Estimating the Future Number of Cases in the Ebola Epidemic
The first cases of the West African Ebola epidemic were reported on March 22, 2014, with a report of 49 cases in Guinea. The outbreak has affected several West African countries and with more than 20,000 cases reported by the end of 2014. To aid in planning the public health response, CDC developed a modeling tool called EbolaResponse to help in estimating the potential number of future cases.
The Ebola epidemic continues in Guinea, Liberia, and Sierra Leone; however, the epidemic has changed over time. Cumulative cases in Liberia now appear to be growing at a dramatically slower rate. CDC’s January 8, 2015 report updates the estimates reported in the September 26, 2014 MMWR of the future total number of cases in the Ebola epidemic in Liberia and Sierra Leone. Estimates are provided for Guinea for the first time.
The new report estimates that by March 20, 2015:
- Cumulative cases in Liberia could range from 8,400 cases (lower bound 8,400 to upper bound 8,500) to 12,700 cases (12,600–12,800 when corrected for possible underreporting using a factor of 1.5).
- Cumulative cases in Sierra Leone could range from 15,700 (13,500–20,100) to 31,300 (27,100–40,100 when corrected for possible underreporting using a factor of 2.0).
- Cumulative cases in Guinea could range from 5,500 (4,600–7,200) to 11,000 (9,100–14,500 when corrected for possible underreporting using a factor of 2.0).
Using information up to December 27, 2014, CDC estimates that the total number of Ebola cases in Liberia and Sierra Leone will be much lower than conditions at the end of August indicated. The model includes confirmed, probable, and suspect cases because the tracking, testing, and re-categorization of probable and suspect cases consume public health resources.
Since the initial September modeling report, a massive public health response by the United States, international partners, and the affected nations’ governments and communities have significantly altered the course of the epidemic. The latest modeling report shows that recognizing cases early, providing prompt infection control and isolation of patients, and conducting safe burials can substantially decrease the number of cases, and are essential in controlling and reversing the epidemic. Changes in the course of an epidemic call for periodic updating to assist in planning next steps in the response. Forecasting the potential number of future cases can help decision-makers plan resource requirements and can help explain why the actions needed to reduce cases are so important.
The new model estimates project a variety of scenarios to assist public health officials and governments in planning resources as well as community outreach efforts. The estimates demonstrate that much work remains to control and end the epidemic. Even a small relaxation of vigilance could result in an escalation in new cases. The report shows that practicing safe burials and having prompt and effective isolation and treatment of Ebola patients is essential.
Several factors contribute to the new estimates being significantly lower. The number of Ebola cases reported weekly in Guinea and Liberia has decreased. These decreases are due, in large part, to an increase in hospitalization rates, use of Community Care Centers as well as effective home isolation, and safe burial practices. The EbolaResponse model allows for changes in clinical and behavioral factors that can affect estimates of future cases of Ebola.
To fully understand the impact of these interventions, periodic updating of the estimates is needed to assist in planning next steps in the response. CDC plans to update the model estimates on a regular basis to assist with planning and public health decision making.
The September 26, 2014, Morbidity and Mortality Weekly Report (MMWR), Estimating the Future Number of Cases in the Ebola Epidemic—Liberia and Sierra Leone, 2014–2015, estimates the future number of cases if current trends continue. The MMWR also adjusts the number of cases based on estimated underreported cases.
- By September 30, 2014, CDC estimates that there will be approximately 8,000 cases, or as high as 21,000 cases if corrections for underreporting are made.
- Without additional interventions or changes in community behavior, CDC estimates that by January 20, 2015, there will be a total of approximately 550,000 Ebola cases in Liberia and Sierra Leone or 1.4 million if corrections for underreporting are made.
- Cases in Liberia are currently doubling every 15-20 days, and those in Sierra Leone and Guinea are doubling every 30-40 days.
- Halting the epidemic requires that approximately 70% of Ebola cases be cared for in Ebola Treatment Units or, if they are at capacity, at home or in a community setting in which there is a reduced risk of disease transmission and safe burials are provided.
- If conditions remain unchanged, the situation will rapidly become much worse.
- We know how to control and eventually stop the epidemic. Halting the epidemic requires placing up to 70% of patients into either an Ebola Treatment Unit or in a community setting in which the risk of disease transmission is reduced and safe burials are provided.
The cost of delay will be devastating. The number of cases is doubling about every 20 days. Every month of delay in reaching the 70% target will increase the number of patients, which means more cases and more deaths and the need for even more beds and other resources.
Instructions of how to use the adapted copy of EbolaResponse model
A previous version of the EbolaReponse tool can be downloaded here.
The version of EbolaResponse has been adjusted to allow for consideration of impact for number of available of beds in Ebola treatment units (ETUs) and Ebola Community Centers (ECCs). That is, as ETU and ECC capacities are increased and become available over time (i.e., a schedule of availability), this adapted version of EbolaResponse allows a user to fit the percentage of patients in ETUs and ECCs such that the patients in each category match the available beds.
What basic question can a user address with this version?
A user can determine, by entering their chosen schedule of ETU and ECC beds, what the impact of such a schedule may have on the course of the epidemic (e.g., is the chosen schedule enough to “bend the curve?”).
EbolaResponse was designed to calculate the beds needed due to user inputted percentages of patients in either hospitals or non-ETU settings such that there is a reduced risk of onward transmission (including safe burials as needed) (1).
In this version, the user can input a schedule of anticipated/ actual beds available, separated into ETU and ECC beds, and then alter the percentage of patients into either ETUs or ECCs such that the number of patients in each of those two categories match (as close as possible) the beds available.
Days in bed: In this version, EbolaResponse is programed for an average of 7 days-in-bed/patient, based on published data (2).
- NAVIGATION: To ensure that a user “sees” all the relevant pages, users may find it easier to navigate through this version using page tabs (bottom of each sheet).
- DATA ENTRY: Users can either use the default values or enter their data in the following tabs:
- Entry cases beds
- BED SCHEDULE ENTRY: User enters desired schedule of beds in ETUs and ECCs on the “Entry cases beds” tab. The page is titled: “DATA ENTRY: Reported Cumulative Cases, Schedule to ETU and ECC beds”
- Recommendation: For each new schedule of available beds, it is recommended that users create and save a separate version of the adapted EbolaResponse tool.
- REPORTED CASE ENTRY: User enters reported cases on the “Entry cases beds” tab. NOTE: In the published paper (1), the authors used Ebola confirmed + probable +suspect cases.
- Recommendation: Since suspected cases use/ require beds and often lab tests, it is recommend that users include such cases when considering “beds needed.”
- USING THE MODEL: FITTING MODEL to reported cases.
- Step 1: Go to tab “Country_uncorrected”
- Step 2: Fitting model to reported data: Alter, if needed, the percentage of patients in each category so that the plot in the “Goodness-of-Fit” graph (on same page) matches that of the reported cases-to-date.
- Note that a user can alter Daily Transmission Risk, by patient category, and “imported cases” – all on the same page.
- Additional information on Data Entry can be found in the model by clicking the button labeled “How to enter data.”
- Step 3: Modeling impact of Schedule of ETU and ECC beds: On the same spreadsheet page “Country_uncorrected”, there are 2 graphs labeled “ETU Beds in Use Vs. Beds available (Uncorrected)”, and “ECC Beds in Use Vs. Beds available (Uncorrected)”. The red dotted line in each of those graphs should plot the schedule of beds as entered by the user in the previous data entry step (Point C, above).
- Step 4: The user then alters the percentage of patients placed in each of the two relevant categories (hospitalized (ETUs); and, “Effective home isolation or equivalent (ECCs)), until the plot of the number of patients in ETUs and ECCs match the plotted beds available.
- Note that EbolaResponse uses 30 day time steps.
- Step 5: Adjusting for underreporting: In the “Entry cases beds” a user can enter a factor to adjust for potential underreporting. Based on the original published report of September 26, 2104 (1), the default correction factor for underreporting is 2.5. The updated estimates (December 27, 2104) used a correction factor of 2.0 for Sierra Leone and Guinea, and 1.5 for Liberia. A user can readily alter the correction factor.
- Step 6: Go to tab “Country_corrected”
- Step 7: Repeat, on the “Country_corrected” page, Steps 2 -4.
Questions about the Tool?
Contact Martin Meltzer, firstname.lastname@example.org
1. Meltzer MI, Atkins CY, Santibanez S, Knust B, Petersen BW, Ervin ED, Nichol ST, Damon IK, Washington ML. Estimating the Future Number of Cases in the Ebola Epidemic —Liberia and Sierra Leone, 2014–2015. Morb Mort Weekly Rept; Supplement; 63(3); Sept. 26, 2014.
2. WHO Ebola Response Team. Ebola virus disease in West Africa–the first 9 months of the epidemic and forward projections. N Engl J Med. 2014 Oct 16;371(16):1481-95. doi: 10.1056/NEJMoa1411100. Epub 2014 Sep 22.