which statistical test to choose?

[bburugap]

Hello all,

I am naïve to statistics. I need help with one of my research projects. Below is my data:

Dependent variable: change in hypertension (I have baseline measure and one year-after treatment measure)-continuous variable

Independent variable: 1. Drug groups (drug a, drug b, drug c): one patient on one drug only

2. Sample groups (sample a, sample b, sample c): one patient might present in more than one groups.

The question is to find out the changes observed from baseline to one year-after treatment hypertension reading when controlling for factors like age, gender, race etc.

What is the appropriate statistical test to use?

Thanks in advance,

BB

[art297]

There is a conflict in what you said. Did any patients get multiple drugs?

If so, they may have to be excluded. Other than that, sounds like a repeated measures ANOVA (probably proc GLM) see: https://www.google.ca/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwj4zcfNw4TTA...

Art, CEO, AnalystFinder.com

[bburugap]

Thanks for your reply.

None of the patients were on multiple drugs.

so, One patient present in one drug group only.

But one patient might be present in more than one sample group.

There are two Independent variables.

1. Drug groups

Drug a

Drug b

Drug c

one patient on one drug only. (patients on multiple drugs were excluded from the study)

1. Sample groups

Sample x: Patients with disease CA/PH

Sample y: Patients with only disease CA

Sample z: Patients with disease CA/PH with no previous drug history

one patient might present in more than one groups. 

[Reeza]

Post sample data that is representive of your actual data. 

[bburugap]

Thanks for your reply!

There are two Independent variables.

1. Drug groups

Drug a

Drug b

Drug c

one patient on one drug only. (patients on multiple drugs were excluded from the study)

1. Sample groups

Sample x: Patients with disease CA or PH

Sample y: Patients with disease CA only

Sample z: Patients with disease CA or PH with no previous drug history

So, one patient might be present in more than one groups.

Below is how the data looks like.

pid	drug	disease	drug_history	sample_group
1	A	CA	1	x
1	A	PH	1	x
2	C	CA	1	x, y
3	B	PH	1	x
4	C	PH	0	x,z
5	B	CA	1	y

Thanks,

BB

[LucyB]

what exactly is the significance of the sample group?