How can I best analyze this data

#1
I have some data I am trying to analyze but was hoping to get some stats help with it. I'm not sure what the best way to analyze it is. I'm interested in seeing if there is an association between CD4 and viral load number, and various different kinds of pulmonary function test characteristics in patients with HIV.

My data is for a medicine research project. I have information from a bunch of patients with HIV. The data includes info on their CD4 cell count numbers, how high their virus load is, baseline characteristics, and what their pulmonary function testing shows. For the purposes of this study, I have decided to recode all of my measurement variables into nominal variables. For example, even though I have an actual number for CD4 count, I care more about whether the number is > 200, between 200-400, or > 400).

My independent variables are cd4 count and virus load (6 options for nominal variable - using ranges). My dependent variables are FEV1 (4 options), DLCO (4 options), and FEVFVC (4 options). I also have other baseline characteristics on the population include age, race, whether they have a history of heart disease, sex, etc). These baseline characteristics are also code-able as nominal variables.

I collected this data retrospectively by looking up patients with HIV and writing down all of these traits. I have spss, which I am okay navigating for the most part, so I have access to many different kinds of analyses.

Any help would be much appreciated.
 
#2
Hi YellowM3,

Seems you are interested in data analyzing. There are some great tools and software products which help to analyze, visualize and present the data in a simple way. I myself have used DataPlay (https://dataplay.us/) and can tell about it. Besides suggesting great tools for analyzing data, it is also integrated to PPT, so you can also make better presentations and present the data.
 
#3
Hello,

well for nominal variables you could use the Chi-Square-Test for dependency and check if there is any significant difference. If you keep the metric variables, then you could use T-Tests. Btw, if you recoded the viral load like 200, 200-400, etc... this is a ordinal scale so you can even do more tests with that!
Im not into that topic, but I would check the literature and see if there are any relevant and similar topics about this. Then you can check the methods used in the past for HIV analyses.

Greets!
 

noetsi

Fortran must die
#4
You can run multinomial logistic regression as well (I assume your DV levels are not ordered. If they are you can run ordered logistic regression instead).
 
#5
If you can transform your IV and DV variables into dichotomous variables, you could run odds ratios or logistic regression. Otherwise, run one IV against one DV using PHI or Cramer's V. Or you could do Chi-Square (or Cochran's Q Test if repeated measures).