Stuck on some basic stats...anyone up for a challange?

[Enik]

---

Members do not see this ad.

Hello stats wizes,

I just started a small side clinical research project and need to calculate something very basic and for some reason I cannot seem to get my head around this. Basically I need to calculate p-values of adverse events repported in patients and make an adverse events table

 

[Mad Jack]

SDN is not your personal tutor.

 

[fakeplastictrees]

Your question is a bit vague. An example paper similar to what you want to do would be helpful. You might want to consult with a statistician and/or consider using more powerful statistical software, such as R or SAS.

 

[zhopv10]

I would highly suggest learning a statistical analysis program if you are serious about doing research. I personally use R as it is free, powerful, and has tons of support and tutorials available but it does have a learning curve that took me quite a bit of work to get proficient at.
Now it looks to me like you are just comparing to see if the number of occurrences is different between the two groups (similar to a table one/patient demographics table). In this case I would suggest chi-square for categorical variables and a two tailed t test for the continuous variables. If you use R the TableOne package is awesome for this.

 

[Krukenberg]

Chi square test for each adverse event because the two variables are categorical (group 1/2 and adverse event yes/no) and the numbers in your table are counts. Set each adverse event up as a 2x2 table and you will see. For the first adverse event the cells of the 2x2 would be 13, 15, 298, and 187 (if you read it like a book).

Look at a paper for a randomized trial that also presents toxicity data. The table is set up like this, with a p value for every adverse event.

 

[zhopv10]

That's embarrassing... I said ANOVA in my post but totally was thinking and meant chi square haha! Thank you krukenburg, I would have lead him astray! (I edited my first post to reflect this so as not to confuse readers in the future, but wanted to make sure krukenburg received due credit for pointing it out)

 

[wholeheartedly]

Biostatistics for Dummies is actually a pretty good basic overview for this sort of thing. You should get a book or take a bascic stats class so you know what you're doing, why you're dong it, and what the assumptions and limitations are when you do it.

I also agree with learning a statistical programming language such as SAS or R, (other options are STATA and JMP) Both have steep learning curves but are very powerful once you get the hang of it and both have tons of online support and resources. R is free and can be used on most systems, SAS costs money and the later versions don't support Macs so you'd have to partition your system and run another operating system to use it. I've had quite a bit of coursework in SAS and am just learning R, but R seems more intuitive and user friendly so far.