Sunday, August 10, 2014

Why are doctors and nurses dying?

When dealing with Ebola, doctors and nurses in the city wear protective gear. Yet they fall sick too. Why? Can’t the layers of rubber or vinyl protect them?

Yes, but…

The first problem is diagnosis. Not everybody showing up has ebola, and you don’t ship people into the isolation ward for triage. That first step can be tricky.

Think about the body’s natural defense system. Never mind the inner ones for now. Most of the body is covered with skin, which isn’t very permeable to start with and which is covered with an armor of scales of dead skin cells. There are perforations—hair follicles, sweat glands, oil glands, and the occasional nick or scrape. The mucous membranes are much more vulnerable, but they cover a much smaller area and produce secretions that at least potentially can sweep out some of the detritus. Even the lungs, which have to be permeable to gas molecules, are only approached by channels lined with cilia and mucus to trap and eject foreign particles.

Any of these systems can be overwhelmed. Smokers overwhelm their lungs’ defenses with persistent smoke (not to mention the gases) and as we all know too well that viruses for cold and flu frequently succeed in arriving in our membranes and starting to work.

In addition, the systems can be compromised. In particular, one essential practice—hand washing—can reduce your dead skin layer and the more aggressive cleansers, accompanied by the scrubbing requited to clean out folds and pores, can inflame the skin over the course of a day—meaning, among other things, that the skin is more permeable.

I almost hesitate to mention this, because not washing is a terrible option. The health care workers are between a rock and a hard place. If they don’t wash, they risk infecting themselves and other people; if they do, they may still infect themselves. Wash. Scrub. Damn the torpedoes.

The calculus differs for somebody taking care of his cousin back in the village. There’s only one person to care for and then no more (one hopes), but any protective gear is entirely ad hoc (cloth dipped in bleach, perhaps?) Because the number of times he needs to wash is smaller, harsher chemicals are OK—in fact probably required since his gear will leak.

It differs again for someone just going about his business in the city. Unfortunately how much it differs depends on the local risk, and I don’t know the details—as fast as I can tell nobody does.

Look at that protective gear again. It should block fluids very effectively. But there’s a reason most good ideas never make it to commercial production: corner cases are common, and so are Murphy Conspiracies.

Look at the suits.

  1. They are stored folded. Are the flexed folds as strong as the rest?
  2. The suits are kept in high heat and humidity. How fast will that deteriorate the material?
  3. Africa is home to many molds and mildews. Which infect the materials?
  4. What will chlorine bleach do to the materials?

Any one of these insults is probably within engineering tolerances. But with 2 or 3 in combination, would you guarantee the integrity of the material?

Most of the time all is well. But a nurse makes dozens of contacts with patients in a day (cleaning the up is no small job!). The probability that her suit or glove leaks is small, and the probability that an exposed bit of skin will serve to infect is not 100% (but with everybody sweating so much I’d not bet the exposed area would stay small); call their product ε. In a simple statistical model the probability that she will be infected is 1-(1-ε)^N where N is the number of contacts. Suppose she has 50 contacts in a day, and each probability that something seeps through to infect her is only 0.1%. The chance that a single day’s work will infect her is about 4.9%. (Beats 100% without protection, though) Numbers are pulled out of the air for example's sake. DO NOT QUOTE THEM!

For comparison, suppose somebody has only one patient to deal with (5 contacts a day) but has cruddy gear with a 1% failure rate. The daily infection probability is the same. Since the patient will probably die in a day or two, the chance of getting infected himself is about 10%. On the other hand the nurse, working day after day after day, will almost certainly get it sooner or later.

Of course, back in the village, our caregiver probably doesn’t know how to disinfect, and was probably exposed already when the patient first got sick; so it isn’t a fair comparison—he’ll almost certainly catch it also. Best bet is to have the village scatter for a week or three and nobody be allowed into any other village. That’s easier to do in the tropics than in climates with winter.

In the “slums” of the city, everybody is cheek by jowl. Also, malaria is endemic, so feeling sick is nothing strange (more people have malaria than ebola—and quite a few die from it too), so rubbing up against somebody sick is likely to be an everyday event. Plenty of contacts can happen before the disease becomes dramatic.

I’m assuming that transmission is by contact. That isn’t guaranteed. There are several claims made about ebola that we need to revisit.

The first is that a person is not contagious before symptoms appear. How do we know that? Skin swabs? We need to know that a person has been infected, and be able to search the samples for presence of the virus. If monkeys could be infected you could try to infect them with human sweat taken at various times of the incubation period, but that’s a lot of monkeys to test if you want to be sure there’s less than even a 10% chance of infection. DNA testing of the swabs? It would be very nice if this was true, but there's not always a hard and fast boundary between symptoms and no symptoms.

Another is that it is not airborne. At the simplest level it certainly is: if a patient sneezes up your nose you’re hosed. So how far can the droplets go in tropical humidity without drying out enough to kill the virus? How long does the virus last when dried out? I assume it deteriorates over time, but it can survive for a couple of days on surfaces. And "In the laboratory, infection through small-particle aerosols has been demonstrated in primates, and airborne spread among humans is strongly suspected, although it has not yet been conclusively demonstrated"

I assume that they spray bug killer all over the hospitals and environs, but flies don’t die instantly. The hospitals try to bag up corpses quickly, but there’s a lot of mess to clean up still. The surface of the mess is probably covered with sprayed bleach, so a fly would probably have to swim to pick up live viruses, but there’s still a window of opportunity for fly-borne transmission—or is there? Most diseases haven’t figured out how to use mosquitoes for transmission yet, so that’s probably not an important vector—but it might be wise to check.

At this point quarantine is probably the best way to handle outbreaks, but what kind of quarantine may depend a bit on the answers to the questions I brought up. I have no good idea of how to do quarantine in one of the high-density slums.

I'm not an ebola researcher, and my knowledge of Liberian isolation wards is based on descriptions I've read. I assume the dedication to the protocols is being kept high.


Texan99 said...

I've been reading everything I can find. You might enjoy "Spillover," which is available at Amazon for download (you don't even need a Kindle any more! Just download to your computer). It's clearly written, well-reviewed, and reasonably current (2012). It's about zoonoses in general, not just Ebola.

I think the authorities say Ebola isn't airborne, or isn't transmissible before symptoms appear, not because they think they can be 100% certain but because these conclusions are borne out statistically. If this think were airborne and infectious before symptoms appeared, there wouldn't be 1771 cases and 961 deaths, there'd be millions. If you expose yourself often enough, the virus gets a chance to beat the odds, so no individual is ever safe. Nevertheless, as a broad pattern in a population, there are activities that are pretty darn safe, like sitting next to asymptomatic people on a bus, and then there are others, like handling the dead or eating a monkey found dead in the forest, that are crazy risky.

It's interesting that popular culture in Africa periodically erupts in paranoia directed at hospitals and all medical workers, because ironically they do make the problem worse, quite often. Ebola is relatively hard to spread, but if you concentrate victims and then have a small group of people tend a number of them in the lax conditions that apparently prevail even at the "good" hospitals in Africa, you're going to spread the disease.

When SARS broke out in Singapore, Beijing, and Toronto, there was an initial spate of heavy casualties among healthcare workers and some alarming rates of spread. (SARS is a lot easier to transmit than Ebola.) But the outbreaks occurred in areas with fairly strong governmental and private institutions, a good bit of money, and pretty high-functioning hospitals once everyone got seriously alarmed. The author of Spillover wondered in 2012 what might happen if the outbreak had occurred in a densely populated chaotic city instead.

james said...

Absolutely if it were airborne before symptoms appeared there'd have been hell to pay already. I was thinking of different modes: the asymptomatic man walking down a narrow alley, and the patient coughing in a (necessarily) closed room.
One source was claiming this version was more easily transmissible than ones from the Congo area. I don't know on what basis they make that claim, but that could be easily achieved if the initial symptoms were mild enough to go unnoticed and the incubation period was on the low end.