Hi. So I'm here at the Data Science Institute at Imperial College London. So up until now, we've been talking and thinking about how we work with these big health care data sets that we have access to due to the evolution of digital health. But here at Imperial College London, we're making use of a data observatory that you can see here around me to study and interrogate these large health care data sets. One of the questions that we're really interested in focusing on is looking at the public health issue of antimicrobial resistant infection. So antimicrobial resistant infection is a huge global health care threat. It's estimated that it costs about 700,000 deaths per year across the world. If left unchecked, so if we don't take any action, it's estimated that the death toll could rise to 10 million by 2050. So here we're using publicly available data sets from England to study the surveillance of these infection and look at the prevalence of antimicrobial resistant infection across regions of England. So I'll take you through some of the health care data sets that we're working with and show you how data visualization can help us to look at aspects and interrogate these data specifically around antimicrobial resistant infection. So here we have some of our surveillance data that's been provided by Public Health England. So here we have a map of England, and each of the geographical boundaries that we can see represented, represents a different health board locality within the country of England. What we're looking at here is the prevalence, so the number of E coli bacteremia infections within each of these regions of England. So E coli bacteremia infections are caused when an E coli bacteria enters the bloodstream. So E coli is a gram-negative bacteria that commonly causes urinary tract infections, but it's a much more serious infection when these E coli bacteria enter the bloodstream. So what we're looking at here is the prevalence of these serious infections across England and within different regions. We can also look at how the prevalence of these infections changes over time, as represented on the visualization here. But what's really great for us as researchers is the level of information that these visualizations produce. So for example, we can begin to look at areas in different parts of the country where there's relatively higher prevalence of these types of serious infection, and as researchers, we want to know more about why there's such a higher prevalence of these infections in these areas. So we can then conduct further analysis looking more in depth at specific locations. We can also look at potential outliers in our data set, so for example, in the northeast of England, colored in purple here, we can see that there's a relatively low prevalence of these E coli bacteremia infections. Again, we want to know more about what's happening in this region. So visualizing the data in this way is a really useful method for us as researchers to interrogate and gain more insight into these infection surveillance data sets. So we've had a look at our prevalence data relating to E coli bacteremia infections at an English-wide country level. Now the next question that we're interested is to try and find out the proportion of those E coli bacteremia infections that are resistant to antibiotic treatment. So we're looking here again at a map of England, again with our geographical boundaries set to being health board. But we're looking at the prevalence of E coli infections that are resistant to commonly used antibiotics that are used to treat E coli infections. So namely they're cephalosporins and the ciprofloxacins. What we can see is that overall generally were quite good at treating these infections. So there's quite a high prevalence of susceptibility to antibiotic treatment for E coli infections across England. But once again, similar to the previous graph, using this data visualization tool, we can begin to pinpoint areas of the country that are of concern. So we can begin to highlight regions such as this one in the northeast of England, where there's relatively higher prevalence of antibiotic resistant infection. We've looked at the prevalence of E coli bacteremia infections across England, and we've also looked at the proportion of those infections that are resistant to commonly used antibiotic treatment. But we can also use data visualization tools to examine risk factors associated with these antimicrobial resistant infections. So here we're looking similar to before, at a map of England with the geographical boundaries corresponding to health board levels. But this time we're interested in looking at antibiotic prescriptions. We know from the literature that antibiotic prescribing is associated with an increased risk of acquisition of an antimicrobial resistant infection. So we're really interested in examining antibiotic use as a risk factor for antimicrobial resistance. Similar to before, we can see that there's a wide variation in antibiotic prescribing across England. What data visualization helps us to do is to pinpoint those areas we may want to concentrate on in our data analysis. So for example, in the northeast of England, we can see a relatively higher rate of antibiotic prescribing, and what we're really interested in seeing is whether or not these regions that have a relatively higher level of antibiotic prescribing, are correlated with those regions which also observe a higher prevalence of antimicrobial resistant infection. Finally, we can bring all of this information together on a bubble chart as indicated here. So what we've done is we've taken all of the information that we've presented on the maps and we've brought it together so we can look at associations over time. So here, each of the bubbles on our chart denotes an individual health board, or CCG as they're labeled here. The size of the bubble indicates the number of E coli bacteremia infections, so similar to the information that was presented in map one. We can also look at the proportion of these infections that were resistant to commonly used antibiotics, and this is depicted along the x-axis here, where we're looking at E coli bacteremia infections that were susceptible, or conversely resistant to cephalosporin antibiotics. The information then about our antibiotic prescribing usage is presented on the y axis here, and this is the number of antibiotics that are prescribed within each individual health board. This is a standardized measurement, so it's weighted according to the proportion of females in a health board, the general age of the population, and also the number of patients within that individual health board. So what this chart allows us to do, is to begin to depict individual health boards or CCGs, and look at what's happening in terms of the association between antibiotic prescribing and antimicrobial resistance over time. We can look at how these trends change and how they change in relation to other health boards. So thank you very much for joining us today. I hope you've enjoyed the time in the Data Observatory here at Imperial College London, and hopefully some of these data visualization tools will be useful for yourselves going forward with your future research.
