and so we need to be putting these kinds of things into our models, especially those models that are being used to generate predictions that are going into the decision-making pipelines at the clinical and/or public health levels."

Scarpino adds, "One of the things we’ve been working on for 10 years is delivering our mathematical, computational and statistical models to the clinical and administrative people making public health decisions. We’ve worked with the CDC, the state of Texas and others to translate our work into things that are actionable. We built a suite of tools for the state of Texas called the Texas Pandemic Preparedness Toolkit that will allow them to use our models to do things such as make strategic decisions about personnel, about stockpiling ventilators, scheduling staff, etc. It's important to start connecting people from different perspectives – people making decisions and people solving the problems and making expensive decisions on how to prioritize expensive resources in a time of uncertainty, I think that’s one area where these types of models can be so useful because they allow us to enumerate our assumptions and to engage in an exercise of structured reasoning to see through these complicated systems and to model the types of uncertainties we may encounter; we can also get a better sense of how an uncertainty is going to relate to potential outcomes to the utility of different decisions that we might make." 

Reference: Scarpino SV, Allard A and Hébert-Dufresne L. The effect of a prudent adaptive behavior on disease transmission. Nature Physics. Aug. 1, 2016.

 

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish