So all of my funding for this trip was paid for by a program from the University of Washington. The idea is to give American students an international research experience in Public Health fields. Students all over the country are eligible and they selected 8 fellows for the program this year. The program has sites all over the world: Mexico, Chile, Peru, Ethiopia, Thailand, Vietnam, and Australia because the University of Washington has established connections with local doctors and/or public health researchers at these sites. However this year they only sent the fellows to three locations: 4 students are in Australia, 2 Students are in Ethiopia and 2 students (Andrea and I) are in Peru! MIRT has been sending students to Peru for ~15 years and they always work with Sixto Sanchez, a local OB/GYN here. Hence our projects are focused on pregnant women!

So my project is officially called “Risk of Spontaneous Preterm Birth in relation to Maternal Depression, Anxiety, and Stress.” The study is a “Case-Control” study. This means that for every “case” of Preterm birth (Birth before 37 gestational weeks) that the study registers in our database, there is another “control” that is registered for women who give birth after 37 gestational weeks (ie. an average pregnancy is 40 gestational weeks ~ 9 months). So before I got here, last year, the Doctor and his team collected data on 959 women who gave birth at 3 local Peruvian hospitals (with the informed, written consent of the women of course!). We collected data on general demographics but also on their levels of stress, anxiety, and depression using a standardized mental health questionnaire. SO all of this data is entered in this big data base and it is my job to open that data base in a specific type of statistical software (called SPSS) and analyze it. For example, say I want to compare the average depression score (using our scales) of a preterm birth case versus one of our normal controls. I do a specific test using the software and the computer outputs the results. What I find is that the mothers who gave birth preterm often have 2 points higher score than the control mothers!

Things get a bit more complex than this when we do something called “Logistic Regression” which is when we adjust for other factors that might affect our results. For example, though that test I described above gives reason to believe that Depression might “cause” preterm birth, there may be some underlying explanation. For example, maybe a woman did not plan her pregnancy and is not economically prepared for it so she becomes depressed. However because she did not plan her pregnancy, she also does not take prenatal vitamins and more frequently consumes alcohol, eventually leading to a preterm birth. Thus even though it seems that higher depression score means preterm birth, it is really the fact that she did not plan her pregnancy, does not take prenatal vitamins, and drinks a lot more. This is just an example that I made up, but thats why we need Logistic Regression, to get rid of these other factors and really see the relationship between just depression and just preterm birth. So in the computer I will set up an analysis factoring out all these additional cofounders (ie. Alchohol consumption, Planned Pregnancy, Maternal Age, Vitamin consumption) and the computer does the plug and chug and voila: I get something we call an “adjusted Odds Ratio.” What this means is that the computer will give me a number such as 2.08 (and a 95% confidence interval of that number). The number (which is about 2) means that there is a 200% increased risk of Preterm Birth with high Depression score and thats AFTER we factored out for all those other things (Vitamins, age…). SO seeing such a high number makes us happy as statisticians! A lower number, such as 1.12 is nearly 1 which makes us not as excited because having a score of 1 means that there is no difference on pregnancy outcome if you are depressed or not.

Then, after doing all of the statistical analyses we have to share and discuss our results with the world! So, now I am working on writing a paper sharing our findings. Our results are pretty exciting but I can’t share them until they are published (maybe sometime this fall) in a scientific/public health journal. So once that gets published (really cool since I am so young) I will definitely post it!

Sorry, I know that was a lot but I guess I just wanted to explain what I am doing here so you know whats up when I say “This summer I did statistical analysis on a public health project!”