Settling the controversy on Haiti’s cholera epidemic

While Japan’s recent earthquake and tsunami have appropriately captured the world’s attention, an earthquake that took over a year ago has continued to produce devastation in another hemisphere. Haiti’s earthquake of January 2010 shattered the country’s already-weak water and sewage infrastructure, and in October 2010, cholera (a bacterial illness spread by contaminated water) was reported in the country for the first time in more than 100 years. Within weeks, the disease had been identified in every one of Haiti’s provinces; by the end of the year, more than 150,000 cases and 3,500 deaths had been reported. Although the cholera epidemic has fallen from public view, the epidemic continues to rage on.

There’s been increasing controversy about what to do in response to the persisting epidemic. Should we distribute vaccines, even though the epidemic has already started? Are antibiotics useful, even though the classical medical textbooks say that hydration alone should be sufficient to avert death? Are clean water provisions sufficient? And, of course, how much money do we need for all of this?

Early on, the United Nations estimated that $160 million was required to control the  epidemic, based on a projection that 200,000 people would be affected by the disease. Where did these numbers come from? When we investigated the source, here’s what we found in the UN’s report: “assuming all of the population (estimated at about 10 million for the purpose of this plan) is at risk of contracting cholera, and estimating a cholera attack rate of 2% (not a conservative estimate, given the prevalence of risk factors for cholera transmission including lack of safe water supply, poor sanitation conditions and the rainy season), the estimated number of cases would be 200,000 (10,000,000 population × 2%)”.

Interestingly, there is no empirical basis for this 2% guess. Even more worrisome, the estimate ignores basic aspects of the disease like how the bacteria are transmitted, how people can be asymptomatic carriers, can acquire immunity, or can receive vaccination and treatment. Two weeks after the publication of this estimate, the projection was doubled, without explanation, to 400,000 cases—a number widely cited by the press and used by aid agencies to allocate resources.

In today’s edition of the journal Lancet, we re-estimated Haiti’s cholera burden using a mathematical model that accounts for existing data from every province, adjusts for underreporting of cases and deaths, accounts for asymptomatic carriers, and projects the impact of antibiotics, vaccines, and clean water provision. Here’s what we found:

[1]  The United Nations projections are likely to be vast, vast underestimates of the epidemic’s true toll. Using complex fitting procedures to calibrate our model against time-series data from Haiti, we found that it’s more likely for cholera to cause nearly 800,000 cases of the disease before the end of this year. That means we’ve likely severely under-budgeted for this epidemic.

[2] Antibiotics can have a substantial impact on the epidemic, because they dramatically reduce the infectiousness of people infected with cholera, thereby preventing downstream infections. If drug resistance develops (which is of very low probability compared with other bugs), alternative generic medications are available for treatment that have a similar impact to the first-line antibiotic. In the past, antibiotics have only been reserved for the most severe cases, but experienced physicians have called for wider use, and our model agrees with them.

[3] Vaccination can make a big difference, even after the epidemic has already started. Even vaccinating only one out of every ten people could avert almost a thousand deaths by the end of this year. That implies a big change from current World Health Organization (WHO) policy, which is only to use vaccines before an epidemic has started.

[4] Rather than quibbling over whether water or antibiotics or vaccines should get the most money, we found that all three interventions had relatively the same impact in isolation, but when implemented together, they would be expected to avert several thousand deaths before the year end. This is also consistent with what physicians on the ground have been calling for.

[5] A decline in cholera cases in early 2011 is due to the slow accumulation of immunity to the disease, and reflects the natural course of the epidemic. Hence, a decline in cases around this time should not be interpreted as indicative of the success of human interventions, unlike what some media outlets have stated. Instead, we find that it’s likely for cholera to persist in Haiti for at least another year, if not become endemic, even though immunity is slowly accumulating and this naturally produces a “long tail” end of the epidemic curve that is going to last for a very long time.

As with all mathematical models, the results of our model are based on assumptions inherent to mathematical modeling. But other independent groups of scientists have come up with nearly the same results when using a totally different kind of model.

The bottom line? Instead of making random guesses, we should try to account for the basic features of a disease when making our projections. In the case of Haiti’s cholera epidemic, we really need to buff up our budgets and plan for a prolonged epidemic that is likely to far exceed initial expectations.


Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s