Comparing blood levels at 2 time points between 2 groups

#1
Hey everyone,

I'm incredibly inexperienced when it comes to biostatistics. For my bachelor's thesis, I need to do statistical analyses, and I want to make sure that I don't mess up.

________

Introduction

My study is on a drug used during a chronic infection.

Study group: patients that have taken drug "A" for more than 5 consecutive years
Control group: patients that have never taken drug "A"

The study is retrospective.

For my thesis I want to find out the effect of drug A on the kidneys. This is measured through the glomerular filtration rate (GFR), which is in turn calculated from serum creatinine levels. A rise in serum creatinine means a drop in GFR, which in turn means kidney impairment. For the calculation of GFR from creatinine, age, gender, and ethnicitiy is used, meaning that these covariates are taken into account already.

I will have a serum creatinine value at t=0 and t=5 for each patient.
t=0 is the first blood measurement before they start treatment (either with drug A or with other medication).
t=5 is the first blood measurement after they have been taking their medication for 5 years

The difference between control and study group is a bit more delicate, since the control group does have to take medication (just not Drug A), but that is not important for my question.

Other cofactors are also recorded, such as CD4 cell levels in the blood at the two time points, age, gender, ethnicity, whether the patient is a diabetic or hypertensive, and how long they have had the chronic infection.

______

Question

So, at the end, I have serum creatinine levels at t=0 and t=5 for each patient, as well as a number of cofactors and demographics. Serum creatinine is converted to GFR.

My question is: how can I statistically assess whether or not there is a significant difference in the (hypothesised) decrement of GFR from t=0 to t=5 between the study group and the control group, and which covariates come into play?
 

Karabiner

TS Contributor
#2
My question is: how can I statistically assess whether or not there is a significant difference in the (hypothesised) decrement of GFR from t=0 to t=5 between the study group and the control group,
I'd think of repeated measures ("mixed") analysis of variance first.
You can incorporate co-variates there.
and which covariates come into play?
You ask for medical information in a statistics forum?
Or what do you mean by "come into play"? Generally
speaking, all factors which are different between groups
and have an impact on the outcome could be of potential
interest. But we don't know why some patients received
only A and others never received B, how much groups
differ with respect to crucial influencing factors, or
even how large the sample sizes are. Maybe propensity
scoring could be an option if it's desired to have
comparable groups.

With kind regards

K.