Hello, SD-cast listeners. My name is Christine Tang. In this podcast, I will interview someone in the System Dynamics or Systems Thinking community. This series is called “What is your SDory?”
This is Episode 7. Titled: “What is your SDory, Sema Çömez?”
Biography
Sema Çömez is a PhD Student in Industrial Engineering & Operations Management at Koç University in Turkey. She earned a master of science degree at Boğaziçi University and bachelors degree at Istanbul Technical University also in Industrial Engineering. She is the 2020 Lupina Young Researchers award winner and earned an honorable mention for the Dana Meadows Award of the System Dynamics Society for her work modeling heart failure.
Thank you for joining us today, Sema.
Sema Çömez: Thank you to invite me.
Interview
Christine: The first question is when and how did you ‘discover’ System Dynamics?
Sema: Actually, when I was a bachelor student, I first heard the Systems Dynamics term in one of [my] seminar course[s]. I don't remember who [was] the guest but he talked about the application of supply chain to System Dynamics. It is about the inventory optimization problem and he showed the results and we see the bullwhip effect on these results. It is about an ecommerce website if I remember right.
And in my bachelor education, I did not have a chance to take a System Dynamics course but after this course, I searched on the internet for System Dynamics and I promised myself to take a System Dynamics course in a masters program.
Christine: What was the first model that you built?
Sema: Except for very small size problems, like homework problems, I think the first model is our project [in a course]. When I [took] Professor Yaman Barlas’s course...we [worked on] one project after the semester and we [chose] a problem about the Turkish independent film industry because my teammate is a really huge fan of independent films. We applied some interventions on the model to make the Turkish independent films more accessible and attractable.
Christine: That’s very interesting...can you adapt that model and tell me how to make SD-cast more attractive and accessible?
Sema: [laughs] I can explain our suggestions for this sector and you can convert your case. Actually, the most problematic part of the Turkish independent film industry is that they are really low budget films and the people won’t know that the films exist because they can’t afford the advertisement costs.
And also, they’re art movies. The content is a little bit hard for most people.
Also, most of the cinema hall owners have contracts with the producers of mainstream films, for example, the Avengers or other action films as you know. And for these reasons, the independent films cannot have a chance to release in the cinema halls. They just release in the film festivals like Cannes or the Berlin film festival, etc. Even if they find a place in the cinema hall, the content of the films does not appeal to the audience. So the box office return is not enough to make a new film. So our first suggestion for directors. They can make more attractive films. They can be between the mainstream and the art house films and people start to know their names and the audience can be their fans and they can follow their other films.
For example, two important Turkish directors, I think they follow this strategy. One is Nuri Bilge Ceylan. The other is Fatih Akin. Maybe the people [who watch SD-cast] know [them]. I don’t know...but they are really famous in Turkey and in the Europe side.
And other...our suggestion is for the Ministry of the Culture. The Ministry of the Culture should have a grant for the directors. In this way, more films can be made and recognition of Turkey and Turkish film industry increases in the festivals which releases more films. And then also, with more films they can get a chance to get an important award so this industry can be sustainable. And also, I have one example, Nuri Bilge Ceylan won the Cannes in 2014 and now he gets funds from the Ministry of Culture.
They’re two suggestions from our work.
Christine: Thank you.
Sema: You’re welcome
Christine: We are also low budget.
Sema: Yeah...maybe for podcasts, Youtube and Spotify can be more proper ways of becoming more famous. I don’t know…
Christine: Yes. Raafat suggested that I list these on YouTube. I only did Google Podcasts, i[tunes] Apple Podcasts, and Spotify as you said...okay and thank you for trying to help me and trying to help others with your work on heart failure. So would you please tell us more about your Lupina Award winning work “Modeling the Dynamics of Vicious Cycles Involved in Heart-Respiratory Failure?” For those who aren’t aware, the System Dynamics Society’s Health Policy Special Interest Group grants the Lupina Young Researchers Award to an early career health researcher. Sema presented this work in the 2020 International System Dynamics Conference.
Sema: Yes, my work is related to heart failure. It is a little bit hard but I’ll try to explain now.
Heart failure is a really common disease in the world because when people are getting older and older, their hearts are also getting older and the disease comes from the deterioration of the heart, so it’s really common. And then, we looked at the medical literature. There are some sources that define the progress of this disease as a vicious cycle. And the vicious cycle in our terminology, it refers to a positive feedback loop (the reinforcing loop). And Yaman Barlas gave me one of these articles and it explains this vicious cycle and my works have been built on this article, this cycle.
My model includes the parameters of the blood circulation (we say the hemodynamics parameters) and water and sodium mechanisms. There are also the respiratory side and the hormonal compensation mechanisms and I added the treatment side. In a physiology model, first of all, you should define the negative feedback loops because the body mechanisms are based on homeostasis.
When a person is healthy, the mechanisms like the hormonal system can compensate for some anomalies in a short period.
So first of all, I built my model based on these negative feedback loops for validation. After the validation side is finished and I see that my results and I validate with the literature, I should simulate also the heart failure side. So I [added] some positive feedback loops. They reflected permanent structural change of the heart. And in this situation, when the heart deteriorates, the compensation mechanisms can not compensate for abnormalities.
For example, if you get an infection in [a] normal body, the body can compensate for the infection of it. But in the heart failure case, this infection creates problems on the heart side.
And in this situation, again, compensation mechanisms are triggered but now they cannot help the body. They force the heart [to] work more and more. And this pressure becomes really problematic for the heart failure patients because this positive feedback loop now acts as a vicious cycle and now this vicious cycle accelerates the deterioration of the heart.
In my model, I simulated acute heart failure, not chronic heart failure. I started my model with the chronic heart failure patients parameter. For acute heart failure, urgent treatment is really important because you should stabilize the patient as soon as possible. If you cannot stabilize the patient, deterioration continues and you cannot make the patient healthy now.
I applied two different treatment methods during my study. One of the treatments is a typical treatment method during the patient's hospitalization. And the other is my suggestion. I used two or three drug types that are very common and I changed the drug doses and the time period the patients take these drugs...and I want to find which one is the better treatment method. This, I think, is a good summary of my work.
Christine: Yes, thank you...and listeners (or readers) can see the links below for Sema’s 2020 International System Dynamics Conference paper and presentation on this topic. Just curious, Sema, did you incorporate any suggestions from the conference?
Sema: Yeah. The conference helped me to meet new researchers and new study areas. I think it makes my study more visible because it is...one of the master thesis in the field of Industrial Engineering. And maybe no one knows my study before the conference.
I got two awards from the conference. One of them is, as you mentioned, Lupina. The other is honorable mention Dana Meadows [Award]. The awards have both financial and academic values and for students, I think both of them is very important. Also, my master thesis period was very exhaustive and hard. And with these two awards, oh, I feel that all my efforts were worth it. Also, I had a chance to meet Professor Özge Karanfil in this conference and now we are working together. I think the conference opens many many doors for the students, for young researchers. During the conference, I tried to watch most of the studies related to the health side and the methodology side, and I got lots of new ideas for my future...maybe for my doctoral thesis. For example, I did not know people use the R language for building SD models and I saw one of the SD applications with the R language...so I [was] really surprised because I also use the R language for the statistical side...so yeah, it is a really good innovation for the SD side, I think. In summary, for staying up to date and following the newest research, participation in the conference is really beneficial, I think.
Christine: Yes. Thank you. So listen to Sema. Attend the System Dynamics Conference. [laughs]
Sema: [laughs] Yes.
Christine: Yes...and people in System Dynamics also use Python. And I know…
Sema: Oh really?
Christine: Yes.
Sema: I don’t know [this]...
Christine: and James Houghton at MIT created some packages for people to use. I think it is called PySD...anyway, that is not related to you. I just wanted to chat about this.
Sema: Thank you for the information. I will look.
Christine: Yes, and…the next question, what are you currently working on?
Sema: Now, I am a second year doctoral student in Koç University, again in the Industrial Engineering field. Koç University is one of the Turkish Universities. Actually, I wanted to apply to this university because when I searched for some professors on the internet, I found that one of the professors is researching about bioinformatics with machine learning. So I wanted to apply to Koç University because of this reason. I wanted to continue my work with more data but now, I still study System Dynamics. Yeah, maybe it is my destiny (I don’t know) and my supervisor is Özge Karanfil and also she is one of the students of Yaman Barlas. She is [from an] older generation. We are working on health system management models and models for medical education. For my doctoral thesis, I want to use a hybrid System Dynamics and Machine Learning approach. I hope I can do it. Most probably, our focus will be about the cancer diagnosis or screening strategies. I don’t know but I think 100% it will be health related.
Christine: Fantastic!
Sema: Thanks.
Christine: Are you also interested in other types of simulation modeling? I am a big fan of Nate Osgood and his work on hybrid or multi-method modeling (whatever terminology you like to use).
Sema: During my bachelor and master education, I worked on discrete-event simulation [and] agent-based simulation. I liked the agent-based simulation. Especially nowadays for the pandemic cases. I think [the] most accurate and most detailed models are using agent-based methodology. I like it but it takes really more time and coding skills (I think) for agent-based modeling.
Also, I don't have any chance to use Network Science. Maybe I can learn more things about Network Science in the future. Now, I am just focusing on System Dynamics and Machine Learning. I am [also] interested in the optimization and statistics fields as an industrial engineer.
Christine: So the next question is “What wiSDom do you have for students or those new to System Dynamics?”
Sema: The System Dynamics methodology can look simple if you are not a model builder.
If you look at the System Dynamics model, you can say, “Okay. I can make it.” But when you start building a model, specifically a hard problem, you understand that the definition of the causal relationships and formulation of the mathematical equations are not easy. So starting with a small model is very important for beginners. Also, Professor Yaman Barlas always said that the best model is the simplest model. So I think that being simple is very crucial...or starting with a simple model is very crucial. Also, for medical problems the key point is reviewing the literature very carefully, very detailedly. Become very patient. [laughs] In the articles and the books (as I see) many problems are still waiting for System Dynamics modeling. As we know and also the medical people know, the body mechanisms are based on homeostasis. This is the negative feedback loops. The body works with many many thousands, maybe thousands, maybe millions negative feedback loops. So I think it would be a very interesting and enjoyable experience to work on physiological modeling if you like to learn new concepts. I am in Industrial Engineering so the medical field is really interesting because when I open the physiology books, everything is really strange for me. But after I finished the research, I learned…[a lot of] things. So I recommend this to everyone because also the outputs are very satisfactory. Because if you validate your work and after your work, you have some results. You got some intuition. You got some know-how now for the doctors--for the physicians to show them the results and these results are mathematical results. They have some foundation with Mathematics and System Dynamics.
So when I look at most of the physiology models and the[ir] results, I think that cooperation with the doctors and the physicians are still not enough. I hope that, in the future, collaboration with the doctors and the engineers becomes more popular and becomes more common so the results of the System Dynamics models become more reliable. I hope.
Christine: We all hope. Yes...so thank you for explaining that. Do you have any tips on how to read medical literature?
Sema: First of all, you should learn the terminology because we are...I am an industrial engineer and most of the SD people are engineers and we are not familiar with the medical terminology. So first of all, I searched the medical dictionaries. And in the internet, you can find the dictionaries and after being familiar with medical terms, you can start to read the physiology books. And during my research, I read Guyton’s physiology book. It is one of the famous (I think) physiology books for medical students.
After reviewing topics from this book, I think that you can start to search for the articles because the articles don’t explain the context very detailedly. Most of the cases, they are just focused on the application so for the following topics you should know the terminology and you should also know the topics. I followed this way. And for searching the medical articles...database examples, I think the most popular is Pubmed and then MedLine, Embase [and] Cochrane are the really famous medical databases.
Christine: Thank you. Do you have any other fun or funny SD stories that you're willing to share?
Sema: I have one story. It’s not funny but I want to talk about it since it is the turning point in my System Dynamic career. When my masters thesis professor, Yaman Barlas, and I were searching for my masters thesis topic, unfortunately, his cat died from heart failure. In my mind, for my research, I had no thoughts about modeling heart failure or any medical topic because I [was] not familiar with this topic and I was researching environmental topics. But one day, Professor Yaman Barlas told me his cat died because of heart failure. And also, he researched this topic many many years ago because his father or uncle passed away due to this disease but he did not have a chance to focus on this topic but knows some articles about heart failure and also they defined these disease as a vicious cycle. So after his cat died, he remembered this topic and he suggested it to me for making this topic my masters thesis topic. And now, I think it is a really strange coincidence. It’s really sad but it’s really strange also. With this topic, I decided to work on health related topics. And now, in my academic life, I am still on this way and I hope in the future also, I will continue to study health related problems. So it’s my story. It’s my beginning of [studying] the health field.
Christine: Thank you for sharing that background story and finding something good out of something sad. And thank you for being our guest today on SD-cast.
Sema: Thank you so much for hosting me on your program. It is my pleasure.
Ending
Sema and other young modelers were inspiring to watch in the 2020 International System Dynamics Conference. The 2021 conference is July 25th - 30th. See the link below for more information: https://systemdynamics.org/conference/
Please also consider attending the free Student-Organized Colloquium which is on Sunday, July 25th: https://systemdynamics.org/2021-colloquium/
Below are links to Sema’s works and the Lupina Award (which we referenced in the interview): https://proceedings.systemdynamics.org/2020/papers/P1353.pdf
https://youtu.be/EJCWOm3MAxY
https://systemdynamics.org/lupina-young-researchers-award/
Now, here is a poem I wrote about Sema:
Sema Çömez
Students should do what she says
Start simple and learn the terminology
Before you jump into physiology
Sema won the Lupina Young Researcher award
And models the dynamics of health issues
To help prevent events that are untoward
By simulating our bodys’ mechanisms
To suggest treatment for our heart tissues
(Which can aid humans and other organisms?)
But we should exercise on a regular basis
To prevent problems like heart disease
Diseases like cancer create unease
Messes up our natural homeostasis processes
Vicious cycles can bring us to our knees
(No reinforcing loops, please!
Unless it is virtuous and creates an oasis)
So we must screen and act early
(Else some patients may become surly)
In order to prevent a potential tragedy
From this unfortunate malady
We can model these diseases that one abhors
To focus on treatment like Sema and her mentors
Thank you for listening to SD-cast. Please subscribe to SD-cast to hear more SDories.
Email me, ctang@wpi.edu, if you would like to be on SD-cast or recommend someone.
See below for the WPI SD Social Media accounts:
https://twitter.com/WPISDclub
https://www.linkedin.com/groups/1916314/
Sign up for the WPI System Dynamics Club mailing list:
https://bit.ly/WPIsdMailForm
Music:
Intro and End
“Limelight” by Podington Bear is licensed under the Attribution-NonCommercial 3.0 International License. I cut and moved the music track to fit the intro and ending.
https://freemusicarchive.org/music/Podington_Bear/Haplessly_Happy/Limelight
https://creativecommons.org/licenses/by-nc/3.0/
