Now we've learned a lot so far in this course. And I want to do this case study. We're going to take an in-depth look, rather in-depth look, at a journal article based purely on the knowledge that we've gained so far. Now we have not looked at deep statistical analysis. Yet, with the knowledge that we have so far, I want to show you what a good piece of research can be put together purely on this knowledge. Now what is this? We looked at research types, whether a piece of research was an observational study or clinical trial. We looked at the two types of classification of data types, whether they be continuous or discrete variables. Whether we are looking at numerical or categorical data. We looked at sampling, how to randomly select, ways of randomly selecting subjects to be entered in a study. And we've looked at some basic descriptive statistics. So purely based on that knowledge, a very powerful piece of research can be put together. So this is the article that I want to take a look at. An audit of trauma-related mortality in a provincial capital in South Africa. From 2014 in the South African Journal of Surgery. So what are we going to have a look at? Certainly, most importantly, what was the research question? What did the authors want to achieve with this research? What study type did they decide on so as to answer this research question? Where did they get the samples from? What subjects were included in this study? What data types did they need to collect to answer, to use analysis to answer this research question. What analysis was done, and then what was the implication, of this piece of research? So let's start with the research question. Now, the authors really wanted to quantify trauma-related mortality in a provincial capital in a developing country. Certainly trauma or trauma-related deaths can be described as a malignant epidemic. And if we look at public health, really one of the first responses to an epidemic is to have a registry to track the disease. And that is how you're going to put together solutions to try and contain this disease, to try and manage this disease. So, from the article directly, this was an observational research designed to provide a comprehensive overview of forensic trauma-related mortality over a two year period. Authors wanted to know why, how many and why people die from trauma in this provincial capital. So what study type would this be? Now there's a few nuances, we can look at it from slightly different angles. Certainly, we are dealing with an epidemic. And we want to know the numbers associated with that epidemic. Now the authors call this paper retrospective audit. In other words, a case series. So anyone is included who died from this disease called trauma. Some trauma was inflicted and unfortunately the patient died. We are discussing an epidemic of those so in some aspects we can also look at this study as being cross-sectional. We purely want to know the numbers of this disease that we are attacking, so you can see that we can look at it from two different points of view, as far as the study type is concerned. Now if you read the article and you think about the sampling, the authors really wanted to be comprehensive. They wanted everyone in the capital who died from trauma-related complications really to be included in this trial. So, they did not take a master list and randomly select from it. They wanted everyone included over the two year period. So, they identified all of the hospitals and the clinics. And even the government mortuary, because remember, unfortunately, some patients would die before even reaching the hospital. So, they even included numbers from the local government mortuary. So that they could have a comprehensive all-inclusive list. So this is actually the master list of trauma-related deaths that they are dealing with. So, what data types? Let's have a look at some of the variables that they wanted to collect. First of all, is gender. Whether someone was male or female. So, remember, although we can count, there's a number that we can count so many males and so many females, that the data type we're talking about here is categorical. The variable is gender someone's either in male classification, either only male or female. And you cannot put one in front of the other. There is no order to male or female. And as such those are two words, descriptive words. Even though you can count how many males there are and how many females there are, this is not numerical data. Data type of the variable gender, being male or female, is nominal categorical. And as we run down the list you'll see, everything they collected is basically nominal categorical. Let's go through them. The mechanism of injury, first of all are we talking about blunt or penetrating trauma? Now you can not say one is worse than the other, one comes before the other, one is ranked higher than the other. Certainly patients die from both blunt and penetrating trauma. You can count how many patients die from penetrating trauma, and how many die from blunt trauma. But the variable itself mechanism of injury, once again is a categorical data type and nominal as such. First of all, they looked at all the trauma at admissions, and just expressed as a percentage how many patients died. So mortality is either, patients survived or they did not. So, it's a yes or no. Once again, that is nominal categorical, data type. Just mortality being yes or no. And lastly, slightly interesting, can look at it also from a slightly different point of view, the place of death. Now if you look for instance at the governmental hospitals, they do mention a small community hospital, a larger community hospital, and then a regional academic hospital. So certainly if you look at the levels of trauma unit that there was, even though this is nominal categorical data, certainly as far as those three hospitals are concerned, there is some order to that in as much as the level of trauma care that they can provide. The trauma unit level certainly does have some order to it in those instances. But remember they also looked at the mortuary, the local clinics, and the private hospitals, which shouldn't have no order to them. So we've really got to call this nominal categorical data as well. So I really want to belabor to this point, we are not going to do any deep statistical analysis here. We don't have that knowledge yet. We haven't covered that in the course yet. All of the variables we are collecting the data on, are nominal categorical data, yet this is a powerful piece of research. So what analysis did they do if we had such a monotonous string of data types? Well, simple descriptive statistics. They mostly looked at proportions, and they expressed those proportions in percentage form. So, again, such useful information from this study. So much that can be done with the numbers that came out of this study, yet analysis was so simple. So what is the implication just of this study? And I really want to call a powerful, yet, with very simple statistical analysis. What the authors learned from this and what the local government learned from this, is really that proportions of blunt penetrating trauma is no different from the United States. And if we look at preventative measures. If we looked at funding programs for the management of this, it's certainly is the same as what we would find in the continent of the United States. But certainly quite different from what is found in Europe, where the levels of penetrating trauma are much less. A while ago in South Africa, there were some changes in the law trying to minimize the number of guns that are available. And certainly this research showed that there really was a decline in gunshot wound, in incidents of gunshot wounds, but still a very high rate of interpersonal violence where stab wounds. just fill up those numbers as far as far as the penetrating trauma is concerned. One very powerful piece of information that came from the study is that the number of patients who died before coming to hospital, are taken to the wrong hospital. Hospital perhaps that does not have the facilities. Remember we spoke about, the governmental hospitals that had different levels of trauma, as far as the units are concerned. And the small little community hospital really received patients for which they were ill-equipped. So they really showed up the immaturity in the prehospital system. And as far as the management of this epidemic is concerned, this paper really showed that that is where a great amount of effort has to go into really bringing up with that prehospital system, maturing it, putting some money and effort into that system. So really, very powerful information if you read through that article, put together through a very simple statistical mechanism. Very simple type of data that was included in this study, yet so much that we could learn from this case study.
