Which tests to use...?

#1
Hi,

I wonder if anyone could give me some advise on which statistical tests I can use to interpret data I have collected for a university project. The software I have available is Graphpad prism8.

I’m trying to find out if there is any correlation between data I have collected such as heart rate, respiratory rate, temperature, length of stay in ICU, length of surgery (and a few other similar values) and see if any of these have an impact on survival rates of my patients after surgery. ( animal patients - I’m at vet school :) )

Any help would be appropriated, I really struggle with statistics in general. Many thank, Lauren
 

Karabiner

TS Contributor
#2
How large is your sample size?

How did you collect data on survival/survival rate. Did you observe your subjects for a pre-specified time (such as 6 weeks, for ecxample) and recorded surviva/death at that date? Or did you collect data in a different way? Was there loss to follow-up of animals during the study?

With kind regards

Karabiner
 
#3
Hello Karabinar,
Thank you for your reply!
It’s a retrospective study. I collected the data from the animal hospital patient records. I used only subjects who had complete medical records with all the data within them. I have 39 cases: 20 cats and 19 dogs. All records contained information regarding survival or death after surgery. All had a diaphragmatic rupture and underwent surgery to repair it. Cases that survived were alive at the time of discharge and ones that died did so before they were discharge ( within the hospital )
Lauren
 

Karabiner

TS Contributor
#4
I’m trying to find out if there is any correlation between data I have collected such as heart rate, respiratory rate, temperature, length of stay in ICU, length of surgery (and a few other similar values)
The variables you describe are all interval scaled. You should create
x-y-scatterplots to visualize the associations between the variables.
You can correlate them using Pearson's correlation coefficient r and
Spearman's rank correlation coefficient rho. If r and rho differ markedly,
then there might be influential outliers in the data.

see if any of these have an impact on survival rates of my patients after surgery.
You have 2 groups: surviving patients and not surviving patients. You
can create box-and-whisker plots to compare the interval scaled variables
(heart rate, temperature etc.) between these 2 groups. If you want to
carry out statistical tests of inference, you can use independent samples
t-tests (with Welch correction, by default).

With kind regards

Karabiner