Ebola Screenings – Reverend Bayes’ Perspective
In class we discussed the medical diagnosis example for “rare condition x,” so it was interesting to find an approach on-line that brought this classroom example to real-life by tying it all to today’s most talked about rare disease, Ebola.
With all the fear, uncertainty, and buzz about the spread of Ebola, it comes as no surprise that in some countries, screening routines have emerged in airports and other ports of entry. Of course, new procedures always come with a wave of skepticism from the crowds, and Ebola screenings are no exception. Because some arguments are difficult to convey in regular terms, the resource I linked to offers a perspective that suggests that we can utilize Bayes’ Theorem to analyze the skepticism and really understand the screenings.
As we saw in class, four situations can arise when a person is screened:
Bayes’ Theorem can then aid us in understanding the relationship between screen results and Ebola status.
The analysis becomes a bit complicated since the formula requires the probability of getting a positive screen if you do not have Ebola, which is related to the specificity of a test. Unfortunately, the specificity and sensitivity of current Ebola tests are not particularly known yet. The CDC (United States Centers for Disease Control and Prevention) does suggest that real-time PCR methods are used, which allows for an educated guess that the “full laboratory test is about 95% sensitive and 97% specific to Ebola.”
Now let’s look at the patient perspective. The following graphs show how confident we should be in a positive diagnostic test (first graph) and how confident we should be in a negative diagnostic test (second graph):
The graphs above demonstrate the value of testing from the perspective of someone who has or has not tested positive. Now we observe the public health perspective, which is to ask how many people would be in each of the four quadrants if we tested everyone.
Collectively, the charts illustrate that the Ebola tests are pretty good, but if we start using them on everyone, there will be a good amount of false positives. Also, even when the risk of disease is very high, even low rates result in cases being missed, which could have scary and serious consequences.
In addition, the graphs above tell us two things:
- Do not rely on testing alone
- Testing should not be used in very low-risk groups (illusion of protection)
Evidently, the decisions involved in screening are complex. There is a good amount of political pressure to do something, but there is also logic behind some of the ineffective-sounding screening activities, such as questionnaires.
It all comes down to the baseline risk, and that is very difficult to estimate. In the end, it is very apparent that the arguments are legitimate, and that there is no simple answer to it all. What the Bayes’ Theorem perspective does is that it allows us to agree on a structure for the arguments so that, in difficult situations, we focus on the factors that should drive our decisions.
Source:
http://monkeyglandin.com/ebola-screening/