How Community Response Stopped Ebola

Figure 1: Photograph of Monrovia, Liberia (credit: Erik (HASH) Hershman)
Figure 2: Mock simulation of pandemic arising from central Africa and spreading through land and air transportation.

In January of 2014 I spoke at the World Health Organization in Geneva, Switzerland. During the presentation I described our work on pathogen evolution and the effect of long range transportation [1]. I included our video about the risk of Ebola due to increasing transportation in Africa as it develops economically (Figure 2). The general expectation is that prior experience is a predictor of future events, but in this case it doesn’t work. A lot of respect was shown for our work. However, to my knowledge, nothing was done related to the Ebola risk.

Two months later, in March 2014, the largest outbreak of Ebola to date began in West Africa. The number of cases increased exponentially, roughly doubling every week. The response was not sufficient, and if an outbreak is growing exponentially, being behind in the response means being completely ineffective [2]. The outbreak become the worst Ebola epidemic to date by a factor of ten.

To combat the disease, public health officials employed contact tracing as a containment strategy [3]. Under contact tracing, a patient admitted to a hospital is asked about their recent direct contacts, and those contacts are monitored for early symptoms (e.g. fever) and isolated if they show them, or even just isolated for the incubation period of the disease [4,5]. This technique can be effective in small communities, or when there are only a few cases, but the spread of Ebola to dense urban centers in West Africa made it difficult or impossible to implement. There are three reasons contact tracing is difficult to implement, and can get behind the growth in cases, making it ineffective.

Figure 3: Timeline for Ebola infections, showing susceptible individuals who become exposed, infectious and “removed.” Symptoms start at about the time of becoming infectious, but symptoms that are unique to Ebola only happen much later so that there are many opportunities to expose others. “Removed” refers to either recovery with some immunity or death which happens in about 50% of cases.

First, the early symptoms of Ebola are like other viral infections, fever, chills, headache, sore throat, muscle pain, … . By the time more serious symptoms happen so that people know they have Ebola a week may have gone by, and there are many chances for them to infect other people (Fig. 3).

Second, there are many contacts that occur in an urban environment that are difficult to track down [6–9]. In West Africa shared cabs were a common way to transmit Ebola. Those with the disease didn’t know who their fellow travelers were. Other forms of anonymous interactions in public transportation and markets can be sources of contact and transmission.

Third, for interactions that can be documented and traced, the number of health care workers needed grows with the number of infected cases. Properly trained individuals are needed for interviewing, compiling lists, seeking out contacted individuals, and performing monitoring and isolation of identified individuals. For an exponentially growing number of cases, the number of responders must grow exponentially [3,10,11]. 10 or 20 cases can lead to thousands or tens of thousands of contacts that need to be found.

Scenario simulations were projecting that the epidemic would not be stopped, i.e., go to “burnout” — which would have meant 10 million people dead in West Africa, even if the epidemic didn’t spread elsewhere [12,13].

Figure 4: Detecting fever by monitoring individuals in the community allows for early isolation of individuals with symptoms even if they don’t have Ebola, preventing the spread of infections.

In light of these problems, we developed an alternative concept for response. We advocated a community-based strategy of monitoring and travel limitations [14]. Rather than focusing on the very sick individuals that come to hospitals and trying to track down all of their contacts, the intervention would take place on the level of communities. Each member of a community would be screened daily for fever and isolated if they show these early symptoms (Fig. 4). Catching infections as the first symptoms develop nearly eliminates the chances for transmission. Travel restrictions would help prevent community to community transmission. Whether or not to implement travel restrictions became a political rather than scientific discussion [15,16], but for the Ebola response it was secondary to screening of people within the community, so we focused on that aspect.

Setting up this approach requires having neighborhood teams do daily monitoring. In collaboration with the Army Corps of Engineers, experts in managing large projects, we put together a specific plan for such an effort.

I was in contact with key individuals and organizations that were involved in the response to Ebola, including the head of the response at the United Nations, members of the CDC responsible for the US response, and of the National Security Council (NSC) at the White House. While they were receptive to our strategy, it was taking too much time to implement. Time was not on our side. The epidemic was growing exponentially.

As we were working on having this strategy implemented, I found out that a my friend Katherine Collins was involved with Last Mile Health, a health organization active in Liberia. They connected me with individuals involved in the response to Ebola on the ground, including Dr. Joa Ja’keno Okech-Ojony and Dr. Paul Steven Ayella Ataro.

I called them up in early October and found out that three weeks before, in mid September, teams of individuals began going door to door in neighborhoods with infrared thermometers to screen for fever. The teams were made up of community members who knew the people in the neighborhood.

Figure 5: Counts of confirmed (red), probable (green) and suspected (blue) Ebola cases in LIberia showing the rapid increase until mid September, 2014, followed by a rapid decrease after community response started including door to door monitoring.

The results were dramatic. The epidemic that was exponentially growing, fell exponentially [17] (see Fig. 5). To the confusion of some international observers, the expected number of sick people weren’t showing up at the special Ebola care facilities constructed in Liberia. Even two months later, reports in the news were saying that they didn’t know where bodies were, that they must be being hidden [18,19].

It took three months, till mid December, for the same approach to be transferred to Sierra Leone, and the same thing happened — the number of cases declined rapidly. By March 2015, these interventions brought the number of active Ebola cases to zero in Liberia [20]. A few small outbreaks occurred later but they were stopped quickly. In the spring, the WHO stated in its reports that ‘community engagement’ was key to stopping the disease [21], and included the importance of community actions as a theme in their lessons learned [22,23]. For the WHO, community engagement includes involving community leaders and groups in the planning and implementation of a range of response measures (case detection and monitoring, contact tracing, and safe burials). While a variety of measures can be helpful, our analysis suggests that one particular measure, community monitoring, by itself can be effective, and without it the success of response efforts is unclear.

Figure 6: Simulation of the spread of ebola. Susceptible individuals (grey) become infected (yellow) then infectious (red) and removed (black). At the point of inervention, individuals are monitored in communities with the disease. Optionally, communities are identified as having active transmission (red blocks), monitored for transmission (yellow blocks) and clear (blue blocks) in order to implement travel restrictions.

We constructed a simulation to show how community monitoring works [24](Fig. 6) and that it is effective even if it is not fully implemented — a property that is necessary to be successful in the real world. The video shows an example of a simulation. After the initial growth of infections (represented by the small yellow and red squares) the intervention begins. The intervention included door to door screening and, in some scenarios, travel restrictions.

Figure 7: Simulation results with only 50% compliance (blue) starting Sept. 15, are consistent with the data on cases from liberia.

Perfectly enforced screening would quickly stop transmission. In the real world, noncompliance is a serious issue for any strategy. The community based strategy does not require perfect implementation. With a 50 percent compliance rate, the simulated outbreak goes into an exponential decline that matches the actual decline of cases (Figure 7). Even 40 percent compliance results in declining cases.

We know of no other similarly validated explanation for the end of the outbreak.

When we add travel restrictions between exposed and unexposed neighborhoods, inter-community transmissions decrease. Neighborhoods are classified as having active transmissions (red block), those being monitored for transmission (yellow), and clear (blue). As you can see, the outbreak is contained to ever-contracting areas, allowing for more targeted care by health workers. Travel restrictions are particularly important for cases were there are very low compliance levels with the door to door screening.

While early on there was a strong resistance to quarantines, by the following summer with the Ebola epidemic still a problem in Guinea, news reports were talking about how communities welcomed quarantine to finally get rid of the disease [25].

As those communities realized, just like isolation of individuals does not mean abandoning them, neither does community quarantine mean abandoning them. It is an opportunity to focus care, prevent people from getting infected and stop the disease. If this had been the earliest response to the outbreak, many more lives would have been saved and unimaginable suffering, as well as the economic and social disruption, would have been prevented.

The same principles of community-based intervention can be applied to a wide variety of potential diseases. Understanding the lessons of Ebola’s containment will allow for these policies to be implemented more effectively in the future, reducing the death toll of future epidemics and limiting the possibility of a larger pandemics.

[1] Y. Bar-Yam, Transition to extinction: Pandemics in a connected world, Medium (July 3, 2016).

[2] M. Hardcastle, Y. Bar-Yam, Effective Ebola response: A multiscale approach, NECSI Report 2014–09–01 (September 19, 2014).

[3] World Health Organization (WHO), Contact tracing during an outbreak of ebola virus disease (September 2015),

[4] World Health Organization (WHO), Frequently asked questions on Ebola virus disease (August 2014),

[5] Centers for Disease Control and Prevention (CDC), Ebola fact sheet (April 2015),

[6] R. Hersher, A ride in Monrovia means wrestling with Ebola, NPR (October 2014).

[7] D. Stamp, Taxis, planes and viruses: How deadly Ebola can spread, Reuters (July 2014).

[8] S. Nebehay, Ebola virus spread by taxi passengers, says World Health Organisation, The Independent (September 2014).

[9] World Health Organization (WHO), Liberia: a country — and its capital — are overwhelmed with Ebola cases (January 2015).

[10] B. Armbruster, M. Brandeau, Contact tracing to control infectious disease: When enough is enough, Health Care Manag Sci. 10(4):341–355 (2007)

[11] Centers for Disease Control and Prevention (CDC), Challenges in responding to the Ebola epidemic—Four rural counties, Liberia, August­November 2014 (December 2014).

[12] M.I. Meltzer, C.Y. Atkins, S. Santibanez, B. Knust, B.W. Petersen, E.D. Ervin, S.T. Nichol, I.K. Damon, M.L. Washington, Estimating the future number of cases in the Ebola epidemic — Liberia and Sierra Leone, 2014–2015. CDC MMWR Surveill Summ 63(suppl 3):1–4 (September 26, 2014).

[13] M.F.C. Gomes, A. Pastore y Piontti, L. Rossi, D. Chao, I. Longini, M.E. Halloran, A. Vespignani, Assessing the international spreading risk associated with the 2014 West African Ebola outbreak, PLOS Currents Outbreaks (September 2, 2014), doi: 10.1371/currents.outbreaks.cd818f63d40e24aef769dda7df9e0da5.

[14] Y. Bar-Yam, DRAFT New Ebola response strategy: Local care team early detection response, New England Complex Systems Institute (October 12, 2014),

[15] T. Frieden, Opinion: CDC Chief: Why I don’t support a travel ban to combat Ebola outbreak, Fox News (October 9, 2014),

[16] Y. Bar-Yam, Response to CDC Director Frieden’s opposition to a travel ban, New England Complex Systems Institute (October 13, 2014).

[17] Y. Bar-Yam, Is the response in Liberia succeeding? Positive indications, New England Complex Systems Institute (October 17, 2014),

[18] T. McConnell, Some people would rather die of Ebola than stop hugging sick loved ones, Global Post (October 10, 2014).

[19] D. Snyder, J. Bavier, Empty Ebola beds in Liberia pose riddle for health care workers, Reuters / Huffington Post (November 7, 2014).

[20] H. Onishi, Last known Ebola patient in Liberia is discharged, The New York Times (March 5, 2015),

[21] WHO, The Ebola outbreak in Liberia is over (May 9, 2015).

[22] WHO, Liberia succeeds in fighting Ebola with local, sector response (April 2015).

[23] M. Chan, WHO Director-General addresses Princeton — Fung Global Forum on lessons learned from the Ebola crisis, Dublin, Ireland, WHO (November 2, 2015).

[24] D. Cooney, V. Wong, Y. Bar-Yam, Beyond contact tracing: Community-based early detection for Ebola response, PLOS Currents Outbreaks (May 19, 2016), doi: 10.1371/currents.outbreaks.322427f4c3cc2b9c1a5b3395e7d20894. arXiv:1505.07020 (May 26, 2015).

[25] K. Camera, Guinea quarantines coastal towns to end Ebola, VOA (June 30, 2015),