Univariate Distributions
Overview
Depression remains a leading problem in the United States, with it affecting an average of 1 in 5 people worldwide and resulting numerous adverse symptoms. What’s more, the World Health Organization found that depression causes the greatest number of Disability Adjusted Life Years (DALYs). Recent research has found promising data showing a link between fiber consumption and depression. I am using data from the National Health and Nutrition Examination Survey (NHANES), which is a nationally representative sample of people throughout the US. I aim to understand the cross-sectional associations between fiber and depression to better understand how fiber deficiency affects people with depression in this sample.
Depression Score Histogram
The following graph shows the distribution of depression in the NHANES data set. Any score greater than 10 is considered to be clinically significant depression. Visual assessment shows that the majority of people have no symptoms of depression, and a small of amount of the sample have clinically significant depression.
Fiber
Fiber Consumption Histogram
The follow graph shows the distribution of fiber consumption in the NHANES dataset. This is using a variable that takes the average of two dietary recalls. Any score greater than 28 grams is meeting the daily value for fiber. Visual assessment shows that the majority of people in this sample are not meeting the daily value for fiber.
As you can see, the majority of people in this sample do not have clinically significant depression and are not meeting the daily value for fiber.
Bivariate Distributions
Fiber consumption by depression scores
By comparing quartiles of fiber consumption to depression scores, we can better understand how these issues affect the population. From this interactive boxplot, we can see that there is a difference between the quartiles, though it is not statistically significant. Further analysis applying survey weights and addressing major confounders would add greater context to this analysis. These results show, from a cross-sectional standpoint, there is only a small difference between depression rates in different fiber quartiles.