Hi, I'm Les Roberts. We're going to have a little chat about disaster epidemiology and the role of data during complex emergencies. I think I'd like to talk about this first in general terms, then I want to spend a little time talking about the difference between surveillance data and survey data. Then finally, I'd like to give you two tangible examples where both data prove to be very, very useful. We live in a somewhat confusing in funny world. Here's a picture of a child taken in the Central African Republic a few months ago. I saw this child with this bat, and I asked could I take your picture, and he was very proud to show it to me. Then this backup passed around a much a bunch of kids, and one of them was wearing it on his back, and another put it on the head of another kid. Because I was part of the response to the 2014 Ebola outbreak in West Africa, I must confess right now I have a little bit of extreme maybe even irrational discomfort about bats. Children in Central Africa have certainly been playing with bats for hundreds of years, and normally it doesn't lead to Ebola. But for me somehow when I see a bat, I'm suddenly afraid of Ebola, and other such fears are running wild in our world. As we look at couple of newspaper articles to the right, when Ebola struck in 2014, oh my gosh, did the rest of the world become terrified that this plague was going to spread through their country even though there was more or less no chance of that. Likewise at this moment in time in my country, the United States, a lot of people are very terrified that immigrants coming across our Southern border are terrorists when in fact, there's absolutely no evidence of that at all. So it's more important now than ever that we have data to help us curb the most fear laden of human tendencies. In front of you is a graphic, that's showing the changes in infant mortality from 1950 through about 2050. It's breaking it down in the middle, the line is in blue, and is for the entire world, and then for other regions of the world. You will note that over this window of time, infant mortality has dropped by a lot. If you look at the middle line in blue with squares that represents the world, from 1950-2000, we've had something like a 60 percent drop in global infant mortality. We have some bad data from Europe in the 1400 and 1500, we've got some bad data from some missionaries in Congo in maybe 1900. Our best guess is probably back before modern medicine was available, about a third of children who were born died in their first year of life. What that means, that we made as much progress from 1950-2000, as our species has made in the 5,000 years that came before probably. Wow, was that an astonishing generation that came before us. So there are lots of good trends in the world regarding health. As we look over that generation that came before us, let's say looking at the United States from 1900 to the end of the 1900s, as this data from the US Centers for Disease Control shows, the causes of death in 1900 were primarily infectious in nature. The big three killers were pneumonia, tuberculosis diarrhea, and other intestinal disorders. In 1997, on the right, we see that the main causes of death were heart disease, cancer, and stroke. So we've moved both in the United States and across the globe from having death profiles driven by infectious agents, to death profiles driven by more chronic or non-infectious agents. But as we look at each of those infectious diseases, there are a host of things happened that made the dramatic reductions we now see. For diarrheal deaths, the reductions have been particularly traumatic, but this has happened with measles and other conditions as well. Each of these improvements is some combination of better evidence is driven by science, plus better social services as organized within a health system, and within a community's ability to pay for and respond and interact with that health system. I am delighted to say that among the various problems that we seem to be getting better at controlling, natural disasters and in particular humanitarian crisis triggered by conflicts are actually part of what we're getting better and better at dealing with. Here in front of you is a graphic, and it's searching through this big database that is based in Brussels, it's called the CRED database. It stopped working a few years ago. So after that, I've been using Google searches. It's showing the highest mortality rate measured in any emergency anywhere in the world, where at least there were a 100,000 people affected, and we're the emergency lasted for at least one month. So if there was a tsunami and earthquake and a lot of people are killed in one day, that's not included here. These are crises that lasted for at least a month. You will notice that back in the 1980s and '90s, we would have 20 deaths per 1,000 people per month, which is two percent of a population dying per month, happening every couple of years. Now, we haven't seen that in two decades. The world has gotten really good at responding to humanitarian crises. As a result, we don't have nearly as many deaths from measles and deaths from diarrhea and death from respiratory infections, and that is the product in part of better and better evidence guiding what we should do. If I look in that CRED database, in this case, at the five most high profile continuous humanitarian crises over this 15 year period of time, Sudan, DRC, Somalia, Afghanistan, and the Central Africa Republic. Here are the number of mortality measures that were undertaken over time in those various crises. All I want you to take away from this, is that we are measuring mortality more and more over time and we are finding fewer and fewer high mortality events as you saw on the last slide. Thus this trend of the world seems to be getting better and better at keeping the negative effects of humanitarian emergencies under control, this trend is not just an artifact that we're not looking anymore. It also happens that we are spending more and more on humanitarian responses over time. The increased humanitarian spending is being spread over many different sectors, and that's probably appropriate and good. So here's a picture. It was taken at the peak of the war in the Eastern Democratic Republic of the Congo back around 2000. This little girl was sitting out in front of her huts and I and a bunch of my Congolese friends were walking by. I saw her and I asked could I go look at this girl and walked over, and I grabbed the skin on her arm and pulled it up and twisted it. It tented up and didn't spring back which is the classic indicator that this girl is very dehydrated. She's also malnourished. Her father said that she was quite sick and that she had diarrhea, but she just didn't want to drink water. So I took out my water bottle and I held it up and offered it to her and she started drinking, drinking, drinking. Oh my gosh, she drank 200 milliliters of water and then threw up. I said to the father please, please, please, won't you or won't your wife take her to the clinic and he said that his wife was dead. Please won't you take her to the clinic? He said, "Yes, I will." So we were doing a survey. We went on and hours later, we came back and the little girl is still sitting outside of her hut. So my colleague went and found the father working in the field. I said, "Please, why haven't you taken this girl for treatment? She's at risk of death." The father said," Well, I don't have the money," and I said, "Well, I will give you the money." He said, "Well, it turns out the clinic here I asked my neighbors say that the clinic here isn't working because there's no drugs. It's just not working anymore." So here's a girl. She's in a rebel held area. Her mother has died. Is she going to die if she were to die? If she dying because her mother isn't there? Is she dying because her health system has collapsed and the local clinic is no longer serviced by the government controlled areas because it's in the rebel held areas? Is it because she has diarrhea? I can't answer that. I can only say that almost all excess deaths is coming from some complicated interwoven, a combination of economic factors, social factors, psychological factors. In this case there's a pathogen inside her, microbiological factors, and we need programmatic responses to deal with those. Thus every crisis is different, and thus the need for data to tailor our, in this case, Oral Rehydration Therapy strategy. The data we need to understand what's going on is going to be different in each crisis. In mind, perhaps simplistic mind, I break data collected in emergencies into broadly three categories: surveillance data, survey data, and assessments which normally are something close to surveys but maybe don't have as much structure or any structure. We're going to spend much of our next few minutes together talking about surveillance and surveys in particular because they both have great value, and they both can highlight the various things that health information can do for us in crisis. Here is an epidemic curve of an outbreak of pellagra. Pellagra is a micro-nutrient deficiency caused by not eating enough niacin. More or less, it only arises in populations who are eating almost exclusively maize meal. If people are eating peanuts, this doesn't happen. If people are eating meat, this doesn't happen. If people are eating rice, generally pellagra will not happen. It seems to be associated with corn based diets. In this particular graph, we can see that among Mozambican refugees in Malawi, in early 1990, when the World Food Program took peanuts out of the ration, there started to be pellagra cases. At that time, there had never been in a humanitarian crisis a pellagra outbreak detected. People in the World Food Program were very skeptical that such a condition existed in a way that would be meaningful and important on the large-scale public health sense. MSF to their credit move from having passive surveillance of people coming into their clinics to having community health workers go out and look for these people with pellagra. Pellagra causes a rash around the neck. It causes diarrhea, it causes dementia. So they went around looking for people with this rash and there's one month, September of 1990, where 10,000 cases of pellagra were recorded with hundreds and hundreds of deaths. This data was so compelling that it convinced the World Food Program to put a source of niacin back into the diet and within couple of months the problem disappeared. Here was a problem that within the World Food Program essentially people didn't believe existed. MSF showed it existed, changed policy and stopped it in really short order. Wow, that's pretty powerful use of data. There are other times when data tell us about something that we thought was going on but we weren't either completely sure or able to articulate. For example, there were some remote satellite pictures taken inside Myanmar in late 2017 suggesting that widespread attacks were happening on Rohingya villages. Then a couple months later, MSF did a survey of the Rohingya villagers who had fled out into Bangladesh and showed that almost certainly over 6,000 people had been violently killed in attacks on the Rohingya. So here's an example of MSF that is, Medecins Sans Frontieres, a large international medical NGO telling the world what they suspected but that they didn't really have the confidence to yell out loud until they documented it well. Sometimes data emergency tells us what to do. Here we are looking at a pie chart showing the causes of death among Kurdish refugees who have fled Northern Iraq, tried to flee into Turkey but Turkey has built a fence and is now keeping them from entering the country of Turkey. They had risen up against Saddam Hussein when the Americans and their allies had invaded in 1991 and after those invading forces withdrew, Saddam really put enormous political pressure and exerted enormous violence on these Kurds, and now they have fled their homes. You will see in this graphic that 75 percent of all deaths were caused by diarrhea and dehydration. So in response to this, the various organizations that were working started creating buckets of Oral Rehydration Solution and going out within these camps from tent to tent trying to find people who had diarrhea and get them to drink ORS. In this particular pie chart, it is so clear that the primary problem is diarrhea that allowed the relief community to essentially focus on that one problem to an extent they rarely do. Sometimes data and emergencies can just provide hope. So here are the three epidemic curves presented by WHO for the countries of Guinea, Liberia, and Sierra Leone that were most affected by the 2013 through '15 Ebola outbreak in West Africa. I went out in September of 2014 to work in Sierra Leone. I must confess it was really scary. We weren't sure what was going on here, and if that disease was just going to spread across Africa because we hadn't managed to contain it. When in October we could see that the numbers of cases in Liberia were steadily going down, and knowing that the Liberian program was roughly doing the same things we were doing in Sierra Leone, was incredibly encouraging to us. So when you're out in a crisis and thinking about perhaps your need to go out and collect data or information, I think there are some basic decisions you need to make. First of all you need to ask yourself, does this information already exists? Maybe it exists in a different form but maybe it already exists. Is the information that's out there about cases, about strain of pathogen that's operating, about the motives by which these people are or are not coming for healthcare? Is the information of adequate quality for you to undertake the decisions you need to? If I'm going to go out and collect new information you should probably ask yourself what exactly am I going to do with it? Data for advocacy has a completely different set of parameters and characteristics than data to guide day-to-day programmatic activities. You also need to understand what is the opportunity cost going to be. Opportunity costs not only being what you're not doing with your time and what you're not doing what those vehicles and with that $5,000 you're going to spend this weekend to go out and visit four villages, but also each time you go to a village and you ask them about some particular issue, "Oh have you been burying your body safely during this Ebola outbreak?" Each time you go and you do that, you might raise expectations within the village. If nothing comes of it, are they going to be disappointed? Opportunity costs aren't just dollar costs. They are social costs as well.
