The Potential Impact of Increased Hospital Capacity to Contain and Control Ebola in Liberia

Eric Lofgren, Caitlin Rivers, Madhav Marathe, Stephen Eubank, Bryan LewisarXiv 2014 · DOI · arXiv

Abstract

West Africa is currently experiencing a severe outbreak of Ebola virus disease (EVD). As part of the international effort to address this outbreak, the United States has committed to building specialized Ebola treatment facilities with 1700 beds. However, the effectiveness of this increase in the available healthcare facilities to treat Ebola is unclear, especially in light of the rapidly increasing number of cases. Adapting a previously validated mathematical model of Ebola in West Africa, we examine the potential impact of an increase in hospital capacity to mitigate the impact of Ebola under several scenarios, ranging from the planned scenario of 1700 beds in 10 weeks to a considerably more aggressive approach of twice the number of beds in 5 weeks. We find that even for the most aggressive scenarios, while increasing the availability of healthcare reduces the number of Ebola cases and slows the outbreak, it is not sufficient to stop the epidemic within the next three months. We find that only a combination of increased hospital beds and a dramatic decrease in the rate of transmission within the community can bring the epidemic under control within the near future.

Introduction

In late 2014, the Ebola virus disease (EVD) outbreak in West Africa represented a global health emergency characterized by terrifying exponential growth. In response to the escalating crisis in Liberia, the United States government committed to a massive logistical undertaking: the construction of 17 specialized Ebola Treatment Centers (ETCs) providing 1,700 beds. While this intervention was unprecedented in scale, its potential efficacy remained a subject of intense debate. This study utilizes mathematical modeling to quantify how different rollout speeds, bed capacities, and infection control efficacies would impact the epidemic's trajectory, specifically addressing whether infrastructure expansion alone is sufficient to halt the outbreak.

Methodology

To simulate the outbreak's progression, researchers employed a two-stage compartmental model, representing an evolution of the classic Susceptible-Exposed-Infectious-Recovered (SEIR) framework. The model was uniquely granular, splitting the "Hospitalized" (H) and "Funeral" (F) compartments into three sub-categories: specialized Treatment Centers (ETCs), local Community Care Centers (ECCs), and home-based care involving the distribution of take-home kits.

The simulation was executed in two distinct phases:

  1. Deterministic Stage: Ordinary differential equations (ODEs) were used to fit the model to historical data up to October 1, 2014, utilizing weighted least-squares optimization.
  2. Stochastic Stage: To forecast the subsequent 90 days, the researchers transitioned to stochastic simulation using Gillespie’s algorithm (with a tau-leaping approximation). This approach, implemented in Python via the StochPy library, allowed the model to account for stochasticity and provide a range of probable outcomes rather than a single deterministic trajectory.

The study utilized a curated time series of laboratory-confirmed, suspected, and probable EVD cases in Liberia, sourced from the World Health Organization (WHO) and the Liberian Ministry of Health.

Results and Analysis

The Baseline and Planned Interventions

Without further intervention beyond the status quo as of October 1, the model predicted a median of 151,245 incident cases over the following 90 days. The planned US intervention—deploying 1,700 beds over a 10-week period—was predicted to significantly reduce this burden to approximately 67,134 cases under ideal infection control conditions. While this represents a substantial reduction in morbidity, the model indicated that the epidemic would continue to grow exponentially.

The "Super Rollout" Scenario

The researchers also tested a "Super Rollout" scenario, which involved doubling the bed capacity to 3,400 and accelerating the deployment to just 5 weeks. Surprisingly, this aggressive expansion only marginally improved the outcome, resulting in 60,560 predicted cases. This finding highlights a critical "exponential trap": when an epidemic is growing rapidly, simply adding capacity cannot keep pace with the rate of new infections if the underlying transmission rate remains unchanged.

The Path to Control

The only scenario in which the epidemic peaked and began to decline involved a multifaceted approach. This required combining the Super Rollout with two additional factors: a 25% reduction in the time-to-hospitalization (the window during which an infectious person remains in the community) and a 50% reduction in community transmission. Under these conditions, the predicted case count dropped to approximately 10,263, shifting the curve from exponential growth toward containment.

Discussion

Two primary insights emerge from this modeling. First, the efficacy of infection control within healthcare facilities is more impactful than the absolute number of beds. The variance in outcomes between "ideal" and "conservative" infection control was greater than the variance between standard and doubled bed counts. Second, the time-to-hospitalization serves as a critical lever; reducing the time an infectious individual spends in the community is essential for breaking the chain of transmission.

Limitations

The model's accuracy is inherently limited by the quality of reported data, which was likely subject to under-reporting during the peak of the crisis. Additionally, the model assumes a constant Case Fatality Rate (CFR), whereas in reality, the CFR might fluctuate as healthcare systems become overwhelmed or as clinical protocols improve.

Conclusion

The findings suggest that the international community cannot "build its way out" of an Ebola epidemic through hospital beds alone. While increasing hospital capacity is a vital component of the response, it must be paired with aggressive community-level interventions. Future policy and research should prioritize the logistics of rapid patient transport, improved public health messaging to encourage early treatment-seeking, and community-based strategies to reduce transmission outside of clinical settings.

Summary updated

Abstract

West Africa is currently experiencing a severe outbreak of Ebola virus disease (EVD). As part of the international effort to address this outbreak, the United States has committed to building specialized Ebola treatment facilities with 1700 beds. However, the effectiveness of this increase in the available healthcare facilities to treat Ebola is unclear, especially in light of the rapidly increasing number of cases. Adapting a previously validated mathematical model of Ebola in West Africa, we examine the potential impact of an increase in hospital capacity to mitigate the impact of Ebola under several scenarios, ranging from the planned scenario of 1700 beds in 10 weeks to a considerably more aggressive approach of twice the number of beds in 5 weeks. We find that even for the most aggressive scenarios, while increasing the availability of healthcare reduces the number of Ebola cases and slows the outbreak, it is not sufficient to stop the epidemic within the next three months. We find that only a combination of increased hospital beds and a dramatic decrease in the rate of transmission within the community can bring the epidemic under control within the near future.
BibTeX
@misc{lofgren2014potential,
  author = {Eric Lofgren and Caitlin Rivers and Madhav Marathe and Stephen Eubank and Bryan Lewis},
  title = {{The Potential Impact of Increased Hospital Capacity to Contain and Control Ebola in Liberia}},
  year = {2014},
  eprint = {1410.8207},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1410.8207},
}
Figure