Using data gathered on the ground in Liberia, Yale researchers have been able to run simulations to answer the question: How best can the world stem the Ebola epidemic?

Researchers who have been following the outbreak, which began in earnest earlier this year in Guinea, were able to model the spread of the disease in Liberia, where the virus later spread. They found that the only way to contain the disease was through a combined intervention that targets all contexts in which the disease could be transmitted.

“We were trying to tackle, in a very general way, [Ebola] transmission in all of Liberia,” said Natasha Wenzel, research associate at the Yale School of Public Health in the Center for Disease Modeling and Analysis.

Researchers obtained their results by looking at data published by the Liberian government on the number of people infected and how each person would have come into contact with the disease. In building the mathematical model, researchers created a virtual population, and divided it into subpopulations of susceptible, infected, deceased and recovered individuals, as well as other categories.

They then modeled the probability of a given individual transitioning between the subpopulations in a given time from the data collected. To model certain transitions, namely from susceptible to infected individuals, the researchers divided the means of infection into multiple types — the general community, patients in hospitals and those attending funerals.

After inputting the data that the Liberian government has collected weekly to track disease transmission, the researchers were able to make a rough prediction of how the disease would spread in the country. They then considered various intervention methods to keep the disease from spreading, specifically hygienic burials for hospital deaths and community deaths, quarantining infected individuals and quarantining those in contact with the infected. They then modified the model to reflect the efficacy of each intervention.

Those simulations showed that a combined intervention implemented in 60 percent of instances would be more effective than any individual method alone implemented in 90 percent of instances. According to their predictions, a combined intervention could reduce the number of newly infected patients to zero within six months.

“You can’t concentrate on one thing thinking that [it] is more important,” said Abhishek Pandey, postdoctoral associate in epidemiology at the School of Public Health. “You really need to actually take a combined approach and take it aggressively.”

The researchers also found that funerals for diseased patients were “super spreader” events. If hygienic burial practices alone were adopted, the number of new cases each day could be reduced by half within two months, much faster than any other single intervention, though this prediction was based on the premise that hygienic burials would be implemented more then may be realistically possible.

“In previous outbreaks, funerals have been shown to be the most important factor,” Pandey said. “Now we see that if you just concentrate on funerals, you’re not going to curtail [the epidemic] very soon, so combining [hygienic funeral practices] with contact tracing and isolation is absolutely necessary.”

Researchers also noted that their estimation of the effectiveness of hospital interventions may be overly optimistic, as the resources available to each hospital vary greatly. In particular, many Liberian hospitals are understaffed and underbedded, so isolation may be infeasible. The researchers further noted that because of the lack of data on the geographic distribution of the disease, they were unable to build a spatial model.

Jan Medlock, professor in the Department of Biomedical Science at Oregon State University and co-author of the paper, said that one of the advantages of the study is that the researchers used well established modeling methods, and that the scientific community will not have to digest novel techniques that may distract from the study’s findings.

Wenzel said that having daily updated data during an outbreak rather than only at the end is extremely valuable.

“We had to update our model almost every week because more and more data was coming in, so … we could actually capture the most recent [situation],” Pandey said.

This year, 2,766 people have died of Ebola in Liberia.