Autism, Ebola, and the Will Rogers Effect

Dr. Coplan reflects on the history of autism, as it may relate to the Ebola epidemic.

I’m going to take a risk this week, and put in my 2 cents on a topic that’s outside of my primary area of expertise – Ebola Virus. What can a Neurodevelopmental Pediatrician contribute to this conversation, and what do autism and Will Rogers have to do with it? Read on….

One of the terms you hear in connection with Ebola Virus (EV) is “Case Fatality Ratio” (CFR). CFR is defined as the number of infected persons who die divided by the total number of persons who are infected. (CFR = Deaths / Total Infected). This ratio is murky for various reasons.

The numerator is the easier value to discover: Death from Ebola is dramatic, and hard to miss. Even the numerator, however, is never 100% accurate: If you are living in the bush, and there are no healthcare workers to document your demise, your death may go uncounted. Or, in non-epidemic situations, an isolated death here or there, even in a medical setting, may go undiagnosed. But in epidemic settings, especially when healthcare workers are on the scene, the numerator can be tallied within a tolerable level of error. 

The denominator is another story altogether. This is where my work in autism enters the picture: In the 1940’s and 50’s, autism was defined as a profound disorder leading to continuous, bizarre behavior and near-total social isolation. Not surprisingly, that definition yielded prevalence figures for autism in the range of 2 to 4 cases per 10,000 children, and the prognosis for these children was correspondingly grim. Over the ensuing 5 decades, however, we gained a better appreciation of the spectrum of the disorder. As we got better at recognizing milder cases of autism, two things happened simultaneously: The reported prevalence of the disorder went up 100 fold (now it’s around 1 in 50 children), and the prognosis improved dramatically – not through any improvement in treatment, but simply because the total pool of children with identified autism contained a larger proportion of milder cases. (Reviewed here https://www.drcoplan.com/feature-publication )

This is where Will Rogers enters the picture: Rogers quipped “When the Oakies (residents of Oklahoma who departed for California during the Great Depression) moved from Oklahoma to California, they simultaneously raised the IQ of both states 20 points.” This is possible if the IQ of the average Oklahoman (excluding the Oakies) was, say, 100, while the IQ of the average Oakie was 80, and the IQ of the average native Californian was 60. The IQ of the Oklahomans who remain in Oklahoma rises when the Oakies leave. Likewise, the IQ of the average Californian rises with the influx of the new settlers (80 being greater than 60). In an insightful paper on the Will Rogers effect in cancer staging, [1] the authors suggest that prolonged survival with cancer is partly due to improved recognition of disseminated but asymptomatic disease at the outset, rather than improved treatment. The same is probably true in autism: inclusion of larger numbers of milder cases leads to improved mean outcome, irrespective of therapy.

All of which leads me back to Ebola.

How big is the denominator (all persons who become infected)? No one has a clue, actually. In an epidemic, the few people with recognized disease who don’t die go into the denominator. But how many additional people are there who develop disease, go unrecognized, and survive? An intriguing paper by a team of researchers in Gabon [2] (and here http://www.virology.ws/2012/01/18/how-lethal-is-ebolavirus/ ) found that fifteen percent of four thousand asymptomatic villagers carried antibodies to Ebola virus, from which they conclude “The high frequency of ‘immune’ individuals with no disease or outbreak history raises questions as to the real pathogenicity of ZEBOV for humans in ‘natural’ conditions.” [ZEBOV = Ebola Zaire]

In other words, researchers studying Ebola may be at the same stage of the game as were researchers studying autism in 1950: Just as the prevalence and prognosis of autism were spectacularly incorrect due to an overly-narrow case definition, it may be that present estimates of Case Fatality Rate are based on a denominator comprising only a tiny sliver of the total number of people who become infected.

Data that would fit this hypothesis (although they do not constitute “proof”) also come from the fact that thus far, none of the secondary cases among Americans has proven fatal. (This will change, of course. Ebola is still a terrible disease.) Patients in top-flight American hospitals get the best possible supportive care, in some cases augmented by experimental antiviral therapy. But these measures may not be sufficient to explain improved survival among these individuals. Rather, it’s possible (“likely” in my view) that vigorous twice-daily monitoring of temperature in at-risk persons is detecting more cases, including a higher proportion of milder cases, than just tallying up people with vomiting and bleeding who show up at a field station or hospital ER.

Is this a good thing or a bad thing?

If you are the person who has contracted Ebola, it’s definitely a good thing: Your chances of survival may be greater than the survival rates based on current estimates of CFR, with their biased ascertainment of both the numerator (the sickest of the sick) and denominator (almost devoid of infected but asymptomatic or mildly symptomatic individuals). On the other hand, if you are a public health official, this should cause concern. The backbone of controlling an epidemic is contact tracing and quarantine. Contact tracing rests on the unspoken assumption that it will be easy to recognize the index case by his or her symptoms: Vomiting, bleeding, high fever, etc. If this is untrue (and I suspect it is), then we may need to revise our thinking. Granted, contagiousness is a function of viral load, and persons who are asymptomatic are probably low-risk as disease vectors. (The risk of dying, and the risk of transmitting the disease to others, seem to go hand-in-hand.) But the prospect of a large pool of asymptomatic individuals positive for Ebola exposure is a potential game-changer. It took the medical community half a century to correct the error of its thinking insofar as the spectrum of severity of autism is concerned. With Ebola, we don’t have that kind of time.

1. Feinstein, A.R., D.M. Sosin, and C.K. Wells, The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med, 1985. 312(25): p. 1604-8.

2. Becquart, P., et al., High prevalence of both humoral and cellular immunity to Zaire ebolavirus among rural populations in Gabon. PLoS ONE, 2010. 5(2): p. e9126.

James Coplan, MD is an Internationally recognized clinician, author, and public speaker in the fields of early child development, early language development and autistic spectrum disorders. Stay connected, join Dr. Coplan on Facebook and Twitter.