Friday, March 17, 2017

Week 6!

So now that the data collection is all done, I figured I would go more in-depth on our actual body of patients: so essentially the demographics of our patient population. As I mentioned in early posts the patients were sorted into different BMI classes based on the WHO standards and % BMI lost classes which I decided on myself. However, even if the results come back and there is some statistically significant association, there is another issue. For example, say we find that we find underweight patients typically have worse outcomes overall. At first we might say, physicians should monitor pre-transplant and not go through with the transplant until the patient has reached a certain weight threshold. However, what if it's simply that underweight patients had simply not taken the chemo as well or their bodies are in a worse condition to begin with? In that case, increasing weight doesn't seem to help the situation. To lower the extent of this issue, I looked further into the demographics of our patient population.
Patient Diseases
Acute Myeloid Leukemia (AML)
Chronic Myeloid Leukemia (CML)
Chronic Lymphocytic Leukemia (CLL)
Acute Lymphoblastic Leukemia (ALL)
Myelodysplastic Syndrome (MDS)
Non Hodgkin's Lymphoma (NHL)
Hodgkin's Disease (HD)
Other
Credit: CIBMTR Summary Slides
As we can see, a large portion of patients die from their primary disease when it relapses.


However, some diseases are far more commonly treated with allogeneic stem cell transplantation than others and as we can expect, survival rates can be vastly different (e.g. with MDS and AML).
Pre-Transplant Conditioning Regimen
There are two types we can see in the diagram below: Reduced-Intensity Conditioning (RIC) or Myeloablative Conditioning.  Myeloablative conditioning is basically using the high doses of chemo and radiation with the goal of getting rid of everything in the bone marrow in preparation for the transplant. RIC is pretty self-explanatory,.
Disease Risk 
The stage of a person's disease also comes into play, with more advanced stages showing inferior outcomes. The  diseases were categorized as low medium or high risk based on CIBMTR standards. 
Here are some examples to help you guys out. Now examining the demographics of the population doesn't really "help" in any way. It's mostly used so that associations between weight loss and say overall survival don't have any underlying causes. For example, if we found a vast majority patients who lost >10% BMI within a 100 days had a low overall survival rate, but most of those patients came in with high risk AML, the association might mean less. 
Hope that helped and see you guys next week!

21 comments:

  1. Hi Justin. I was reading your blog and was curious to see if you found any major connections to BMI and survival rate. It seems that there are a lot of other factors to consider, but does BMI have a consistent effect on people or does it mostly matter on the person?

    ReplyDelete
    Replies
    1. Hi Michael. The data has not been analyzed yet so it's hard for me to answer this. Right not though, I have a feeling that it has more to do with the person. People at extremely high and low BMI's are of course less healthy and therefore, less likely to survive.

      Delete
  2. Hey Justin! Nice job organizing the data collected and coming to an answer on seeing a "new" factor that helps with your research. This post was definitely helpful seeing your analysis on the data collected.
    I am still kind of confused on the fact that the survival rates depends on both the type of disease and BMI right? So how is this going to help you with your research, what are you going to take from this data that is implemented in your research? Sorry for the confusing question. See you next week! :)
    - Sruthi Murala

    ReplyDelete
  3. Hi Justin. It's so great that you finished collecting your data and that you looked into the relationship between the demographics of the population and your project. Your charts gave me better insight into the relationship between survival rates and BMI. Can't wait for next week!

    ReplyDelete
  4. Hi Justin! Thanks for giving us graphs as examples to help us understand. They really helped me understand the relationship. How would you create a relation between BMI/diseases and survival rates?

    ReplyDelete
    Replies
    1. HI Richard. Disease really isn't one of the factors I'm comparing in the study. My study is comparing weight loss to overall survival, GVHD, and transplant related mortality. Checking disease afterward is to make sure that the relationship between BMI and the factors listed above are due to BMI and not disease. Hope that helped!

      Delete
  5. Hey Justin and welcome back! Could other conditions than those stated affect weight loss in this study. Are there more ways to keep as much patients in the study without allowing confounding variables interfere withe the results?

    ReplyDelete
    Replies
    1. Hey Akash. I think that's a good question. Since it's a retrospective study, it's really hard to get rid of the confounding variables because they are already set in place. At this point, I just hope to see the results and have an association between BMI and overall survival, transplant related mortality, and GVHD and have minimal other variables affecting this relationship.

      Delete
  6. Hi Justin! That seems interesting. While you were researching did you notice some diseases caused more weight loss than others.

    ReplyDelete
    Replies
    1. Of course! Many diseases such as AML are much more aggressive and needed to be treated with Myeloablative chemotherapy regiments which definitely cause a more severe drop in weight.

      Delete
  7. Hey Justin!
    The graphs are well done and help explain the topic. Did you have more research than the others for certain diseases? Also, out of these, was the data precise, or was it all scattered. Good luck with the rest of your project!

    ReplyDelete
    Replies
    1. Definitely. The source I got these graphs from is CIBMTR which basically takes data from transplants all over the world. Many diseases are much more common than others so there wasn't an even distribution of data across all of the disease. Other than that, the data was pretty precise.

      Delete
  8. Hey Justin! Woohoo! That is some pretty cool info, and it seems like you're doing a lot at your site! Do you expect you'll be done with everything by April 15th to see the data?

    ReplyDelete
    Replies
    1. I think so. The data should be back within a week or two so I'll have a week or two left to analyze it.

      Delete
  9. Hey Justin! Glad to see you've collected all of your data. Just wondering, did certain conditions of the patients affect the weight loss for each different patient?

    ReplyDelete
    Replies
    1. Of course. We might expect more aggressive diseases (e.g. AML) or stronger chemotherapy would cause more weight loss. However, hopefully, if the study goes well, we can see a relationship between pre-transplant BMI and weight loss that isn't due to these other factors (e.g. chemo regiment disease, etc.)

      Delete
  10. Hey Justin! You definitely have a lot of graphs and stuff. More than I have. I like how you are accounting for patient demographics to get a more complete understanding.

    ReplyDelete
    Replies
    1. The patient demographic section is kinda like a final check for the project. If there is a relationship between BMI and factors predicting the success of the transplant then we need to make sure there is no underlying, confounding variable (e.g disease, chemo regiment, age, etc.)

      Delete
  11. Hi Justin. You have quite the talent for presenting data. Why might the survival rates differ for some of the primary diseases?

    ReplyDelete
    Replies
    1. Thanks! It's just that certain diseases are more aggressive than others and treatments are more advanced for some diseases. For example, AML often requires high dose chemo and has a very low survival rate because it's more aggressive. However CML can be often be treated by taking an oral pill (e.g. Gleevec) for the rest of your life.

      Delete