In this section, we're going to talk about surveillance as an observational process. So a public health surveillance system is a system that's designed to uncover the truth, if you will. Let's get a little philosophical about what public health surveillance is trying to do. Within any given population, disease exists. The disease that you're interested in exists out in the world. Our ability to see or detect that disease depends on our surveillance process. Things like the case definition we use, the methods we use to identify those cases. Outbreaks can go unnoticed, or unreported, even if they still occur, right? So what's going on out in the world is the truth. And our surveillance process has to be designed to maximize our ability to see that truth if we want to be successful. Our understanding of a public health problem is only as good as this public health surveillance process. And I want to give you an example in this section about how that process can be suboptimal, particularly as it relates to access to care. In this slide, we see the population, and we see a health care facility, which is a common place where we conduct public health surveillance. Within this population we have some people who have disease. You see them here in green boxes. And what we typically assume is that people with disease are going to do what? They are going to come to our health facility and seek care. And then we assume, we hope, that we have a case definition that will find all these people. And so our surveillance will detect all of the cases of this disease within the population. But it often doesn't work that way, does it? So instead, what we may have is something that looks like this. Some people seek care and are captured by our observational process, our case definition. We're going to capture, as we discussed previously, some people who really do not have true disease. And some people who have disease may not seek care at all at our health facility. They may stay in the population. Why might that be? Maybe they can't afford care. Maybe they're not very sick at all. Maybe they don't think they need any healthcare. Or maybe they didn't make it to the healthcare facility. Perhaps they died before they got there, so we never saw them. It's just important to remember that case definitions are operating within your healthcare facility, typically. So you rely on patients to at least seek care before you can capture them with that case definition. And often, you don't even have the chance to do that, if patients do not seek care. And this is a very common phenomena in many parts of the world, where people have inadequate access to medical care. This is a limitation of using facility-based surveillance in places where people don't have the kind of appropriate access to care. I want to give you a further example from some of my work on meningoencephalitis surveillance in Bangladesh, and how access to care can influence our observations of disease. So if you look at this figure, along the x-axis on the bottom, it shows the distance to hospital in kilometers. It runs from 0 to 120. On the y axis, it shows the cumulative number of encephalitis cases, from 0 up to 400. And there are two colors on this map. You can see the green, which shows the total number of cases, and the purple, which is the observed. Let me tell you how we got the data for this figure. So the observed represents the number of encephalitis cases who sought care at our surveillance hospitals, and were detected. And where they live with respect to the hospital. So if you see the purple shaded area, you can see that most of our cases live within, let's say 30 kilometers of the surveillance hospital. Some live farther away, but not that many. People really didn't seek care greater than 30 kilometers distance from the hospital. And when we went out into communities, we did surveys, all in this area surrounding the hospitals. And we asked people about signs and symptoms of encephalitis. Now, lay persons can report this. It includes fever and coma. And most people can remember if they, or one of their family members, recently suffered from coma, and unconsciousness. So when we went out and we counted all the people who actually had this disease, we saw that there are many more cases occurring in the community than those that came to our surveillance hospital. And the cases that we missed, or who didn't come to surveillance hospital, are represented in green. So there's no reason for us to think that people who live farther away from hospital have a lower incidence of encephalitis. And we saw here that they didn't. But what is true is that people who live farther away were much less likely to come to the hospital, and to be observed by our surveillance system. So this was an important reminder for us to always keep in mind that our surveillance is an observational process. It gives us an indicator of what's happening in the population, but it rarely tells us the whole truth.