Ebola - Some Hope for Control in West Africa

Ebola, as we all know, is out of control. As an example, every time I turn on the TV there is Dan or Mike or Dan staring back at me. Eventually, we will calm down in the US and begin to focus our attention on the critical outbreak in West Africa. My prediction is that this will happen sometime soon after November 4th (Election Tuesday). In the meantime, there is some possibly, maybe, hopefully good news out of West Africa in today's NYT. As of a few days ago, fewer than half of the 649 available treatment beds in Liberia were occupied. Of course this could be good or bad, but I'm holding out for good.

There was also some potentially good news in a report published in the Annals yesterday. Dan Yamin et al. analyzed a stochastic model of Ebola transmission populated with parameters from a 2000-2001 Uganda outbreak and the current outbreak in Montserrado County Liberia. The authors used the model to determine the number of secondary cases infected by survivors or non-survivors and also evaluated the effect of isolating/hospitalizing patients. I have included the key figures from the paper below. In Figure 1a, they estimate the Ro stratified by whether the index case was a survivor or non-survivor. For the whole cohort, the Ro was 1.73. However, the difference between non-survivors and survivors is striking. It appears that non-survivors infect four times as many people as survivors (2.36 vs 0.66). This may explain why the two Dallas nurses were infected after being exposed to a non-survivor while no secondary cases have yet occurred in other US hospitals, where everyone else (so far) survived.

Figure 1a

In Figure 1c, the authors provide an estimate of the average number of secondary cases per day of symptomatic disease. You can see that there is very little transmission in the community at day 1 and it remains very low for survivors but jumps up after day 2 for non-survivors. This implies that waiting for symptom development is a scientifically valid strategy for preventing community transmission of Ebola even in Africa. (We expect these numbers to be far lower in the US where our communities are less crowded and we are fortunate to have toilets, indoor plumbing and clean water.)

Figure 1c

Finally, in Figure 2 the authors evaluated at what time point non-survivors (very sick individuals) must be actively isolated to prevent community transmission. They estimate that if 75% of the non-surviving cases are detected and isolated by day 4 this results in a 74% chance of disease elimination and if 100% are detected and isolated by day 4 then there is a 94% chance of disease elimination. Currently, the authors report that the average time from disease onset to hospitalization in Liberia is 5 days, so there is some room for improvement. However, I suspect that the current expanded efforts could achieve 4 days. When I put the results of this Annals paper together with the NY Times report of empty beds, it suggests that there is available capacity to hospitalize and isolate patients within 4 days of symptom onset and it might even suggest that current efforts are already working. I'm certainly hoping this is the case.

Oh, and if isolating patients 4 days after symptom onset works in West Africa, it means WE DON'T NEED TO QUARANTINE ASYMPTOMATIC FOLKS IN THE US. So please stop it...and sorry for shouting.