Pages Navigation Menu

The science site for Baby Boomers

2014’s Ebola Prediction Fail

Credit: Graph by Susan E. Swanberg based on CDC and WHO data

Credit: Graph by Susan E. Swanberg based on CDC and WHO data

An unprecedented epidemic

A devastating Ebola epidemic hit the people of West Africa in 2014, killing almost 8,000.

Approximately 20,400 suspected and confirmed Ebola cases have been reported so far in Guinea, Liberia and Sierra Leone,* a far cry from the 1.4 million worst case scenario calculated by a Center for Disease Control and Prevention (CDC) research team.

Whether the CDC’s worst case scenario was in error, intervention by international medical teams altered the trajectory of the epidemic or the press exaggerated the significance of the worst case scenario will likely be debated for a long time.

Dire predictions of an expanded epidemic

The World Health Organization (WHO) was notified of an outbreak of Ebola in March of 2014, and the epidemic grew in strength over the course of eight or nine months. West Africa’s medical community was overwhelmed, and cultural practices, including the care and respect afforded the dead by surviving relatives and friends, contributed to an upsurge in cases.

At some point during the epidemic, accusations that the world was ignoring a potential global catastrophe began to appear in the news.

In July of 2014 Ken Isaacs (Vice President of Programs and Government Relations for the international relief organization, Samaritan’s Purse) wrote a piece in the New York Times asking “Why Are We Ignoring a New Ebola Outbreak?” The Telegraph published a story titled “How the world ignored Africa’s Ebola tragedy.”

Former U.N. head Kofi Annan, quoted in several news reports, claimed that the international community ignored Ebola until the disease traveled to the U.S. and Europe.

The epidemic was already unprecedented, but in the fall of 2014, predictions of a dire nature started to be published. “Ebola could infect 1.4 million in Liberia and Sierra Leone by end of January,” was the headline for a September, 23, 2014 Washington Post story. “Every Second Counts: New Ebola Report Predicts Huge Spike Without Action,” said Maggie Fox of NBC News. “CDC: Ebola cases could reach at least 550,000 by January [2015],” said a CNN headline.

The source for these and similar news articles was a September 26, 2014 report published in the CDC’s Morbidity and Mortality Weekly Report (MMWR). The report, titled “Estimating the Future Number of Cases in the Ebola Epidemic — Liberia and Sierra Leone, 2014-2015″ was authored by nine CDC scientists.

Digital image of the Ebola virus. Credit: Shutterstock http://tinyurl.com/l4lud4d

Digital image of the Ebola virus. Credit: Shutterstock
http://tinyurl.com/l4lud4d

According to the article’s abstract, the CDC developed a modeling tool called “EbolaResponse.” The purpose of the model was “to provide estimates of the potential number of future [Ebola] cases.” The abstract said, “[i]f trends continue without scale-up of effective interventions, by September 30, 2014, Sierra Leone and Liberia will have a total of approximately 8,000 Ebola cases.”

Looking at the actual statistics, this prediction was not far off. By the end of September, Sierra Leone and Liberia actually had about 5,220 Ebola cases. Using a “potential underreporting correction factor of 2.5” the model estimated that approximately 21,000 total cases would have occurred in Liberia and Sierra Leone by the end of September, 2014.

The correction factor was derived from a comparison of the EbolaResponse model’s prediction of how many hospital beds should be occupied (based on cumulative Ebola cases) to the numbers of hospital/clinic beds actually occupied by Ebola patients. The two figures differed by a factor of 2.5.

The CDC researchers plugged into their model numbers from several possible Ebola scenarios, including differences in levels of intervention.

With no increase in intervention, the worst case scenario predicted 1.4 million Ebola cases by January 20, 2015. Increasing the numbers of patients in treatment was the factor most likely to avoid the worst.

The worst case scenario has not come to pass, and is now highly unlikely to occur.

This protective gear would be inadequate for handling Ebola patients because the goggles might not completely cover the skin. Credit: Shutterstock EfFbi4ytORmX

This protective gear would be inadequate for handling Ebola patients because the goggles might not completely cover the skin. Credit: Shutterstock EfFbi4ytORmX

The World Health Organization’s Alternate Prediction

The CDC was not the only health authority making predictions about the course of the Ebola epidemic. On October 16, 2014, the WHO published an article about the trajectory of the epidemic in The New England Journal of Medicine. The article, “Ebola Virus Disease in West Africa — The First 9 Months of the Epidemic and Forward Projections,” made a much more cautious and measured prediction than the CDC.

According to WHO researchers, “at the current rate of increase, assuming no changes in control efforts, the cumulative number of confirmed and probable cases by November 2…will be 5740 in Guinea, 9890 in Liberia and 5000 in Sierra Leone, exceeding 20,000 cases in total.”

This prediction is more in line with the status of the epidemic as of January 2, 2015.

Limitations of the CDC model

In fairness to the scientists who developed the EbolaResponse model, it’s clear that they were transparent about their model and its limitations. The modeling tool is available in a spreadsheet here. The authors described in detail three levels of intervention that were included in several of their scenarios.

In their article, the CDC researchers reported five limitations:

1. Extrapolating future cases from (then) current trends “might not be appropriate,” said the researchers.

2. Assuming incubation and infectiousness periods based upon previous epidemics “might not be accurate.”

3. Reliance on expert opinion regarding bed occupancy might not account sufficiently for other factors relevant to the correction factor.

4. The correction factor might change over time.

5. The predictive scenarios didn’t consider the logistics requirements required to improve intervention levels.

The Role of the Press in the Prediction “Fail”

In spite of the study’s clearly disclosed limitations, and the inclusion of more possible scenarios than the worst case, many news reports emphasized the most extreme of the CDC’s predictive scenarios. (See below for links to articles.)

This is not the first time some science journalists have chosen to focus on one aspect of a study, excluding or overlooking other relevant factors. Going back to the original autism/vaccination reporting and looking at current news reports touting “cures” and “breakthroughs,” it’s clear that we need to do a better job of reporting science.

Looking critically at my own reporting on the CDC’s predictions, I see that I also referenced the worst case scenario. I did, however, qualify my interpretation and provide a detailed analysis of the significance of the prediction. As science writers and journalists, it behooves us all to keep our subjects under a magnifying glass — reading the primary literature thoroughly, skeptically and including as much detail as possible in our analyses.

One item that stands out to me now, is the CDC’s own statement about the accuracy of its predictions. In a Q & A about the predictions, the CDC asks and answers as follows:

How accurate are the estimates (using uncorrected or corrected data)? They are only estimates, and CDC cannot guarantee their accuracy; however, the estimates illustrate that if conditions remain unchanged, the situation will rapidly become much worse. As such, the estimates are a warning and a call to action.”

What role should journalists play in spreading the word about somebody’s call to action?

Was it right to broadcast the CDC’s worst case scenario the way we did? How could we do better in the future? Did hyping the CDC’s most dire Ebola prediction ensure that the public will ignore the calls for help when the next epidemic comes along?

The 2014 Ebola epidemic is not yet over, and we must continue to be vigilant and to help improve health conditions for West Africans and people around the world.

Science writers and journalists can help, too. We must avoid the temptation to hype or overextend the findings of scientists. We must look to the literature, read it carefully and ask questions.

If we fail to be the watchdogs the public needs instead of cheerleaders for those with an agenda, we will lose the trust and faith of our audience — trust and faith that might be needed to avoid the next epidemic.

__________

I encourage my readers to debate these and other questions the CDC’s Ebola predictions raise.

__________

*In addition to the cases reported above, 20 cases (including eight deaths) occurred in Nigeria, eight cases (including seven deaths) occurred in Mali, one case each occurred in Senegal and in Spain and four Ebola patients were diagnosed and/or treated in the U.S.

One of the four U.S. patients died from Ebola.

Note: This piece represents my interpretation of the reporting on the CDC’s predictions. At some point in the future, I might undertake a more scientific article on the issue.

__________

Dire Prediction Reporting

http://www.webmd.com/news/20141023/experts-predict-catastrophic-ebola-epidemic-in-west-africa-if-aid-delayed\

http://www.newscientist.com/article/dn26448-future-scenarios-show-how-easily-ebola-could-explode.html#.VKcw-Isgl94

http://www.nytimes.com/2014/10/15/world/africa/ebola-epidemic-who-west-africa.html

http://news.nationalgeographic.com/news/2014/09/140923-ebola-virus-west-africa-cdc-projections/

http://news.sciencemag.org/africa/2014/09/who-cdc-publish-grim-new-ebola-projections

http://www.inquisitr.com/1579002/ebola-predictions-scientists-see-cases-in-every-major-u-s-city-possible-by-years-end/

Share

2014’s Ebola Prediction Fail

Credit: Graph by Susan E. Swanberg based on CDC and WHO data

Credit: Graph by Susan E. Swanberg based on CDC and WHO data

An unprecedented epidemic

A devastating Ebola epidemic hit the people of West Africa in 2014, killing almost 8,000.

Approximately 20,400 suspected and confirmed Ebola cases have been reported so far in Guinea, Liberia and Sierra Leone,* a far cry from the 1.4 million worst case scenario calculated by a Center for Disease Control and Prevention (CDC) research team.

Whether the CDC’s worst case scenario was in error, intervention by international medical teams altered the trajectory of the epidemic or the press exaggerated the significance of the worst case scenario will likely be debated for a long time.

Dire predictions of an expanded epidemic

The World Health Organization (WHO) was notified of an outbreak of Ebola in March of 2014, and the epidemic grew in strength over the course of eight or nine months. West Africa’s medical community was overwhelmed, and cultural practices, including the care and respect afforded the dead by surviving relatives and friends, contributed to an upsurge in cases.

At some point during the epidemic, accusations that the world was ignoring a potential global catastrophe began to appear in the news.

In July of 2014 Ken Isaacs (Vice President of Programs and Government Relations for the international relief organization, Samaritan’s Purse) wrote a piece in the New York Times asking “Why Are We Ignoring a New Ebola Outbreak?” The Telegraph published a story titled “How the world ignored Africa’s Ebola tragedy.”

Former U.N. head Kofi Annan, quoted in several news reports, claimed that the international community ignored Ebola until the disease traveled to the U.S. and Europe.

The epidemic was already unprecedented, but in the fall of 2014, predictions of a dire nature started to be published. “Ebola could infect 1.4 million in Liberia and Sierra Leone by end of January,” was the headline for a September, 23, 2014 Washington Post story. “Every Second Counts: New Ebola Report Predicts Huge Spike Without Action,” said Maggie Fox of NBC News. “CDC: Ebola cases could reach at least 550,000 by January [2015],” said a CNN headline.

The source for these and similar news articles was a September 26, 2014 report published in the CDC’s Morbidity and Mortality Weekly Report (MMWR). The report, titled “Estimating the Future Number of Cases in the Ebola Epidemic — Liberia and Sierra Leone, 2014-2015″ was authored by nine CDC scientists.

Digital image of the Ebola virus. Credit: Shutterstock http://tinyurl.com/l4lud4d

Digital image of the Ebola virus. Credit: Shutterstock
http://tinyurl.com/l4lud4d

According to the article’s abstract, the CDC developed a modeling tool called “EbolaResponse.” The purpose of the model was “to provide estimates of the potential number of future [Ebola] cases.” The abstract said, “[i]f trends continue without scale-up of effective interventions, by September 30, 2014, Sierra Leone and Liberia will have a total of approximately 8,000 Ebola cases.”

Looking at the actual statistics, this prediction was not far off. By the end of September, Sierra Leone and Liberia actually had about 5,220 Ebola cases. Using a “potential underreporting correction factor of 2.5” the model estimated that approximately 21,000 total cases would have occurred in Liberia and Sierra Leone by the end of September, 2014.

The correction factor was derived from a comparison of the EbolaResponse model’s prediction of how many hospital beds should be occupied (based on cumulative Ebola cases) to the numbers of hospital/clinic beds actually occupied by Ebola patients. The two figures differed by a factor of 2.5.

The CDC researchers plugged into their model numbers from several possible Ebola scenarios, including differences in levels of intervention.

With no increase in intervention, the worst case scenario predicted 1.4 million Ebola cases by January 20, 2015. Increasing the numbers of patients in treatment was the factor most likely to avoid the worst.

The worst case scenario has not come to pass, and is now highly unlikely to occur.

This protective gear would be inadequate for handling Ebola patients because the goggles might not completely cover the skin. Credit: Shutterstock EfFbi4ytORmX

This protective gear would be inadequate for handling Ebola patients because the goggles might not completely cover the skin. Credit: Shutterstock EfFbi4ytORmX

The World Health Organization’s Alternate Prediction

The CDC was not the only health authority making predictions about the course of the Ebola epidemic. On October 16, 2014, the WHO published an article about the trajectory of the epidemic in The New England Journal of Medicine. The article, “Ebola Virus Disease in West Africa — The First 9 Months of the Epidemic and Forward Projections,” made a much more cautious and measured prediction than the CDC.

According to WHO researchers, “at the current rate of increase, assuming no changes in control efforts, the cumulative number of confirmed and probable cases by November 2…will be 5740 in Guinea, 9890 in Liberia and 5000 in Sierra Leone, exceeding 20,000 cases in total.”

This prediction is more in line with the status of the epidemic as of January 2, 2015.

Limitations of the CDC model

In fairness to the scientists who developed the EbolaResponse model, it’s clear that they were transparent about their model and its limitations. The modeling tool is available in a spreadsheet here. The authors described in detail three levels of intervention that were included in several of their scenarios.

In their article, the CDC researchers reported five limitations:

1. Extrapolating future cases from (then) current trends “might not be appropriate,” said the researchers.

2. Assuming incubation and infectiousness periods based upon previous epidemics “might not be accurate.”

3. Reliance on expert opinion regarding bed occupancy might not account sufficiently for other factors relevant to the correction factor.

4. The correction factor might change over time.

5. The predictive scenarios didn’t consider the logistics requirements required to improve intervention levels.

The Role of the Press in the Prediction “Fail”

In spite of the study’s clearly disclosed limitations, and the inclusion of more possible scenarios than the worst case, many news reports emphasized the most extreme of the CDC’s predictive scenarios. (See below for links to articles.)

This is not the first time some science journalists have chosen to focus on one aspect of a study, excluding or overlooking other relevant factors. Going back to the original autism/vaccination reporting and looking at current news reports touting “cures” and “breakthroughs,” it’s clear that we need to do a better job of reporting science.

Looking critically at my own reporting on the CDC’s predictions, I see that I also referenced the worst case scenario. I did, however, qualify my interpretation and provide a detailed analysis of the significance of the prediction. As science writers and journalists, it behooves us all to keep our subjects under a magnifying glass — reading the primary literature thoroughly, skeptically and including as much detail as possible in our analyses.

One item that stands out to me now, is the CDC’s own statement about the accuracy of its predictions. In a Q & A about the predictions, the CDC asks and answers as follows:

How accurate are the estimates (using uncorrected or corrected data)? They are only estimates, and CDC cannot guarantee their accuracy; however, the estimates illustrate that if conditions remain unchanged, the situation will rapidly become much worse. As such, the estimates are a warning and a call to action.”

What role should journalists play in spreading the word about somebody’s call to action?

Was it right to broadcast the CDC’s worst case scenario the way we did? How could we do better in the future? Did hyping the CDC’s most dire Ebola prediction ensure that the public will ignore the calls for help when the next epidemic comes along?

The 2014 Ebola epidemic is not yet over, and we must continue to be vigilant and to help improve health conditions for West Africans and people around the world.

Science writers and journalists can help, too. We must avoid the temptation to hype or overextend the findings of scientists. We must look to the literature, read it carefully and ask questions.

If we fail to be the watchdogs the public needs instead of cheerleaders for those with an agenda, we will lose the trust and faith of our audience — trust and faith that might be needed to avoid the next epidemic.

__________

I encourage my readers to debate these and other questions the CDC’s Ebola predictions raise.

__________

*In addition to the cases reported above, 20 cases (including eight deaths) occurred in Nigeria, eight cases (including seven deaths) occurred in Mali, one case each occurred in Senegal and in Spain and four Ebola patients were diagnosed and/or treated in the U.S.

One of the four U.S. patients died from Ebola.

Note: This piece represents my interpretation of the reporting on the CDC’s predictions. At some point in the future, I might undertake a more scientific article on the issue.

__________

Dire Prediction Reporting

http://www.webmd.com/news/20141023/experts-predict-catastrophic-ebola-epidemic-in-west-africa-if-aid-delayed\

http://www.newscientist.com/article/dn26448-future-scenarios-show-how-easily-ebola-could-explode.html#.VKcw-Isgl94

http://www.nytimes.com/2014/10/15/world/africa/ebola-epidemic-who-west-africa.html

http://news.nationalgeographic.com/news/2014/09/140923-ebola-virus-west-africa-cdc-projections/

http://news.sciencemag.org/africa/2014/09/who-cdc-publish-grim-new-ebola-projections

http://www.inquisitr.com/1579002/ebola-predictions-scientists-see-cases-in-every-major-u-s-city-possible-by-years-end/

Share