November 27, 2024

Exclusive interview with artists on their new piece added to the EPI 320 Museum Collection

By Epi Reporter

UW News

One of the artists’ of the group wearing the cloak.

The cloak, crafted by artists Kelly Le, Sophie Walker, Cora Cashmere, and Hathaway Bush, has been newly added to the EPI 320 collection at the Museum of Epidemiology. It was unveiled earlier today in the gallery alongside many other art pieces and projects as part of the showcase to introduce the concepts of epidemiology.

Pictured are the front (left) and back (right) sides of the cloak.

The opening of this new collection is part of the museum’s mission to make epidemiology more accessible and intuitive to understand. Today, I was able to talk to the group to gain some exclusive insight into their inspiration and get some clarity behind the meaning of certain details on the piece.

To start, can you all introduce your piece and tell us what inspired this cloak? What message were you all trying to convey?

What inspired us in making this artwork was first thinking about how this piece of clothing mainly surrounds our body around the upper torso. This led to us thinking about how our lungs are an important part of our health and very entwined with people’s well being. Such a large organ is susceptible to numerous different diseases and conditions. Smoking is a largely researched public health topic and we wanted to convey the association it has to lung cancer.

Our cloak tries to depict some of the tools epidemiologists use to calculate specific measures of association and represent what that measure might look like visually. We want to also highlight some of the strengths and limitations of the study design that was used from the data we chose.

Could you all go more into detail about the front of the cloak and what this imagery means?

Closer view of the front sides.

On the top, you can see the two variables that are the focus of this all. Smoking being the exposure on the left and lung cancer being the disease or outcome on the right. They are both in hazy clouds that come from cigarette smoke spanning across the front and back of the piece.

Of course, the two lungs on the front draw the most attention. We hope viewers at the museum notice that the pair of lungs are placed exactly where they would be on your actual body when you wear it.

Overall, we wanted to highlight the severity of this public health problem so we kept our color palette to reflect that. Using a combination of grays, blacks, and reds, we wanted to visualize the danger through this color language.

Balance scale imagery sewn and embroidered on left. 2x2 table depicted on the right side.

On the left side of the cloak is a balance scale. The cups on either side of the scale show the status of the exposure with that word embroidered across it. The arrows represent how exposure status may weigh more or less heavily in relation to the odds of the outcome. The reason that the two statuses are on either side of the scale is because we are making a comparison between the two groups. The exposed group may be more or less as likely to experience the outcome compared to the other non-exposed group on the other side. In our case we are discussing lung cancer as the outcome.

On the right side is a 2x2 table. This table is used to calculate measures of association in epidemiology. It is a useful tool when trying to determine how exposure and outcome are related to one another. A 2x2 table can answer questions such as does an exposure affect the outcome? Are the two variables positively or negatively correlated to one another? In order to set up the 2x2 table you must have data that relates the exposure and to the outcome. E+ represents populations that are exposed and in our case, smokers. E- represents unexposed populations, nonsmokers. The first column of the 2x2 are the possible outcomes: D+, those who have the disease/outcome, in our case lung cancer, versus D-, those who do not have the disease/outcome, in this case, do not have lung cancer. Then you can fill the data in accordingly based on your data, including totals. This 2x2 makes it very simple to calculate the odds ratio.

What is an odds ratio? Is there a reason you all chose this specific measure rather than any other used in this field?

An odds ratio is a statistical measure of association that is used to compare the odds of an outcome occurring between an exposed group and an unexposed group. Basically, it’s a tool epidemiologists use to help understand how much more or less likely the outcome is to happen in one group as compared to the other. We wanted to use this specific measure because we got our data from the NSDUH. The NSDUH is an online data survey that collects data nationwide using cross-sectional study techniques.

Since this data comes from a cross-sectional study where the exposure and outcome are taken simultaneously, we are limited to what kind of measures we can use. In epidemiology, some other measures you may commonly see are risk ratios or rate ratios. These two measures differ from odds ratio because they use incidence data. Incidence measures the rate of new cases over a period of time. Cross sectional studies gather information on an individual’s exposure and outcome status at the same time, so we are missing this time factor that would allow us to use these measures instead.

I’m not quite sure I’m understanding this all yet. What is the difference between risk and odds then?

That’s completely normal! It took us some time to grasp as well since the two are used interchangeably in everyday conversation.

Risk is more similar to a probability. Say for instance, 200 people are smokers and during the study, 40 of them develop lung cancer. We can say that the risk of this outcome is 40 out of 200. On the other side, odds is used to compare. The odds in this case would be 40 people to 160.

The big point here is that with risk, we can directly measure the relationship between exposure and outcome since we tracked that data over a time period and saw how it happened. With odds, the data being gathered is prevalence. We don’t know the order in which the outcome or exposure happens so we can’t directly conclude or measure the risk between them.

This is why it’s integral to consider what sort of study you’re dealing with and what kind of data you have. It’s important as epidemiologists to be careful in drawing conclusions of causality and associations!

You all mention something about ‘rate ratios’ earlier. What is that?“

Rate ratios also use incidence data like risk ratios, but they consider the question of how much faster or slower an outcome occurs in the exposed compared to the unexposed group. They require us to know the person-time of the population at risk, which is how much time a person has been at risk for a certain outcome. Again, since we don’t have this time component, we can’t use this measure of association.

It would is useful for cases where we want to see something like how many cases of lung cancer are happening over a certain amount of time in the two exposure groups. If we had a different data set that recorded the person-time measure, this would be interesting to look at!

I see! So in this case, how did you all calculate the odds ratio that’s written on the right side of the cloak?

In order to calculate the odds ratio you must divide those who are exposed and have the disease by those who are exposed and do not have the disease. This number will give us the odds of disease among exposure. The next step is to calculate the odds of disease among unexposed which is done by dividing those who have the disease and are unexposed by those who do not have the disease and are unexposed. After you get these two values, divide the odds of disease among exposure by the odds of diseases among the unexposed to get your final odds ratio.

Here, we used data gathered from 2022. As we drew on the right side next to the table, it would be 499,999 smokers who had lung cancer divided by 10,186,000 smokers who did not have lung cancer. The same thing is done for non-smokers and then you have the odds for both groups which is sewn on the back side. The final odds ratio comes out to be 1.99.

Okay, we have the odds. But, what do these numbers really mean in the context of everything?

Closer view of calculation results sewn on the back.

The odds of disease among those exposed to cigarette smoke is 0.0489. The odds of disease among those not exposed to cigarette smoke is 0.0246. 1.99 is the odds ratio between the two, sewn right in between the cloak.

Our interpretation of these numbers is that, “Those who are smokers are more likely to experience lung cancer. The odds of lung cancer was 1.99 times AS likely in smokers compared to non-smokers.”

The individual numbers are placed in two areas to represent their exposure status and there are figures of people colored red and white to represent their outcome status. We visualized our interpretation by showing more red people within the black cigarette smoke area compared to the little amount seen in the space outside of the smoke.

If the odds between the two groups were the same, the odds ratio would be 1. This means lung cancer would be equally likely in both smokers and non-smokers. We got an odds ratio of 1.99. This means for every 1 non-smoker who develops lung cancer, approximately 1.99 smokers also experience it.

We think a key point to consider when interpreting numbers in this study and any others that viewers may encounter is the ‘as likely’ wording. Again, odds ratio is not directly measuring risk or causation. In casual conversations, people may say ‘more’ or ‘less’ likely but this unintentionally suggests that there is causation between smoking and lung cancer that has not been confirmed yet with this specific measure of association.

I noticed some more symbols on the back side. Could you speak more about the imagery and meaning of those?

Closer photo of imagery sewn and embroidered.

We embroidered a camera with an unspecified date, a person running, a magic 8 ball, and a money sign are depicted. These represent the benefits and limitations of the study we chose.

As mentioned before, the study we used is a cross sectional study. These types of studies use prevalence data and are from a specific point in time. To represent this, a camera with an unspecified date is shown. Because these studies are from a certain time frame, it is impossible to discern if the disease or exposure came first. Because of the “snapshot” nature of cross sectional studies, there is a chance of temporal bias, or an error that occurs when the time of data collection affects the studies outcome. The magic 8 ball also shows this idea of guessing. The benefits of this study are that it is fast and cheap, represented with the bill and person running. Finally, the thought bubble is to show that cross sectional studies are limited to hypothesis generation. More research and different study designs should be use to answer further questions outside of association.

We have read other studies and it is widely known that smoking affects the risk of getting lung cancer. Unfortunately, as stated before, if researchers were to use this study alone, they would not be able to definitively come to this conclusion because of the hypothesis limitation. Despite this, cross sectional studies are a good way of finding potential associations quickly without a big budget.

Was there anything challenging about creating this piece?

Finding the materials and making do with what we have was a struggle. We were able to source the fabric from friends and family, but there were some materials we had to buy such as fabric markers and embroidery floss. The process of coordinating a group project was also difficult since there were four of us, but we were satisfied with how our piece turned out. Each of us worked on the four separate panels that make up the cloak. Our skills varied but we were able to collaborate to cover each others strengths and weaknesses. The process of making this piece together reflects a lot on how projects and research is done in the public health field. Communication and collaboration are integral so this experience in itself was valuable to each of us.

Thank you all for your insightful answers! To wrap up the interview, what do you hope viewers can take away at the end of this experience?

We hope that our cloak was visually interesting to look at and is able to capture your attention. We also hope that viewers leave feeling like they have a better understanding of one of the measures that epidemiologists use to make sense of data and create interpretations of real world public health issues.

We want to make it clear that epidemiological studies don’t necessarily have to be confined to tables, graphs, and reports but that they can also be translated into many creative ways beyond academic reporting.

That wraps up the interview with the four artists. The Museum of Epidemiology is open on weekdays from 8am to 6pm and the EPI 320 collection will be available to visit until the end of this autumn quarter. Tickets are free for UW students so please do visit and take a look at the showcase to get an introduction to epidemiology.

For more information, contact Epi Reporter at EpiReporter@uw.edu